"Either a zoonotic virus crossed over to humans fifteen miles from the biggest coronavirus laboratory in the Eastern Hemisphere."
As they say, 'you couldn't make it up'. Or, 'truth is stranger than fiction'. It does *seem* very suspicious co-incidence that a new virus that would turn into a world-wide pandemic just happened to pop up out of nowhere on the doorstep (as it were) of a specific institution dedicated to doing this kind of research, but there was no connection. And yet we seem to be forced to accept that this is so.
Congratulations and gratitude for sitting through 15 hours of video to produce this post! And I did appreciate the Linear B/Lineage B joke.
I would also add that concerns about a lab leak in Wuhan are not post hoc. There were concerns about lab safety in Wuhan in the scientific literature published from 2015 onwards. If you were to ask people in the know prior to the pandemic - where in the world would there most likely be a lab accident with a coronavirus, Wuhan would have been top of the list given that it has the world's largest collection of bat coronaviruses, and did incredibly dangerous experiments on them at BSL-2.
Concerns about the wet market are not post hoc either. An Australian virologist posted pictures and video from visiting the HSM back in 2014, noting it would be an ideal breeding ground for coronavirus pandemic.
Actually, Wuhan was used as a negative control for coronavirus antibodies in a WIV study prior to the pandemic. i.e. It was thought a v unlikely place for a spillover prior to the pandemic.
Shi Zhengli has said she never thought a spillover would occur in Hubei
Also, the Australian virologist you refer to (Eddie Holmes) was in Wuhan because he was visiting the Wuhan Institute of Virology! He collaborated with the WIV on RatG13, one of the closest known relatives to covid, in 2018 (and then apparently forgot that he had done this_
...and therefore... what? I don't understand your point here at all. I know he was visiting WIV. We already knew about RatG13. So?
Gonna combine 2 threads to reduce confusion:
> Shi Zhengli has said she never thought a spillover would occur in Hubei
Did she think a spillover would occur in the WIV? It's kind of weird to have this sentence right after your prior one... are we just cherry picking experts now? Some people identified wet markets, including HSM, ahead of time. Some people identified WIV ahead of time. Seems like kind of a wash to me.
There's extensive bat populations in the cave systems of Enshi in western Hubei aren't there? And the whole province is within the range of the rufous horseshoe bat. There wouldn't be large populations in the area around Wuhan itself since it's heavily urbanized, but hunters would bring them into the city to sell them given that its the provincial capital and richest city in the nearby area by a far margin.
Also raccoon dogs are common across rural areas of central and eastern China and are another possible source.
Obviously I don't know what the internal reasoning of the Chinese CDC was for deciding where to put their research into coronaviruses. But Wuhan, as 7th largest city in China, fairly centrally located, and in the region of the relevant animals would seem a good choice. Chengdu is the only larger city in central China, but is much further west and has less tech sector, so presumably would be harder to get good people
You also couldn't make it up that the virus appeared in a wet market that was identified years ago as a plausible place for a pandemic to appear. These kinds of coincidences are the rule, rather than the exception, for real-world events.
I was once at Disney World (over a thousand miles from my hometown of population 3000) and met a person from my hometown who also happened to be there that day. We had 0 idea the other was going.
This is a lot less weird than if I met them at a random gas station in rural Mongolia, but it's still pretty weird!
While on a high school trip, my classmates & I ran into someone else from our class (not associated with the extracurricular organization that had arranged the trip) on the second level of the Eiffel Tower.
I once was on random free city tour in another country ~5,000 from my home with maybe 10 people in total and met another person who lived within five miles of me back home (same suburb, so not like a city of millions either).
Or as Jon Stewart put his perfectly natural skepticism:
“‘Oh, my God, there’s been an outbreak of chocolaty goodness near Hershey, Pa. What do you think happened?’ “Like, ‘Oh I don’t know, maybe a steam shovel mated with a cocoa bean?’ Or it’s the [expletive] chocolate factory! Maybe that’s it?”
I've seen the argument that it's suspicious that COVID appeared near one of China's only two BSL-4 labs, but in this debate the *proponents* of the lab leak hypothesis said that the WIV recklessly did coronavirus research in a BSL-2 lab, not in the BSL-4 lab. (The WIV has a BSL-4 lab as well as BSL-3 and BSL-2 labs.) It's less suspicious that COVID popped up near a coronavirus lab if dozens of labs do such research than if this is one of the few such labs in the world.
Yeah, the emphasis on the pro-lab leak side puts less emphasis on the BSL-4 designation and much more weight on "there are/were two labs that were at the forefront of bat coronavirus research, WIV and one in North Carolina." Plus WIV seems to have been doing GoF research.
Cities are designed to make distance and geography irrelevant. They aren't perfect about it, but the intent is that your interactions are defined by your social and economic network, and the physical infrastructure allows you to hop between those nodes without concern for what boring irrelevant-to-you terrain is between them. You probably barely notice that stuff when you're on the freeway, and you can't see it at all from the subway,
Again, that's only an ideal which is imperfectly approached. But what we need to know isn't "what's the distance from the WIV to the HWSM", it's "how many WIV lab techs have families with a taste for fresh seafood or exotic bushmeat?" If the answer is "lots", then the 15 miles won't stop them and/or their wives from schlepping their virus-laden selves over to the market.
Or maybe there's some class or cultural divide that means nobody working in the WIV's labs is going to shop at the HWSM; I don't know. That information, which would be very informative here, is hard to come by. The geographic distance, which isn't nearly as informative, can be read off any online nap; a few clicks and calculations, So that's what most people use; looking for their lost keys under the streetlight.
My prior, starting from ignorance, is that the odds of an average WIV lab tech('s wife) visiting the HWSM, is the same as the odds of an average Wuhan resident visiting that market. Which is something we can calculate.
> Cities are designed to make distance and geography irrelevant.
They really aren’t. Cities are the slowest way to travel from A to B by car - which is what we are talking about with the 15 miles. It’s shorter as the crow flies. As for public transport we would have to look
Into it I suppose but if the lab Typhoid Mary was taking the infection to the wet market on a train then you expect the spread of the disease along the track of the line.
My belief is a zoonotic answer but if the virologists come out with a conclusive peridot that it’s artificial the lab leak isn’t the only game in town.
Several recent papers challenge the core arguments relied on for Huanan Seafood Market origin in the Rootclaim debate. Also, Miller incorrectly claimed the N501Y mutation would result from passage in hACE2 mice (he mixed them up with BALB/c mice).
1. Samson et al (2024) estimate likely point of emergence between August and October 2019. Rendering the December 2019 cases and January 2020 market samples irrelevant to origins.
They refer to horizontal gene transfer of the spike gene from BANAL bat CoVs in Laos to pangolin CoVs occurred in mid 2018. "Horizontal gene transfer" can occur in the wild (rarely) or can be induced in he lab artificially "transgenic". In other words a nice code word for lab engineered virus leak.
Samson S, Lord É, Makarenkov V (2024) Assessing the emergence time of SARS-CoV-2 zoonotic spillover. PLOS ONE 19(4):
2. Dietrich Stoyan, Sung Nok Chiu, Statistics did not prove that the Huanan Seafood Wholesale Market was the early epicentre of the COVID-19 pandemic, Journal of the Royal Statistical Society Series A: Statistics in Society, 2024;,
3. Lv et. al. (2024) found new intermediate genomes so the multiple spillover theory is unlikely (it was anyway given lineage A and B are only two mutations apart as François Balloux, Virginie Courtier-Orgogozo and Jesse Bloom have observed). Single point of emergence is more likely with lineage A coming first (as Nick Longrich points out below). The market cases were all lineage B so not the primary cases. Their findings are consistent with Caraballo-Ortiz (2022), Bloom (2021).
Jia-Xin Lv, Xiang Liu, Yuan-Yuan Pei, Zhi-Gang Song, Xiao Chen, Shu-Jian Hu, Jia-Lei She, Yi Liu, Yan-Mei Chen, Yong-Zhen Zhang, Evolutionary trajectory of diverse SARS-CoV-2 variants at the beginning of COVID-19 outbreak, Virus Evolution, Volume 10, Issue 1, 2024
4. Jesse Bloom (2024) published a new analysis showing that genetic material from some animal CoVs is fairly abundant in samples collected during the wildlife-stall sampling of the Huanan Market on Jan-12-2020. However, SARS-CoV-2 is not one of these CoVs.
Jesse D Bloom, Importance of quantifying the number of viral reads in metagenomic sequencing of environmental samples from the Huanan Seafood Market, Virus Evolution, Volume 10, Issue 1, 2024,
5. Michael Weissman (2024) shows a model with ascertainment collider stratification bias fits early Covid case location data much better than the model that all cases ultimately stemmed from the market. George Gao, Chinese CDC head at the time, acknowledged this to the BBC last year - they focused too much on and around the market and may have missed cases on the other side of the city).
Michael B Weissman, Proximity ascertainment bias in early COVID case locations, Journal of the Royal Statistical Society Series A: Statistics in Society, 2024;
6. Chen et al (2024) discuss how using the Grunow-Finke assessment tool lab origin looks at least as likely.
Chen, X., Kalyar, F., Chughtai, A. A., & MacIntyre, C. R. (2024). Use of a risk assessment tool to determine the origin of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Risk Analysis, 1–11. https://doi.org/10.1111/risa.14291
7. The anonymous expert who identified coding errors in Pekar et. al. leading to an erratum last year has found another significant error. Single spillover looks more likely. t.co/GAPihZu51P
8. The argument that an engineer wouldn't make the furin cleavage site with the features of SARS-CoV-2 overlooks it resembles that of MERS in several structural and functional ways, and the sequence looks quite similar. In 2019 WIV researchers were involved in MERS research. Dr Andreas Martin Lisewski discusses similarities with a MERS infectious clone described in 2017 here. t.co/fAVUlJu0TK
9. Broad Institute biologist Alina Chan also observes the S1/S2 FCS PRRA insertion in SARS-CoV-2 generates a Class IIS restriction enzyme site (BsaXI). This was used by WIV and Ralph Baric at UNC previously. The full DEFUSE proposal available since the debate strengthens the argument of Bruttel et al. Specifically, the use of BsmBI, 6 fragments, and leaving the sites in). https://usrtk.org/covid-19-origins/scientists-proposed-making-viruses-with-unique-features-of-sars-cov-2-in-wuhan/
What you say about raccoon dogs here is mistaken. Raccoon dogs are not a plausible intermediate host for sars-cov-2 on the basis of information that has been known since 2021. There are several considerations.
1. Xiao et al (2021) - https://www.nature.com/articles/s41598-021-91470-2%E2%80%8B%E2%80%8B%E2%80%8B , which includes a co-author of Worobey et al (2022), a leading zoonosis paper states in table 1 that the raccoon dogs were wild caught in Hubei, not farmed as you assert in the piece. This alone rules out raccoon dogs as plausible hosts for two independently sufficient reasons. Firstly, there is unanimity in the literature that the bat ancestral virus to SARS-CoV-2 is in southern Yunnan or South East Asia. Everyone agrees with this, including Shi Zhengli. If a species was wild caught in Hubei, then there would be no explanation of how it acquired the ancestral bat virus, given that Hubei is 1000 miles from southern Yunnan.
Secondly, a mystery of sars-cov-2 is how it acquired the furin cleavage site that makes it so transmissible. There are 850 known sars-like coronaviruses, and only one with a furin cleavage site. According to private messages exchanged by proponents of zoonosis, the furin cleavage site could not have been acquired in the market because the density of animals was too low (only 3-4 per cage). When avian influenza acquires a furin cleavage site that occurs on farms with thousands of chickens densely packed, i.e. not in the wild and not when there are a handful of animals in cages in a market. https://usrtk.org/covid-19-origins/visual-timeline-proximal-origin/
2. Wang et al (2022) https://academic.oup.com/ve/article/8/1/veac046/6601809 also confirms that the raccoon dogs were wild caught in Hubei. What's more, Wang et al (2022) tested 15 wild raccoon dogs of suppliers of Wuhan markets, including the Huanan market, in January 2020 and found them to be negative for SARS-CoV-2. On average, 38 raccoon dogs were sold across the four markets in Wuhan from 2017 to 2019. So, the 15 raccoon dogs likely comprised nearly the whole inventory of raccoon dogs that would have been supplied to the Huanan market at the time.
In short, the raccoon dogs supplied to the market were tested and were negative. It is very strange that the raccoon dog narrative has persisted for this long given this public information. Even the strongest proponents of the raccoon dog hypothesis have walked back their bold claims that raccoon dogs are thost.
Xiao et al (2021) has a list of species sold at the Huanan market. I would encourage you to read that list and suggest which animals you think are plausible, and I will tell you why they are not actually plausible.
I don't think there's a consensus that the ancestral virus is from Yunnan or SEA. No plausible immediate ancestor has been found anywhere. For what it's worth, high seroprevalence for SARS-related viruses have previously been found in Hubei:
Second, the line between wild-caught and farmed animals in the wildlife trade is blurry. For instance, there was a wildlife farm in Hubei where SARS was detected in palm civets. The farm contained a mix of wild-caught and farm-breed civets:
While it is true that Sarbecoviruses have a wide range, direct bat ancestors to SARS-CoV-2 have a much narrower range considering bat and viral dispersion patterns. Still under investigations though, and politically sensitive topic.
My point was specifically about SARS-COV-2, which can infect mammals from several different orders. And we don't know whether the restricted bat lineages are descendent from an ancestral generalist lineage, or if SARS-CoV-2 is descendent from a more restricted lineage.
Si et al has a preprint out (zhengli shi is a co-author as well) adressing some of it by analyzing what makes RBD have broad or narrow ACE2 tropism. Seems broad tropism is ancestral
There are SARS-like coronaviruses in bats in Hubei, but they are too distant from sars-cov-2 to be the ancestor virus.
Shi Zhengli of the Wuhan Institute of Virology stated in July 2021:
“We have done bat virus surveillance in Hubei Province for many years, but have not found that bats in Wuhan or even the wider Hubei Province carry any coronaviruses that are closely related to SARS-CoV-2. I don't think the spillover from bats to humans occurred in Wuhan or in Hubei Province”. (Cohen interview)
This is confirmed by wider bat sampling efforts by the Wuhan Institute of Virology and EcoHealth Alliance prior to the pandemic. (Fan et al 2019 - WIV team; Hou et al 2010, Latinne)
The WHO Joint Study report noted:
“Tests on samples of more than 1000 bats from Hubei Province showed that none was positive for viruses related to SARS-CoV-2” (p106)
The samples tested had 74% to 90% sequence identity to SARS-CoV-2,(annex F table 4) compared to 96.1% for RatG13 (from Yunnan), and 96.8% for BANAL-52 (from Laos).(Temmam)
From January 7th to 18th 2020, Wang et al (2022) collected samples from 334 bats around Wuhan. All were negative for SARS-CoV-2 or its progenitors.
Hassanin et al (2024) reported the “the results of several field missions carried out in 2017, 2021 and 2022 across Vietnam during which 1,218 horseshoe bats were sampled”.(Hassanin) They conclude that “For SARS-CoV-2 however, phylogeographic indicators provided very high support for an origin in the zone covering northern Laos, southern Yunnan and north-western Vietnam”.
Pekar et al (2023) argue that the ancestral host for SARS-CoV-2 is ‘very unlikely’ to be from Hubei. *Pekar et al are the main proponents of the zoonosis hypothesis*
“Considering the large distance separating the closest-inferred bat virus ancestors from where SARS-CoV-1 and SARS-CoV-2 each emerged (Fig. 3i,j) and the high dispersal velocities necessary to traverse the aforementioned distance (Extended Data Fig. 18), it is very unlikely that the lineages descending from the closest-inferred bat virus ancestors of SARS-CoV-1 and SARS-CoV-2) (which reached, respectively, Guangdong Province and Hubei Province solely via dispersal of these viruses through their bat reservoirs.”
Another point - If the animals came from a farm, why were there approx 1-2 spillovers only in Wuhan city sufficient to cause an ongoing outbreak? If the farm had hundreds of raccoon dogs, why were they not numerous spillovers at other locations the farm was supplying? This is much easier to explain on the lab leak hypothesis.
What is the standard structure of exotic meat suppliers to wet markets? Are they usually big operations that supply many markets on a weekly basis, or are they usually small mom-and-pop operations that turn over their entire inventory in a single transaction to a single market? I'm sure I'm presupposing a million false things in the way I've phrased this question, but I think some of this information would be relevant to knowing how much and whether your question is a significant one to ask.
Hello, farm size varies a lot - for SARS-1 civet farms had around 100 civets per farm, but some bamboo rat farms have tens of thousands of animals. So, a lot depends on what the intermediate host was and how big the farm was.
which is cited by Worobey 2022. They say that the suppliers of the Huanan market were farms in Hubei.
They reported that reported that authorities closed 290 farms in western Hubei that accounted for between 450,000 and 780,000 animals. This suggests that there were 1,551 - 2,689 animals per farm. The number of spillovers posited by zoonosis proponents is around 8. None of these spillovers have actually been ascertained, they are the product of a (bad) model, and none of the official confirmed cases in the wet market worked at wild animal stalls.
Is there any reason to expect that the farms that were closed were relatively equal in size, rather than some being tens of thousands of animals and some being only a few dozen? Is there any reason to expect that the farms that were closed had a distribution of sizes that is representative of the set of all potentially relevant farms?
There is a lack of information on this front. The farmed live animals sold at Huanan were mainly from Enshi and "A clampdown on Enshi's wildlife trade at wet markets began on Dec. 23, 2019, according to state media, eight days before China publicly acknowledged the new virus. The head start in Enshi doesn't mean officials found something amiss: It could have been preventive, as rumors emerged of market vendors falling mysteriously ill in Wuhan. But it means evidence regarding Enshi's wildlife trade was erased before the world was aware of the existence of a novel coronavirus."
It would be very nice to know what the specific farms supplying the market were, but we don't have much information. But certainly one has to posit that the farms could not have been very large, otherwise there would have been numerous zoonotic spillovers at other markets, which were not observed
So, do you think the Wang et al (2022) and Xiao et al (2021) papers are fraudulent or mistaken? Have you thought about calling for a retraction. Wang et al literally sought out the suppliers of raccoon dogs to the Wuhan markets and tested the wild caught raccoon dogs. All negative.
They only tested for viruses, not antibodies. If you tested 100 random humans today, your chance of finding any specific endemic virus (e.g SARS-COV-2) would be quite low.
They also only state that some raccoon dogs were wild-caught, not that all were.
As noted above, they tested the entire inventory that would have been supplied to Wuhan and said they were all wild caught - they were not farmed. Xiao et al (2021) also says the raccoon dogs for sale at Huanan were wild caught. Xiao et al (2021) includes a co-author of Worobey et al (2022), the main zoonosis paper.
They tested their tissue and blood samples by PCR. Zoonosis proponents posit that the approx 20 raccoon dogs for sale at the Huanan market in the month of November 2019 caused approx 8 zoonotic spillovers, of which 2 led to an ongoing outbreak (per Pekar et al (2022). Despite testing of the suppliers to that market in January, they were all negative.
There are, to recap, three independently sufficient reasons to rule out raccoon dogs (1) wild caught, so no explanation of FCS acquisition; (2) wild caught in Hubei, so ~impossible that they were exposed to the ancestral bat virus; (3) the animals at the market suppliers were tested and tested negative.
You seem very focused on a very narrow scenario for zoonosis, that is very similar to the consensus position for the SARS-CoV-1 outbreak. If you accept any scenario for one hypothesis, but any scenario for another, of course you will end up favoring the second hypothesis. If you could prove definitively that it couldn't be raccoon dogs, that's evidence against a zoonotic origin. But it's far from devastating. There were plenty of other susceptible animals at the market. Additionally, it was a dense area with an unusual amount of traffic from people with close contact to wildlife. I think you really need evidence of engineering in order to believe in a lab leak, and I with a molecular biology background, find the evidence on that profoundly unconvincing.
I would also argue that since it's CCP policy that the outbreak didn't happen in Wuhan, we should be skeptical of statements from Chinese researchers working in China.
Xiao does not seem to declare that all raccoon dogs sold at the market were wild-caught.
>Approximately 30% of individuals from 6 mammal species inspected (labelled W in Table 1) had suffered wounds from gunshots or traps, implying illegal wild harvesting (Table 1).
>Raccoon dog fur farming is legal in China; however, due to a drop in fur prices, raccoon dogs are now frequently sold off in live animal markets, augmented by wild-caught individuals."
If raccoon dogs in any specific population were subject to any particular viral epidemic that burned itself out due to herd immunity, it's possible that no virus would be sampled two months later.
Xiao et al Table 1 marks them as wild caught on the basis of injuries and vendor interviews.
As noted above, on average, 38 raccoon dogs per month were sold across the four markets in Wuhan from 2017 to 2019. So maybe like 20 were sold at the Huanan market in November 2019, which is less than one per day. Wang et al 2022 states
"We immediately started a surveillance investigation on the
origin of SARS-CoV-2 during 7–18 January 2020. Consequently,
the lung, liver, and intestinal tissue samples were collected from
mammals, which were captured in the rural area (Changxuanling and Yaoji towns) of Wuhan by three local traders for
vendors at animal markets including Huanan Seafood market
during 7–18 January. "
I.e. these are locally captured wild animals. The 15 raccoon dogs likely comprised nearly the whole inventory of raccoon dogs that would have been supplied to the Huanan market at the time.
"If a species was wild caught in Hubei, then there would be no explanation of how it acquired the ancestral bat virus, given that Hubei is 1000 miles from southern Yunnan." How often do viruses travel a thousand miles? I was under the impression it's common.
"Even the strongest proponents of the raccoon dog hypothesis have walked back their bold claims that raccoon dogs are thost." Who are these strongest proponents?
It is true that viruses often travel more than 1000 miles, the point is that they do not do this via bats which do not migrate that far. e.g. it seems that SARS traveled ~1000 miles from Yunnan to Guangdong via the civet trade. But with SARS-CoV-2, for animals wild caught in Hubei, there is no explanation of how they would have been exposed to the ancestral bat virus.
eg Angie Rasmussen is in many ways the face of the zoonosis hypothesis and a co-author on many zoonosis favoring papers. She said to the Atlantic “This is a really strong indication that animals at the market were infected. There’s really no other explanation that makes any sense.”
Then she said this after the Bloom critique "As a co-author of the “raccoon dogs saga” report, it’s disappointing to see a journalist outright lie about our work. We never claimed to find infected animals."
> She said to the Atlantic “This is a really strong indication that animals at the market were infected. There’s really no other explanation that makes any sense.”
> Then she said this after the Bloom critique "As a co-author of the “raccoon dogs saga” report, it’s disappointing to see a journalist outright lie about our work. We never claimed to find infected animals."
I'm not sure what this is supposed to prove. There's no contradiction.
If I go into a bathroom and smell shit, there's a really strong indication that someone took a dump in there. There's really no other explanation that makes any sense. But that doesn't mean I found a turd in the toilet.
I was responding to this: "Even the strongest proponents of the raccoon dog hypothesis have walked back their bold claims that raccoon dogs are thost." Who are these strongest proponents?"
She said "there is no explanation that makes sense apart from an infected raccoon dog". She then said that this doesn't imply that a raccoon dog was infected.
No. She's saying that all signs point to an infected raccoon dog, and that they have not found an infected raccoon dog.
A detective finds a body. All signs point to a gunshot wound. He has not found a bullet or a gun. It doesn't mean he's changed his mind, or that it's now necessarily more likely that the guy lying on the ground with a big hole in his head died of strangulation.
Except that animals were not all wild caught in Hubei. This may be part of what tripped up Saar during the debate -- lab leak supporters have spent the past four years repeating things that aren't necessarily true and creating an epistemic bubble for themselves.
In reality there are many ways a virus from southern China could have made its way to a market selling wildlife in central China. For instance the markets in Wuhan seem to have sourced raccoon dogs and civets from farms in a mountainous region called Enshi about 630km to the west of Wuhan, nearly halfway to Yunnan province. Enshi has some of the largest cave systems in the world and there's no testing of bats or farmed animals from there that's been released so far. This is the region, western Hubei, where animals were found infected with SARS-CoV-1 back in the early 2000s. https://www.independent.co.uk/asia/china/covid-coronavirus-bats-caves-hubei-b1940443.html
We also know that the Huanan Market also imported at least some animals such as bamboo rats directly from southern Yunnan province, as well as Guangxi. And these are only the sources we know about...it's unclear exactly where the raccoon dogs sold at the time of the outbreak came from.
Meanwhile the evidence at the market does seem to point to raccoon dogs. The store where they were sold appears to be the most likely source of the outbreak within the market. Positive samples were found on a cage used for holding them, as well as a tool used for stripping fur off the animals. That doesn't prove they were the species that spread the virus to humans but there's definitely not a good basis to rule it out.
As noted above, there is consensus in the literature that the ancestral bat virus is not in Hubei. So, you are going against all of the scientific literature here, including Pekar et al (2023), who are leading proponents of zoonosis.
Worobey et al cite the washington post article which says that:
“A person with knowledge of the Wuhan market supply chains, who spoke on the condition of anonymity to protect his contacts, told The Post that live animals sold at markets in Wuhan were sourced from Hubei, particularly Enshi and Xianning prefectures, as well as from Hunan and Jiangxi provinces.”
Further confirmation for this is provided by the fact that:
“A clampdown on Enshi's wildlife trade at wet markets began on Dec. 23, 2019, according to state media, eight days before China publicly acknowledged the new virus. The head start in Enshi doesn't mean officials found something amiss: It could have been preventive, as rumors emerged of market vendors falling mysteriously ill in Wuhan. But it means evidence regarding Enshi's wildlife trade was erased before the world was aware of the existence of a novel coronavirus.”
I don't understand why you think the raccoon dogs were not wild caught locally. Xiao et al Table 1 says this explicitly, as does Wang et al 2022
It seems to be saying that many animals were from farms in Enshi where civets and raccoon dogs were raised. That's a significant distance from Wuhan and yet you're pointing to testing of the 15 raccoon dogs from within the Wuhan city limits itself.
As far as the ancestral virus not being from bats in Hubei that may be true but like I mentioned SARS-CoV-1 also somehow infected animals in western Hubei. So there are certainly some unexplained details but I don't think any of this makes a spillover to humans from raccoon dogs or civets impossible at all. Remember the zoonotic origin doesn't need to be something as simple as it spilling over from bats to one species of animal that then infects humans at the market. It could have spread between multiple species just as COVID did after infecting humans.
Also the reported shutdown of these markets in western Hubei a week before the Huanan Market was closed is interesting and that alone would seem to call the lab leak explanation into question.
The 15 raccoon dogs were traced from suppliers of raccoon dogs in the Huanan market. It's not that they randomly sampled raccoon dogs in Wuhan. They did this over 11 days. Around 38 were sold on average per month. so, the raccoon dogs they tested likely accounted for the entire raccoon dog inventory in Wuhan at the time.
The wapo article is about the farms supplying the market, but Xiao et al says that not all of the animals sold at the market were farmed. Worobey et al also says this!
fwiw I think the infected animals for SARS in western Hubei is disputed though haven't looked into it.
Shutting down of the farms is not evidence against lab leak. Some parts of the government clearly thought it was a zoonosis and acted accordingly in Dec and Jan, which is eg why they disinfected the market in Jan.
The Forestry Bureau responded with a list of sources for the store -- which included raccoon dogs sourced from a farm where mink and foxes were also bred. So in at least some cases, they were sourced from farms.
2) This study doesn't say that all raccoon dogs were wild-caught and testing 15 animals from one location is a ridiculous number to draw a conclusion from.
Overall the wildlife trade in China is difficult to track with animals being raised on farms but also poached from the wild, often shipped long distances, with some such as pangolins trafficked illegally. We never got a full accounting of animal sales at the Huanan Market -- in spite of the post above by the Chinese NGO, this information was left out of the eventual WHO report. There was no mention of the sources for raccoon dogs and civets, two of the most-suspected species, and China denied there was any evidence live mammals had been sold at the market at all anytime in 2019.
The article you link to refers to an advertising sheet from 2015. Xiao et al (2021) is a study from 2017-2019 involving observation of the animals and interviews with sellers.
15 is not a ridiculous number to draw a conclusion from. If the monthly sales were 38 per month across four markets in Wuhan, then the January testing by Wang et al (2022) of the supplier is half of the inventory that would be supplied to that specific market in a month.
Worobey et al said of the raccoon dogs photographed in Dec 2019 "“appear to be local, wild-caught common raccoon dogs rather than farmed raccoon dogs and that their plush coats are consistent with those observed in the winter”. edit - reference is the worobey preprint - https://zenodo.org/records/6299600 fig 3
No the article I linked involves an environmental NGO asking the Wuhan Forestry Bureau about that advertising sheet, which lists nearly a hundred species for sale, far exceeding the total number of species you're allowed to sell in China.
The Wuhan Forestry Bureau responds by giving a list of sources for that store at the market, and it includes raccoon dogs from a specific farm. This is the ministry that was responsible for regulating the animal trade in Wuhan.
Meanwhile the Xiao study you mentioned simply indicates that some of the animals observed appear to be wild-caught. So this is a good example of why it's important not to draw such a strong conclusion when there are all these different facts swirling around. Let's at least wait until China gives a source for the raccoon dogs that were sold on the western side of the market -- they've never provided this information amazingly.
Still, the referenced farm is in the east of China, which is even further from the relevant bats than Hubei. In general, fur farms in China are overwhelmingly in the east of China.
Yeah I agree but this particular store, located on the east side of the market, was already less likely to be the origin. The question is why this information, sources for the most suspected species, was left out of the WHO report. There are no sources listed for raccoon dogs or civets at all, in spite of them clearly having been sold there and in spite of China already having both sampling data and seemingly also sales records showing they were.
WHO investigators later said CCP officials pushed them to exclude any mention of these animals from the 2021 WHO report and denied there was any evidence of any live mammal sales at the market during 2019. So China seems to have censored almost all of the information (beyond the 15 locally tested animals you mentioned) about the most likely intermediary species.
Roko (the one infamous for the basilisk) has argued that what might have appeared as a shift in expert consensus is in fact "manufactured consensus." This is worth considering. Link: https://www.lesswrong.com/posts/bMxhrrkJdEormCcLt/brute-force-manufactured-consensus-is-hiding-the-crime-of | I also believe an algorithm somehow sorting, weighting evidence cleverly could solve the problem of "too much evidence." The CIA, in fact, at least has used one such algorithm, invented by Richard Heuer.
I think the fact that the expert class went out of their way to suppress the lab leak hypothesis, and that lab leaks of this nature have occurred in the past, is probably damning enough. My impression is that Roko hasn't really introduced any new arguments that Peter didn't cover in the debate, and Peter isn't being funded by the CDC, to my knowledge.
Everybody plays "Cover Your Ass" whether they're innocent or guilty. So that's not a viable argument. Nobody likes to be accused of having made a mistake, whether they have or not.
No, it's supposed to say that suppressing a hypothesis is not damning evidence of the hypothesis being correct. The hypothesis might be correct or incorrect, suppressing it is bad anyway.
I regret to inform you that technocracy is unusually *good* on this point. If you can reliably find a strain of human that doesn't reflexively hide their misdeeds, we sure could use them!
Everyone? Some naive people talk to cops believing they will get their side of the story out there
Destroying evidence, like oh idk dna sequences the wuhan insitute were working on at the time; suggests a estimation that they expected they couldve been at fault, I would expect innocent people to hide the evidence and do a very private ananlysys and ask a lawyer if it helps them.
>So [playing "Cover Your Ass"] is not a viable argument.
It *is* a viable argument in the sense that it dampens and modulates the strength of other arguments.
Arguments about what the WIV can or cannot do, arguments about the number of cases centered around the wet market, and arguments about which DNA sequences were found and when they were found, all need to have their Bayesian impacts divided by a selection-bias-from-a-Maoist-Totalitarian-Government factor.
OK, but you've got to include EVERYONE, not just the people you don't like. It wasn't *just* the various segments of the government of China, it's also the investigators, people who have a public positions ahead of time (and "public" can be a argument with another expert).
Just about everyone who made a public statement on any factor related to COVID during the early days had at least one mistake that they wanted to "de-emphasize". There were "public experts" in the US claiming that we could keep COVID out after it had already been circulating for months. And the border closings were a farce. The quarantines were not really enforced, and lots of people were allowed to essentially skip them. (Not that it really mattered since COVID was already circulating.)
The expert class went out of their way to suppress the lab leak hypothesis in part because Trumpist Republicans were promoting the lab leak hypothesis and the expert class is reflexively opposed to anything associated with Donald Trump. And in part because discussion of the lab leak hypothesis, regardless of its factual basis, would imperil the funding and prestige of themselves and their colleagues.
Neither of these depend on the lab leak hypothesis being true or probable, so the fact
of the suppression carries little information regarding the truth of that hypothesis.
More generally, innocent people reflexively engage in coverups often enough that "Coverup, thus Guilty!" is a weak argument at best. The bit where innocent people welcome thorough investigation of every accusation made against them because they are confident it will prove their innocence, is sadly disconnected from reality.
Also important to realize that the idea of lab origin was pushed by a number of different groups, not just Trump Republicans. The first version of it was actually from the CCP themselves, who tried to claim SARS2 was a bioweapon the USA attacked them with at the Wuhan Military Games. From there it was picked up by anti-CCP groups in the Chinese diaspora, which had a surprising degree of influence on certain Trump officials, and the storyline changed to focus on the Wuhan Institute of Virology. Other governments such as Iran and Russia came up with their own versions blaming the USA.
So there's really an ocean of BS on this topic and you can't blame scientists for saying it's less likely when it actually does appear to be less likely. None of this resulted in the idea itself being suppressed. It's probably gotten something like a hundred times or more coverage in the media as the animal trade origin. Governments and other groups are incentivized to promote some version of this storyline for geopolitical reasons -- in China's case to evade responsibility for the pandemic, in the USA's case to more directly blame China. Meanwhile the real origin gets lost in the confusing back and forth of competing claims and accusations.
"It's probably gotten something like a hundred times or more coverage in the media as the animal trade origin"
Really? You want to dig into the archives of NYT and show me a hundred-to-one ratio of articles propounding the lab leak theory vs. zoonosis?
I also think 'Biorealism' made a fair point at the end of the comments section-
"There is at least one argument they used knowing it was misleading. They used Ron Fouchier's argument (without acknowledgement) that WIV would have used a well known reverse genetics system despite Andersen saying that wasn't the case on at least two occasions. He noted they had been created "on a whim" and on 20 Feb if anyone thought these were hard to create from scratch a group of researchers had just created one in a week. They should have acknowledged this.
The Nature reviewer clearly took too hard a line as well. Their initial manuscript was far more balanced. Moving from lab origin was not "necessary" to "not plausible" was not justified. As David Relman emailed Francis Collins there need to be an impartial assessment which acknowledged the lack of evidence either way. A Dept of Defense Working Paper dated 26 2020 noted their conclusions were based on "not on scientific analysis but, on unwarranted assumptions"."
Just doing a quick search on Google, even NYT, which has been criticized by lab leak supporters for their supposed biased coverage, has about twice as many stories that mention "Wuhan Institute of Virology" vs "Huanan Market." And overall in the media the number of articles on the animal trade origin -- i.e. looking into the animal trade in China in relation to the outbreak -- is almost nothing vs. those covering the lab leak controversy.
This is strange because even if you think it leaked from a lab there are specific people in China, vendors and workers at the market, who could tell us much more about what happened, and yet the focus has shifted almost entirely to American scientists and politicians. You then have to think about the fact that many experts believe the lab leak never happened at all!
As far as the Proximal Origin paper I think it's fine to say the conclusion should have been less strong, maybe they should have said "lab origin is less likely" rather than "not plausible." But what's funny is you can also find a quote from Richard Ebright from the same time period (but before Proximal Origin was published) where he says "Based on the virus genome and properties there is no indication whatsoever that it was an engineered virus."
So clearly this was not an unreasonable conclusion to come to, unless you are saying Ebright has no idea what he's talking about, and where it falls apart IMO is this idea that in writing this paper scientists were somehow suppressing the lab leak theory. Just weeks after the paper was published the president of the USA was repeating stories about how someone at the lab had started the pandemic. The State Department launched an investigation that resulted in the stories about "sick WIV scientists with COVID symptoms." Steve Bannon ran a PR campaign to promote Li-Meng Yan's bioweapon allegations. Congress issued multiple reports and held hearings about it. We're now going to see yet another investigation by Rand Paul, who doesn't exactly seem neutral. You can't say this idea was really suppressed that much...instead there's a narrative people repeat over and over about how it was suppressed.
"Just doing a quick search on Google, even NYT, which has been criticized by lab leak supporters for their supposed biased coverage, has about twice as many stories that mention "Wuhan Institute of Virology" vs "Huanan Market.""
If you use those exact terms, yes, you get 450 results vs. 170 in favour of the Wuhan lab being mentioned. If you search for "wet market" vs. "lab leak", you get ~40,000 results for the former vs. ~16,000 for the latter, so the balance of coverage flips. If you search for the same terms on google outside of the NYT website, you get about 3 times more mentions of the Wuhan lab or 'lab leak' than you do for the wet market, but on the other hand, "chinese animal trade" returns 89 million results as opposed to 274,000 for "chinese biolab". So I'm not sure the zoonotic origin story is being downplayed, precisely.
My argument is not "cover-up, therefore lab leak is true". My argument is "cover-up, therefore untrustworthy"- that the effort to suppress the lab leak hypothesis was a crime in itself, and the expert class are still 'guilty' in that sense.
My understanding was that the Proximal Origins paper was already published by the time Trump made a public statement in early April, but I don't have a ref handy.
Yeah that's correct, although it had been promoted earlier by other GOP politicians such as Tom Cotton, and outside the government by former Trump officials like Steve Bannon. Actually I don't think Trump himself was ever too invested in this idea. There's a hilarious interview he did with Sherri Markson where it quickly becomes clear he doesn't have much more knowledge than a random person. It was more a project of Mike Pompeo, the Secretary of State and former CIA director, and a group of people working under him
Ultimately the Proximal Origins controversy becomes much less interesting IMO if there was no actual inside knowledge about what happened, and even Emily Kopp of US Right to Know told me she doesn't think that's the case, although she does think scientists engaged in a "coverup" of some kind. Nothing ever prevented anyone from writing a different paper saying lab origin is plausible and many did.
Hence my specifying "Trumpist Republicans" rather than "Donald Trump". As everythingism notes, Trump himself was not really the problem, but there were plenty of people in Trump's orbit advancing the Lab Leak theory.
And advancing it well beyond what the evidence would have supported at that early date. It was absolutely reasonable for the "expert class" to push back against that, and at least understandable that they would overcorrect in the other direction.
This is one of the situations where I don't feel any more confident trusting the expert consensus than looking at the papers and/or debates and trusting my own reasoning to evaluate the arguments.
A good example I personally like to use of this definitely being the case would be from COVID as well: during early 2020, the Finnish Institute and Wealth and Welfare (THL) officially claimed IFR of [alpha-variant, in immunonaive population, before best practices for treatment were discovered, as average within a population with a lot of elderly people] COVID to be <<.01%, purportedly repesenting the best available scientific evidence, while at the same time some .5% of entire population of Bergamo in Italy had died (and similar IFR could have been calculated from Diamond Princess data, etc). Here, rational evidence easily trumps claims made based on scientific evidence: a child could see if explained to that no matter how many unidentified cases there might be, there's no way for IFR to be lower than total excess mortality within the population. And additionally, I have a clear model where the people working in the institute with their virology doctorate doctorate degrees went wrong: they rely (indeed, might be contractually obligated to, I haven't checked) on scientific studies, but as none had been published at such an early stage, they pattern-match to the most closely analogous study, which might be about flu or something. Most people, even experts, just aren't epistemologically nuanced enough to use rational evidence, as Scott has often written about in "The Phrase "No Evidence" Is A Red Flag For Bad Science Communication".
When expert scientists are discussing statements within established science, I would trust their claims to not misrepresent the science, and since the scientific method approximates Bayes' theorem pretty well, I have a high degree of trust for the claim. This isn't such a case! Science works, but it advances one funeral at a time more or less, and there's no way this debate has gotten to a point of established science, indeed the polls lean to one direction but also specifically show the question ISN'T settled.
Sure there can be. Not all deaths are caused by COVID, and both Diamond Princess and the population of Bergamo were heavily age-skewed. Not a representative population, that's why serosurveys were useful to determine the true number. You say the Finnish government were also making claims about the elderly but did you check that the numbers line up there?
There are other issues with early IFR estimates, for example in Italy there was mass abandonment of care home residents by the mostly foreign carers who were scared and often wanted to return home in anticipation of border closures. Deaths reported as COVID were actually in some cases caused by abandonment e.g. dehydration. Early on doctors were also way too ready to put people on ventilators, partly because they thought that would keep the virus out of the hospitals and themselves (it didn't). But ventilators are a very destructive last resort option. Once doctors stopped doing that so much the death rates went down.
The point made in the link is classic the-social-aspect-of-opinion-formation. The world of domain experts can be analysed as a community (more-or-less closely integrated), complete with opinion leaders, followers, rebels, go-betweens, outcasts and every other (social) role related to the goings-on in any community.
The negative spin on these (social) processes in the linked post is unneccessary, though. Partly because such social processes are inevitable/unavoidable, partly because they may also be conductive in getting the joint reseach process from A to B. (The importance of all social factors on opinion formation is Janus-faced.)
(...and ACX/self-declared rationalists is also to a certain extent a community, of course.)
I have no horse in this particular race, but I do have a lot of expertise in some of the areas rootclaims "investigates" (especially the stuff related to Syria and chemical weapons) - where their analysis is so shoddy and laughable it's indistinguishable from Youtube conspiracies - and the biggest surprise to me here is that anybody really bothers with rootclaim in the first place? The more you learn...
Even if you do not agree with the values of the (prior) probabilities, you can choose your own, and replicate the process rootclaim uses. It is far from "shoddy" and research on Bayesian methods modelling these kinds of events is highly valuable.
I have nothing against Bayesian reasoning. Use it all the time. But the priors, the "evidence", the weights attached were so out-there, I just assumed arguing with them was like arguing with Alex Jones who has taken a stats course. Unfortunately statistical methods don't cure brain worms (and in fact might even provide them with a rationalist veneer).
Can we then Aumann agree that: the Rootclaim method with better priors and handling of evidence would not necessarily be shoddy, the current analysis is or might be, and the method itself remains valuable enough that it would be good if they continued to work on it (in Scott's framework, (2)).
With good-enough priors, yes it would be good enough (by definition). But how good would those have to be? Is it possible to have priors good enough that the method isn't just laundering assumptions through some fancy statistics, as meta-analyses or financial projections or suchwhat often are? I'd answer these questions as "at least 10x better than we have" and "not really, no."
You happened to laugh at an analysis where we turned out to be spot on. There is no now clear verified video evidence of the opposition carrying out the Ghouta sarin attack.
If you are going to claim your analysis is spot on, please link to a credible independent source. Otherwise this comes across as we believe this unlikely thing and used our analysis you find shoddy to conclude we were right so you should not consider our analysis shoddy.
I did follow the link. My point is if you were vindicated I would expect to be able to read about it in an independent source. Otherwise I need to look at all of your sources and see both which are in support of your claim and which you are claiming support you and look at the evidence in detail. Lab leak as a plausible theory has been printed in the NYTimes since at least 2021, has any mainstream newspaper published anything about this?
How often do mainstream newspapers publish anything about Ukrainian misdeeds in the war? Probably something like 1/1000 ratio compared to the Russian ones, whereas a reasonable prior of their prevalence would be like 1/10 at best. Similar sort of bias is probably at play with Syria.
I'm not discussing the ratio of coverage but here is an article from le monde that is easily searchable that backs up the claim that Ukranian's committed war crimes. https://www.lemonde.fr/en/international/article/2022/04/09/ukraine-s-military-accused-of-war-crimes-against-russian-troop_5980121_4.html. If there was anything that resembled reasonable evidence, it would get some coverage the trouble is Saar has motivated reasoning that is not convincing without external collaboration. Otherwise the best you can claim is that there is some evidence that the syrian weapons were a false flag and again it's not your (Saar's) method that proves the analysis is spot on, its how strong you trust the conclusions from the grayzone/wikileaks that the appearance of impopriety means the conclusions are false.
Sorry, I thought it was obvious. These outlets have no incentive to publish such findings. Sadly these are the kinds of things you need to verify yourself.
How do you account for “one of the most powerful political actors in the world which has a history of controlling information flows runs the area in question and can be manipulating what evidence is available?”
I get that you have to work with the evidence you have. When the evidence you have is passing through a filter that has pretty strong incentive to hide its own wrongdoing (if that were to exist) how does that fit into your analysis?
Did they assess a probability that, ie, certain research records where the researchers tried that ~exact~ furin site were destroyed by the CCP? Wouldn’t that just be a “naked prior”?
The CCP were trying to cover up zoonotic origin as well at first, with arresting Dr Li and cleaning up the site. Because it was seen as politically embarrassing, as China said it had fixed wet market issues after SARS. So that applies on both sides.
(The Chinese official position now is to heavily imply it came from abroad)
Also a question of whether, in the lab leak hypothesis, they'd have been able to cover up large amounts of evidence effectively, given how inept the initial covid response was. And officials at different levels acting at cross purposes to cover their asses
This is not true. George Gao said in January 2020 that it was from the wet market. After leading sampling there and the case search, he changed his mind in summer 2020
If you look at what usually happens with research-related accidents, they are usually shrouded by obscurity for years. This is likely even more true in China. Compare e.g. the 1977 Russian flu, sverdlovsk anthrax leak, various SARS leaks. In some cases finding out that an accident occurred can take decades.
I think you're expecting a government to act in a unified way. I'm sure if you searched you could find statements supporting every side. And I *have* encountered (reports of) official Chinese statements claiming it came from abroad. So that is one of the positions that they are (or were) taking. I've also encountered denials that COVID exists (e.g. by the mayor of Wuhan), though that was early in the epidemic.
I was responding to the claim that the government was trying to cover up the wet market origin in early 2020. This is inconsistent with the head of the China CDC saying it came from the wet market in January 2020.
Obviously, all of their incentives are to cover up a potential lab accident.
The problem is that everybody has incentives to cover up, whether or not they are at fault. If someone things something MIGHT be read as placing the blame on them, they'll try to suppress it. This makes things quite murky.
FWIW, I believe that if the disease originated (as a human disease) in, say, New York, but didn't get the last mutation that it needed to become pandemic until someone in Wuhan caught it, we'd see largely the same response. You can't judge things by "who's trying to cover things up".
But you have to ask why George Gao would change his view to denying the wet market origin if it came from the lab. That's a strange thing to do if China's intention was to cover up a lab leak. What these statements from China and the various lab narratives/hypotheses have ultimately done is leave the origin unresolved within the domain of global public opinion, an outcome that may not be the worst from China's POV. They're able to avoid officially taking the blame for the pandemic and can control the narrative domestically to say the virus came from outside China.
That would only work if the research was completed but not published before COVID happened. If someone was studying furin cleavage sites in 2018, they would want to publish a paper saying "hey, we tried cleavage sites X, Y, and Z and only X worked" and China would have no reason to think that should be kept secret because COVID hasn't happened yet.
Even a 3 minute skim of anyones work here suggests they probably believe "math is so great you get the best answer anyway", and something something wave at criminal forensics existing
This is discussed in the debates and in Scott's summary. It is a common theme with conspiracy theories (term used literally) to allege that the conspiracy is both capable of and motivated to make the evidence come out to whatever happens to be the case. However, just because you can imagine a hypothetical in which this might happen, doesn't mean it's remotely likely, especially when characteristics like who is involved and how competent they are change from argument to argument. China is not omnipotent and frankly probably doesn't care much about lab leak vs zoonosis. Let's consider the epidemological data pointing to the market, which is often alleged to be tainted.
1. The Chinese government first tried to cover up the existence of the pandemic at all, and failed. Also, as Peter points out, if it started much earlier, you would have many times more deaths, which is even harder to cover up.
2. They also tried to cover up (or at least, were incompetent) the wet market, by claiming animal testing showed no results, but then it turned out they hadn't actually tested live animals (roughly, I might be getting details wrong here).
3. Now they claim the virus started in the US; if they were trying to fake evidence of this, they would have had cases cluster near an airport or hotel or something, not a market that is almost exclusively visited by residents.
4. There's multiple different lines of evidence that point to the market, which makes it increasingly unlikely that China could manipulate all of them in a way that's undetectable.
5. If you don't believe the market data, why do you believe the pandemic started in Wuhan at all? Couldn't China be trying to hide, like a secret bioweapon plant in a different city? This is possible, but of course there's no evidence for it.
Maybe you can help me understand how this would fit into a Bayesian framework, but here’s how I’d think of it:
Suppose someone did indeed to a research and create that exact strain of virus that was ostensibly “not what you’d do” if you were doing research. This would be pretty direct evidence overwhelmingly in favor of the lab, right?
Yet if that evidence existed it would be far easier to track down and destroy, precisely because someone would know “holy crap I did this” and that specific person would have a very strong incentive for that evidence not to go public.
I don’t believe that “China make the evidence go however it wants”, I just don’t see how to account for “people who knew they caused this could destroy that evidence far more easily than we could find it, and would have a very strong incentive to do so.”
> Suppose someone did indeed to a research and create that exact strain of virus that was ostensibly “not what you’d do” if you were doing research. This would be pretty direct evidence overwhelmingly in favor of the lab, right?
I don't follow what you're saying here. It looks like "Suppose it was created in a lab, that would be evidence in favor of the lab." Can you rephrase? Is there a statement missing here?
> Yet if that evidence existed it would be far easier to track down and destroy, precisely because someone would know “holy crap I did this” and that specific person would have a very strong incentive for that evidence not to go public.
I agree that *if* some lab workers knew they had created covid, and knew how they had done it, they probably would and probably could destroy at least some evidence. It's not at all clear to me how well this would actually work (i.e. would any evidence remain? would there be evidence of this sort of cover up?) Do you have more specifics?
(Even in the case of lab leak, there might not have been any such cover up--I have seen the theory that the leak was totally accidental and came directly from samples, so they might not have even realized at first even if they were to blame, and WIV workers were behaving pretty normally through December at least. Conversely, even in the case of zoonosis, panicked workers could have engaged in some sort of cover up if they had some vague idea they might be considered responsible. So FWIW I don't think that either "there's some evidence of a cover up at the lab" or "there's no evidence of such a cover up" would be definitive evidence either way.)
> I just don’t see how to account for “people who knew they caused this could destroy that evidence far more easily than we could find it, and would have a very strong incentive to do so.”
Criminals try to hide evidence all the time, but still get convicted (or convinced to take plea deals). It's actually really hard to successfully hide, destroy, or fake all of the evidence for something this big, and also do the same for the cover-up, and for that cover-up, etc. There were whistle blowers for previous outbreaks, including lab leaks of Sars 1, as well as for Sars 2, from the hospitals, in spite of Chinese crackdown.
You can certainly put a cap on how confident you are in a theory--it's not like the judges, or Miller, or Scott, or me, or the superforecasters, are >99.9% confident in zoonosis. Seems like mostly in the 80-95% range. But it's not correct to just throw all the evidence out.
I agree with most (all?) of what you’re saying but still don’t understand something. I couldn’t glean this from the article.
How did “people have an incentive and ability to destroy evidence” influence the computed odds here?
Yes i understand that destroying evidence would likely happen either way (due to panic, not wanting to be blamed, not knowing the true cause) and that it can be hard to totally cover something up and that people often are convincted despite trying to cover up evidence, etc.
I’m not questioning the outcome of the debate.
I’m just trying to understand how “people had an incentive to destroy evidence” influenced the computed odds. It seems like the conclusion is “that doesn’t matter because there are multiple trails of evidence that point to the zoonotic case”, but I can’t tell if this is “it shifts odds by 5%” or “we just didn’t compute that because it doesn’t seem necessary” or what.
> How did “people have an incentive and ability to destroy evidence” influence the computed odds here?
Miller spends a fair amount of time in the debates on why the early case data is reliable enough, in ways that can't easily be faked, as well as reasons to believe that China is neither capable of nor willing to fake the data in a way that explains the evidence. So I think the conclusion is "that's theoretically true but it seems unlikely to explain anything."
I think the highest level answer is "everyone rounded the BF used for each piece of evidence down, based on how likely they thought the evidence to be wrong." If you want the specific numbers the judges used, you would have to dig into their reports, or perhaps ask them directly. As far as I can recall, I don't think Rootclaim really dove into this point, but rather just used it at a very high level ("those data could be wrong, so the evidence provided is weak") so the judges may have just taken it has most likely that the data are accurate.
Keep in mind that the data merely being limited (which is known to be the case) doesn't necessarily tell you much, because you have to explain why the cases cluster at the *market* specifically. If want to claim the pandemic started at the lab, it's not enough to say that maybe there are cases missing, because it would still be a hell of a coincidence that the cases we do know about are all at the one place in the city most likely to be the origin of a zoonotic pandemic.
> 1. The Chinese government first tried to cover up the existence of the pandemic...
It's a bit more nuanced than that. When people started coming down with mysterious cases of a pneumonia-like disease during the first weeks of December, the local hospital MDs and administrators alerted the health authorities of Wuhan, and they in turn notified the Mayor of Wuhan and the Governor of Hubei. The mayor and the governor tried to hush it up because a big party convention was scheduled to be held in Wuhan (early in Jan IIRC). The mayor and governor of Hubei Province decided to go forward with the conference. They felt a small outbreak of some unknown pathogen would reflect poorly on their administration, and they didn't alert the national authorities to the outbreak—until people were crowding the emergency wards, and they couldn't cover up all the hospitalized patients. (This would have been around the third week of December when they couldn't cover it up anymore—although it didn't reach the Western media until the first week of January.)
Local dithering and silence wasted two critical weeks, but some of the hospital MDs notified the folks across the river at the WIV, and they were able to alert the National Health Commission of the PRC that they suspected they had a SARS-like outbreak on their hands. Not long after the MDs were beaten and arrested by local Gong An Bu, probably at the behest of the local authorities who were embarrassed about the outbreak happening on their watch (again, it's unclear who directed them to do this, but the Mayor and the Governor had the motive).
The national health authorities took the warning seriously, and they informed Chairman Xi that the situation was potentially quite serious. He immediately fired the Mayor (and then the Governor) for incompetence and he sent in a hand-picked troubleshooter to handle the situation. (I wasn't able to find out what happened to the mayor and governor. I don't know if they were allowed to fade into private life or were maximally disciplined.) Once the National-level health authorities assessed the situation (and it took them about a week), they alerted the WHO at the end of December, and the Chinese started sharing data with the WHO, the ECDC and the CDC. The local coverup wasted two critical weeks, but it was at the local level, and not at the national level.
> 2. They also tried to cover up (or at least, were incompetent) the wet market...
Due to the case cluster, the Chinese national health authorities immediately focused on the wet markets in that district. SARS1 and some Avian Flu outbreaks had started at wet markets, so the wet markets within the initial case cluster were immediately suspect. And once it was clear that they had something dangerous on their hands, public health authorities immediately shut down the wet markets in Wuhan (I believe they shut them all down all over the city). The animals were disposed of, and the market stalls were disinfected. From a public health perspective, this was a wise decision—but from a data-gathering perspective, they lost the chance to test the animals when they incinerated the carcasses. Again, I may be misremembering some of the details but they invited the WHO to visit the site in mid-January. Either the Chinese or the Chinese with the WHO collected swabs of DNA and RNA from the stalls. So the disinfection happened after they collected the DNA and RNA. The Chinese sequenced the virus and they sent the genome to the WHO, to ECDC, and the CDC on January 15th. It wasn't until March when the CIA director and Trump started accusing the Chinese of covering up a lab leak that the Chinese stopped cooperating with us. Then naturally, they blamed us for creating the virus.
It's important to note that wild animal farming is a 7 billion dollar business in Hubei, and Wuhan is the second only to Guangzhou as a transshipment center for the animals and their byproducts. It's a big business, and the authorities don't want to see it shut down because it lines their own pockets. So their impulse is to cover up instead of disclose because of all of the negative attention they've garnered worldwide.
What's important to realize here is that the CCP doesn't want evidence to exist for *either* of these theories. In both scenarios, the CCP would be partially responsible, which is obviously bad for their optics. Ultimately, COVID didn't hurt China relative to other countries, and so they don't really have any incentive to prevent this from happening again. Any evidence that would narrow down the truth, any truth that would implicate China, would have been destroyed already.
I enjoyed this piece - I tried to follow this story in 2020 and have always leaned more zoonosis than lab leak (I believe I said 90-10 publicly in 2021, probably closer to 95-5 now).
Despite this, I've often been frustrated and annoyed by mainstream virologists who dismiss LL with poor reasoning, and almost always find myself defending (more reasonable) LL proponents.
I guess that this annoyance has led lots of people in rationalist-adjacent circles to lean further LL than they should have. I wonder if this is a common rationalist failure mode we need to update on.
Without caring about which one is actually true, we can look at the costs of the beliefs of they were true or false. If LL is true, and you fight against some sort of conspiracy, that is some positive U. If LL is false and you fight against an imaginary conspiracy, that will worst case cause skepticism and more investigation. Believing that it's zoonosis means you will be enforcing a conspiracy if you are wrong, which is negative U. Otherwise slightly positive U to most people, but less than what you'd get if you were right and fighting a conspiracy. Eliezer says one should believe that which is true, but actually this sort of analysis of costs and gains of being wrong / right is very useful, because it can help to navigate the unvertainty of what you should believe in.
I did not take signaling that precisely into account but this is absolutely true and might have some wider implications (I did not have solid concrete examples at time of writing).
It certainly is an important point. Every statement you make about the probability of something is at the same time a signal you send about what kind of person you want to be perceived as, and/or want to perceive yourself as.
Unless you go all Pascal's wager (which is obviously bullshit), your beliefs don't have immediate consequences in the world.
In the brains-as-Bayes-engines model, beliefs are probabilistic in nature.
What you do about your beliefs is then another question subject to cost/benefit analysis regarding your utility function. You might want to state your true beliefs, exaggerate them, stay quiet, pay lip service to the opposite side or even do your very best to argue against your own beliefs.
There are some corner cases where you might chose to stay strategically ignorant due to direct effects of your beliefs (e.g. infohazards, unconditionally knowing about your top 100 health abnormalities might encourage hypochondria in all the cases where you don't have serious health problems, or psychiatric disorders being at least somewhat memetically infective in populations), but for the most part the Litany of Tarski holds: if X is the case, then you want to believe X.
I think Pascal's Wager is true. Probabilistically speaking, life is meaningful and consequential enough to treat it as such
If I had to give a number on a test I might say that our life is 60% likely to be materialistically determined. To strengthen the point, we'll go with 99%
Now multiply the likelihood by the consequences. Again, to strengthen the point, let's say it's probably not very consequential to both exist in a 'meaningful' and non- random universe and correctly believe that it is. So 99% random or probabilistic existence x 0 consequences of being correct or incorrect about that. And 1% non-random x .1 consequences of it being true. Point one units of consequence weighed against 0 seems very noticeable in an absolute sense
With my own weights of likelihood and suspicion that human beliefs and actions can be wave like, propagating and amplifying despite small discrete effects, then my take is that I am compelled to act as if my life and beliefs are potentially consequential and subject to free will
It's a little weird to cite Eliezer as the person who says one should believe that which is true! This really seems like something that should be much more fundamental than one guy who's famous for an unrelated issue (AI safety).
He wrote less wrong and also a screed on what "the simple truth" is, as well as "Harry Potter and the Methods of Rationality", not "Harry Potter and the sealing of horcruxes" or similar.
I don't think you can say he has *no* claim on being the truth guy, or he isn't famous for wanting to believe true things.
Thank goodness he came along in the early 21st century and upended thousands of years of philosophic practice by inventing the idea that people should believe true things to be true!
I mean, that might not be a novel idea, but it's not really universal either. You'll find plenty of people insisting that sometimes believing or teaching false things -- even tricking yourself into believing things you don't know to be true -- is a good idea. (Plato and Pascal being two famous examples.)
To be the Truth Guy, he doesn't have to say it, he has to be the first or the best in some way. A lot of the problem is that EY's fans tend to be naive about the history of thought, so it all seems new to them. It's not even the first iteration of out-of-the-mainsteam philosophy that claims to answer everything.
The thing called "rationalism" has little to do with rationalism as the term is understood in mainstream philosophy. It is better understood as part of a pattern where a succession of small, insular groups engage in a very self confident style of amateur philosophy. Unlike other cult-like groups, they claim to have a strong orientation to science, logic and reason(and to be able to implement them much better everyone else).
Examples include Yudkowsky's "rationalism", Critical Rationalism (Karl Popper's philosophy, articularly David Deutsch's version), Ayn Rand's Objectivixm and Korzybski's General Semantics. There are often several such movements overlapping each other, some are reactions to others, and some borrow ideas from others.
In more detail, This kind of Thing is characterised by:-
• Being science-orientated , but having much more specific claims than "science good"
* Being strongly to stridently irreligious and anti religious. Also opposed to New Age thinking and continental pilosophy.
• Being largely outside of mainstream academia etc
• Being an insular group that mostly talk to each other
• Having difficulty in communicating with outsiders , in any case, because their own theories are expressed in a novel jargon.
• Centering on a charismatic leader, with a set of mandatory writings
• Having an immodest epistemology..which claims to be able to solve just about any problem..
• ..which is based on a small number of Weird Tricks. Nothing requires the same level of deep study as mainstream academia, and the key ideas are often brief slogans.
> To be the Truth Guy, he doesn't have to say it, he has to be the first or the best in some way.
I used lower case truth guy, clearly in reference to him becoming well known in some respect for this. I use it in the same way that I would call a somewhat competent local mechanic the "car guy" or I can call Bill Nye "the science guy" without claiming he's the best person in the world at science, or that he invented science.
Or that I was replying to the statement
> It's a little weird to cite Eliezer as the person who says one should believe that which is true!
Which, like, do you believe that you should only cite the things from the best formulation or inventor? Maybe start auditing Kenny about what other dastardly crimes against citation he has committed would be better.
I feel it's not a coincidence that this type of misreading is common to, as one would say, "haters",whom all seem to share the following traits, excuse me they are Characterized by:
> Perceive any group seeking improvement to be automatically suspect
>Incapable of reading even slightly positive sentiments about their target and stay silent
>seem to predictably misread incredibly common use of language, seeing "good vibes" and assuming it means GOAT or inventor being a common example
>Incapable of putting forth or discussing any object level example, preferring instead to talk in innuendos about how there's a large population of informed disagreement, yet fails to link or write something. If an example is provided, it is full of guilt by association and sneering arrogance and little else, and they converge on it like little flies, congratulating each other about how many good points have been made.
> does not seem to be aware of their own constant isolated demand of rigors, or that many of their objections can be answered with less than five minutes of thought
> become obnoxiously self congratulatory about how they used middling to low wit to dunk on their targets
You see, listing a bunch of negative traits is a great way to be truth seeking. If someone tries to engage, it's because they are insecure and you, yes you, personally are correct, no matter how obviously derailing or irrelevant that list is. If you can play to an audience, that's all that matters. :-)
Well, you know, one can choose the quality of their friends, but not their enemies. There are ways to choose the quality of your enemies, but the world is not ready to hear them.
I am puzzled that you include Karl Popper here, his view on the demarcation criterion between science and non-science (i.e. that a hypothesis should at least in principle be falsifiable) is very mainstream among scientists. (I do not know Deutsch' interpretation, though.)
If you believe in malicious plots that aren't real you reduce social trust which is one of the main predictors of sociatal succes. Messing with that has huge potential negative U.
I think this is definitely a common failure mode in general, and I observe it way more often in rationalist circles.
Personally, I experienced the exact same thing with the mask and airborne debates. I still get an emotional reaction every time I see someone wearing a cloth mask.
FWIW, cloth masks are a major help in many environments. They slow down the speed of the air ejected by someone who is infected, so it tends to remain near them. They also absorb moisture, so any wet globules ejected are likely to be absorbed.
Note: Almost all masks are more effective at preventing an infected person from spreading the disease than they are at preventing someone from catching it.
I fully agree with you, but the standard response is that (1) wearing masks is "dehumanizing", (2) we will all catch the diseases anyway, sooner or later, and (3) many people wear the masks incorrectly anyway. Therefore, the proper way to signal sophisticated thinking is to refuse wearing the masks. Especially when the cost of doing so will mostly be paid by the others.
Yeah, I think. There's a common failure mode of overcorrection from "don't blindly follow the mainstream" to believing the most popular contrarian option by default.
You see a similar thing in the left with people over correcting from "America is always right" to supporting genocidal dictators as long as they oppose the USA.
As Scott described in the article, Lab Leak is the "mainstream" view among the general public, even if it's contrarian in academia and prestige / elite / media / whatever you want to call it circles.
JFK conspiracies are also the "mainstream" view. I think it's possible for "mainstream" to be a slight "term of art" even if the raw survey data says otherwise.
Well, there is a big Oscar-winning Hollywood movie heavily implying that the official JFK narrative isn't the whole story, so calling that view mainstream seems reasonable in any sense.
> You see a similar thing in the left with people over correcting from "America is always right" to supporting genocidal dictators as long as they oppose the USA.
Ironically, nowadays the right tends to do the same thing, due to associating the US with Evil Wokeists. It really is a trip, seeing politics flip to basically the exact opposite of 20 years ago.
Well there's the question of how far this was an overcorrection. It may very well be that it was correct to put (as a random number) 80% on it given your state of information & willingness to investigate & priors, while 90% was too much.
(I think it is easy for social-posts, like notable blogposts, pointing out that some belief of the community was likely wrong to end up having the same problem of overcorrecting to "don't disagree" rather than "what degree did we do worse than we should have")
I think that there's almost no difference between 80% and 90%. Even if one believes in objectively correct priors, someone who was off by this much is basically as right as possible.
And there is a danger if offering more precise probability estimates, in particular when offering predictions about the future (and even if you offer them only to yourself): The danger that you may convince yourself that you have more precise knowledge that you actually have, or can possibly ever have.
> I wonder if this is a common rationalist failure mode we need to update on.
Absolutely. I sometimes feel like a very large fraction of comments on any topic on this blog are reaction formation by rationalists who are annoyed at something that MSNBC watchers are saying and have talked themselves into the diametric opposite to rationalize their annoyance.
It's often worse than that--it's rationalists who are annoyed at a caricature of what MSNBC watchers are saying, and have talked themselves into the diametric opposite of that caricature.
Nate Silver justifies his belief in lab leak pretty much entirely with frustration with the mainstream media's attempts at narrative control in 2020 (nevermind that the mainstream media has been breathlessly *promoting* lab leak ever since). It really is sad to see this from someone who should know better.
I don't think this grapples with the problems the Mr Chen case poses for the zoonosis argument. Both sides in the debate agree that Mr Chen contracted symptoms on 16th Dec. This suggests he was exposed around 10th Dec. He thinks he was exposed on the train or in hospital on 8th Dec.
This raises two problems:
1. Mr Chen lived 30km away from the Huanan market on the south of the river, and never went anywhere near it in December, according to his interviews. His case was only recorded because he happened to be transferred to the top tier Wuhan Central Hospital, which is a sentinel hospital for respiratory outbreaks in Wuhan. It is in the north of the river from his local hospital because his relative happened to work there. This is evidence of a geographical bias in he case search. How many other cases were missed on the south of the river because they happened not to visit top tier hospitals, which were largely near the market.
Geographic bias in the case search is also confirmed by:
2. Given that there was a confirmed case 30km from the market around 10th of December, this suggests widespread community transmission across Wuhan in early December and late November. i.e. not that there was a zoonosis in early Dec. The first confirmed case was on 10th/11th Dec (even for that case, it is unclear whether the first case was market linked). There was no published retrospective case search. The first case was initially thought to be 1st Dec. Usually if you do contact tracing and a case search, the first official case goes backward in time. In this case, it went *forward* in time to 10th Dec.
As I wrote in an earlier post, I think it was circulating in rural China long before it was officially detected. Catching it on the train fits in with this quite well. My main hypothesis is that a "merchant" carrying produce (of some sort) to the wet market was the source of that epicenter. (That could actually be widened a bit, as all that is required is that someone infected in a rural village for some reason visited the wet market area for awhile.)
Remember that prior to special tests being developed, COVID was being diagnosed as "atypical pneumonia", at least in a Washington (I presume state rather than DC) hospital.
>Remember that prior to special tests being developed, COVID was being diagnosed as "atypical pneumonia", at least in a Washington (I presume state rather than DC) hospital.
Good point!
Personally, I'm somewhat skeptical about the "geographic focus" evidence because there so very many asymptomatic cases. Even if there had been consistent diagnosis from patient 0 onwards, we would have only been seeing the tip of the iceberg.
( I have to admit, I haven't been following LL vs. Z terribly closely. )
It's common for COVID to have an incubation period of only a day or two, so the fact that he developed symptoms on December 16th doesn't exclude that he was exposed on December 14 or 15th.
Rai et al (2022), a meta-analysis of the covid incubation period for patients infected by early variants from January to March 2020, found that the average time from exposure to symptom onset is around 5.7 days, with a 95% confidence interval of 5.2 to 6.3 days. This suggests he was likely infected around the 10th December.
I think you're misreading the paper. The 95% confidence interval means that there's a 95% chance that the mean incubation period is between 5.2 and 6.3 days. It doesn't say anything about how likely a 1-2 day incubation period is in an individual.
I haven't read the paper (I only read the abstract), so perhaps there's information there about how common a 1-2 day incubation period is.
Does when he thinks he was exposed matter at all? Why should we assign any weight to his speculation? He had symptoms in the hospital on Dec 16th which seems very consistent with contracting COVID in the hospital a few days before that.
I agree it is consistent with that, I am just saying that the average time to symptom onset is 6 days. he could also have been exposed ten days earlier, per this logic, which is a further update from zoonosis and these two effects cancel out, no?
The first officially confirmed case, the shrimp vendor, got sick December 11. So she was probably infected December 5. She probably wasn't the first human case, because she didn't interact with the raccoon-dogs; more likely a raccoon-dog vendor got (quietly) sick first and spread it to some other vendors. It's not even clear she was in the second batch of cases - only 5-10% of COVID gets detected, probably less when it was first starting, so she could be anywhere from one to four transmissions away from human Patient Zero. So human Patient Zero probably got sick in late November or early December.
Let's say Patient Zero got COVID November 27 (with high error bars). He infects other people at his workplace (the market), and some of those people infect still other people at the market, or other people elsewhere (for example when they're on the train home from work). COVID case numbers double every 3.5 days, so by 12/11 when the shrimp vendor starts feeling sick, there are somewhere around 16 cases (with high error bars - and maybe more if there was a second spillover for Lineage A, but ignore that for now). Any of those people could have passed it to Mr. Chen on the train on the 8th/10th/11th/whenever.
My final comment. It is important to consider how different SARS-CoV-2 is to other zoonoses. I have a challenge for zoonosis proponents to find me a zoonosis with all of the following features
- Spillover occurred after 2000 when sequencing became much cheaper
- There were more than a hundred human cases
- There are zero infected animals.
This characterises SARS-CoV-2, but no other zoonosis meets these criteria. Why?
What evidence do you have that there are zero infected animals? I only see evidence that none of the animals that have been tested were infected with an essentially-genetically-identical pathogen, but there are many animals that were not tested. (I assume the claim you're making is about essentially-genetically-identical pathogens, since there are plenty of samples from animals of coronaviruses closely related to SARS-CoV-2.)
I think there are also some viral hemorrhagic fevers that haven't had their animal reservoir identified yet, though I'm not very familiar with this, and I don't know if any of them have had over 100 human cases.
Yes I am saying that they haven't found an infected animal in their testing efforts, which were very extensive, and included testing of species sold at the market in Hubei and across China. Usually, for a zoonosis with more than 100 human cases, they find numerous infected animals and the first people infected mostly interacted with those animals. This did not happen for SARS-CoV-2.
This is true for: SARS-1, MERS, H7N9 avian flu, 2018-2021 Langya Henipavirus outbreak in China, NIPAH, and all other zoonotic coronavirus spillovers with more than 100 human cases.
Some important context here is that according to the WHO, China also claims to have found zero infected animals total, during the course of the entire pandemic, even after COVID swept across the country infecting (most likely) hundreds of millions of people.
It seems wildly implausible that there are really zero animal infections (including those passed on by humans) in a region as large as the entire continent of Europe, that also has millions of susceptible species being raised on farms. More likely China is either failing to identify infected animals or failing to report this information to the world.
2013-16 West African Ebola outbreak; almost 30k cases, no animal intermediate. This is probably the most notable zoonotic episode of the last 20 years apart from SARS-2, I'm surprised you missed it. There are also 7 other Ebola outbreaks that match your criteria. We've been studying Ebola for over 40 years and have yet to determine the animal reservoir. It took 20 years to identify the reservoir for HIV-1's progenitor. Sometimes finding the reservoir is easy, sometimes it's hard. Typically it is easy when you have lots of cases and the virus is not very efficient at human-to-human transmission, because that necessitates lots of separate zoonotic events, which necessitates lots of infected animals. For something that spreads fast (i.e., the kind of virus likely to start a pandemic), you don't need a big reservoir, so you have a smaller target. For example, we *did* find the reservoir for the 2009 flu pandemic, but it took 7 years: https://elifesciences.org/articles/16777
There is strongly suggestive evidence for bats -Marburg virus has been isolated from bats, we've been able to identify antibodies in various bat species that react to Ebola, and in one case we were able to isolate a short bit of RNA that was similar in sequence to Ebola, but nobody's been able to isolate the virus, nor is it clear how it might get to humans from bats.
HIV reservoir was confirmed to be chimps in the mid-2000s, well after sequencing technology was cheap enough to do it. Main limiting factor was just that many chimp populations don't have it, plus having to go into the jungle and track chimp troops and isolate fragile RNA from their poop...not at all trivial to do.
Wikipedia says they found "SARS-like" viruses in civets, which made them think civets were the natural reservoir, but I don't think they ever caught SARS itself in the wild.
There's different types of evidence for infected civets, some of which comes from higher seroprevalence among civet traders.
In May 2003, Guan et al (2003) identified SARS-CoV-like virus in animals in a live-animal market in Shenzhen, Guangdong Province, China. Guan et al (2003) also tested for antibodies among workers in the market. They note that “8 out of 20 (40%) of the wild-animal traders and 3 of 15 (20%) of those who slaughter these animals had evidence of antibody, only 1 (5%) of 20 vegetable traders was seropositive.” This suggests that the majority of the infections of the 11 people with close contact with animals were zoonotic.
Among 508 animal traders, 66 (13%) tested positive for IgG antibody to SARS associated coronavirus by ELISA, while the control groups including hospital workers, Guangdong CDC workers, and healthy adults at clinic had an antibody prevalence of 1–3%. Among animal traders, the highest prevalence of antibody was found among those who traded primarily masked palm civets (72.7%), wild boars (57.1%), muntjac deer (56.3%), hares (46.2%), and pheasant (33.3%). Those for cat, other fowl, and snake were 18.6%, 12%, and 9.2%, respectively.
We have no such evidence for covid. As I have noted, none of the cases in the market worked at the animal stalls; there is no evidence that they had close contact with animals.
That's an interesting result and something similar was later found for wildlife workers in Myanmar and Laos -- antibodies to SARS viruses but the workers themselves having no memory of being sick. However one reason we have no such evidence for COVID is because this kind of testing for workers at the Huanan Market was never released by China. There also is an article stating that one of the first to get sick at the market interacted with animals, although it's unclear to me whether this person was included in the WHO report or not. The first known case at the market, the shrimp vendor, has said she thinks she may have been infected by sharing a toilet with the wild game traders.
Thanks for your deep and well-reasoned dive into a truly important topic. For me, your most important take-home messages were a) Bayesian reasoning has serious practical limits; b) we need to take the potential dangers of viral lab leaks seriously; and c) we need to take the potential dangers of natural zoonotic spread very seriously. One point I haven't seen mentioned is that the Earth harbors roughly 10**27 virus particles.
Probably, and most are marine. But it's a big number even if you cut it x 1,000. It's a big number even if you cut it x 1,000,000. It's a big number even if you...
It's also interesting how b and c support somewhat opposite reactions to gain-of-function research. If zoonosis is the bigger threat, we'd want to learn more about how this happens.
good point. I'd say that on the one hand there probably *should* be more gain-of-function research, but it needs to be performed far more cautiously. ALL such work on virus should be performed in BSL-3 conditions, using only virions that have been genetically crippled to be intrinsically non-virulent. I guess the question is whether the scientific "community" can ensure that some jerk somewhere doesn't do something really stupid. Unfortunately, we tend to underrate the raw power of Stupid.
a) is particularly important. What this blog post has shown beyond reasonable doubt is that "real" Bayesianism is not for amateurs.
A little knowledge is a dangerous thing. Better to only teach students who will be tasked to make difficult decisions in their professional lives (such as medical doctors, psychologists, investigators, judges and child protection officers) awareness of cognitive biases & where heuristics may go wrong, and NOT to also teach them formal Bayesian analysis - as they are unlikely to have the time & information & sufficiently advanced methodological skills to do it properly (plus be aware of all the pitfalls).
My rule of thumb for something like this is 'whose evidence stands up to scrutiny'. I am not smart enough to understand any of this stuff, but I vaguely knew about the evidence beforehand from podcasts and social media. And then Peter just explained it all away effectively without significant pushback. This never happens in UFO debates. Or political conspiracies. Someone spends a billion hours researching sounds smart makes 30000 claims and then you pick 5 at random and 4 are nonsense and 1 is true or misleading or misinterpreted. I notice that my plausible theories of lab leak after hearing Peters argument about the implausability are *not* those advanced by rootclaim and so I assume they are obviously wrong I just don't know why but I'm confident Peter could convince me instantly. Rootclaim's inputs fell apart under scrutiny', which makes me doubt his statistical analysis. Garbage in garbage out. Note my real confidence about who is right is super low because I am dumb. But the debate definitely moved me away from lab leak 70-30
You are conflating the spoken debate part, where Peter had far superior memorized knowledge, to the written parts, where there is basically no zoonosis evidence that survived scrutiny. We'll elaborate in a post soon.
“Okay we got our asses kicked in the mutually agreed format *that we ourselves proposed*, but now after the fact we want to re-litigate in a format that plays more to our advantages” *really* feels like shifting the goalposts in an unfair way.
If you take that route I think it would only be fair to give Peter access to the same resources you have in doubling down on the research and time to refute your refutations.
I agree it should reflect poorly on Saar that he didn't realize that sophisticated debaters are in fact a huge problem (see: https://web.archive.org/web/20190129225836/http://commonsenseatheism.com/?p=1437 and to be clear I don't begrudge him for not knowing about a link rotted post from 2018 by a niche atheism blogger).
But I feel you should be engaging in the object level too, and realize that a spoken debate in fact is not the best at eliciting truth, for usual status reasons as well as the reasons Saar cited:
1.Debate importance is rated via time spent on a topic, not its actual importance, see Scott's point about the UFO detail debater.
2. It mostly explores how many layers of point-refutation each person has at the moment of debate, rather than what the reflectively endorsed endpoint would be. I would rather people take the time to decide how much each point matters to their main point, be explicit about this ahead of time and then decide if hashing it out is a good use of time. This is difficult to do in real time format.
3. If it is indeed true that Saar has found the ~one epistemology~, it seems much better to agree upon that frame (but not its conclusions) beforehand. This is not just to advantage it, but if it turns out to be as dismal as debate afterwards, well, that should give us a pretty good idea how well it generalizes.
That said, you are correct that retooling the die is a social faux pas, and that we should downrate the second round result.
I think if you are as confident on lab leak as Peter is, you should be rubbing your hands together, salivating that Saar is going to eat crow in HD this time.. Maybe not by Peter in particular, because Peter sick of this and won.
I think Saar should state up front that he commits to doing only this one round of welching / sore losering, and have it framed for everyone to see after, if he wants to dodge the "after wit who keeps coming up with reasons why he is right 3 weeks after losing the argument" accusations.
Yes, this was included in my initial assessment of "penalty for not understanding that debates suck". I don't know why people keep thinking it's good to debate! Judges predictably flop the other way they started at the beginning, we know about the existence of truth independent debaters! This is an opinion I've had before I even heard it was a debate, and to my eye this mostly looks like every other case where an amateur debater loses, but had a better case (not that I think rootclaim had a better case, but rootclaim at least pattern matches to non sour grape, honest dismay at debates sucking).
I don't think sophistication is a good way of tracking truth, because sophistication is often displaying that you understand deep nuances and details, regardless of how it connects to the bigger picture. So anything that tracks sophistication tracks more of your ability to fight on specific sub topics, rather than subtopic weighted by centrality to claim.
I think I can be convinced away from this position if Saar goes on to repeat this similar behavior, or if he reflects and decides he was wrong after all. Or if Saar wasn't willing to additionally show his reasoning, or welches on the actual bet.
This post has actually left me with a deep curiosity on Rootclaim itself. To be honest, rather than the Nth post on the topic of lableak I'd very much rather read a primer on how to use the method.
A lesson Scott draws is that Bayes is effectively impossible to apply in complex real-world situations, and that ultimately we are forced to rely on our intuitive reasoning. We’ll just throw a little math in at the end to make sure we aren’t making some clear and obvious blunder. I suggest this means the true epigraph for ACX should not be “P(A|B) = [P(A)*P(B|A)]/P(B), all the rest is commentary,” but rather “commentary, and all the rest is P(A|B) = [P(A)*P(B|A)]/P(B).”
Well, I would argue that Hillel's standing-on-one-foot summary of the Torah was similarly too simple to apply in reality and largely just ends up being a rationalization for our intuitive aesthetic judgements, so I'd say that just makes the epigraph all the more fitting.
Like cost-benefit analysis that is not done by a very skilled and aware-of-own-biases practitioner, hard/real Bayesianism seems to create a danger that uncertainties are hidden in the priors & underlying assumptions & the empirical material used for the calculations, rather than appearing openly for all to see.
> When psychoanalysts claim their therapies work, they don’t mean that someone who just read a two page “What Is Psychoanalysis?” pamphlet can do good therapy. They mean that someone who spent ten years training under someone who spent ten years training and so on in a lineage back to Freud can do good therapy.
And Robyn Dawes in House of Cards showed the evidence doesn't support the idea that training makes them any better.
Jeremy Howick (The philosophy of evidence-based medicine) cites the derogatory acronym GOBSAT related to the problems with "Metis": Good Old Boys Sat Around A Table (and agreed on diagnosis & treatment).
“One read out of 200,000,000 is completely statistically insignificant,” said Pond. “It really had no SARS-CoV-2. There is no evidence based on genetic analysis there was SARS-CoV-2 in that sample. One read out of 200,000,000 — it could have been a low level of trace contamination.”
What’s more, as Bloom’s preprint reports, Q61 was the only swab above a certain threshold for raccoon dog genetic material that contained any SARS-CoV-2 RNA at all: “13 of the 14 samples with at least 20% of their chordate mitochondrial material from raccoon dogs contain no SARS-CoV-2 reads, and the other sample [swab Q61] contains just 1 of ~200,000,000 million reads mapping to SARS-CoV-2.” When Bloom plotted the quantity of animal genetic material found in the swabs with their SARS-CoV-2 RNA content, he determined that there was in fact a negative correlation between the abundance of SARS-CoV-2 and genetic material from raccoon dogs in the swabs.
But also.... What happened to discussion of "NIH-funded gain-of-function research" by Peter Daczak and EcoHealth alliance? Is this the "DEFUSE" grant I'm question? It makes not mention of University of North Carolina - that's new to me, though of course I didn't delve deep into the grant.
I also wish it would be possible to go and collect all the Twitter posts on this topic from when the situation was unfolding, before and after the pandemic left China, including all sorts of claims: spread through plumbing, claims of corona virus circulating months (!) before the official first case, etc. -- do an analysis of how much is real, how much is fake. For example I remember following very closely everything I could read and don't remember racoon-dogs ever being mentioned, but pangolins were.
Pangolins were mentioned in March 2020 as a hypothesis. Raccoon dogs became the more prominent hypothesis by early 2021, when the WHO investigation got samples from the market. Obviously, there are severe limitations on what you can conclude with any confidence from year-old samples, but they're not as much as the limitations on what you can conclude without any samples.
Can someone explain to me Peter’s argument "doubling rates at the wet market points at zoonosis rather than a superspreader event" ?
Intuitively, the two scenario (zoonosis / superpreader event) looks the same to me : an infected (lab worker/racoon dog) (goes at/is brought to) the market and infects market participants. I expect the same doubling rates in both cases.
A least insofar as I remember, "superspreader" was typically used in the public Covid debates to describe cases where the disease vector, for whatever reason, would spread the disease to particularly many new targets. Ie. if a typical Covid patient ended up spreading it to two others (I can't remember the actual numbers), a superspreader would spread it to twenty.
Yeah, that argument was weak and I would have liked to see it either elaborated or conclusively refuted - it's a high-load-bearing part of the zoonosis argument.
Early research showed that the original COVID variant spread bimodally: most of the spread (~80%?) was caused by a small number of vectors. How is the evidence incompatible with the idea that a WIV employee seeded infections that developed over the course of the following week?
Furthermore, even if that model isn't correct, what plausibility advantage does "infected raccoon-dog" have over "WIV employee infected a HSM vendor who then became an vector (asymptomatic or not), infecting multiple people over the course of a week or two"?
The plausibility advantage of “infected raccoon dog” is that there is a good reason for the racoon dog to infect someone (or someones) at the wet market, but nowhere else - raccoon dogs don’t take crowded public transport to and from their office job and their meat shopping.
For the WIV employee, there is an extra layer of coincidence that needs to occur for them to only infect someone else (or at least, only infect someone else in a way that “catches on” and spreads) at a wet market that isn’t particularly close to the WIV itself.
Basically, conditional on the probability of the wet market being the epicenter of the subsequent epidemic, the probability is higher that this is because it originated there, vs originating at the WIV and being carried to the market, not creating other outbreaks elsewhere in Wuhan.
That's fair: I agree that the "infected a vendor" model deserves a penalty relative to "infected raccoon-dog" (all else being equal). A small penalty for the extra layer of coincidence, and a somewhat larger one raised by the question of "why HSM and not anywhere else?" (I don't think zoonotic-origin proponents have convincingly refuted the possibility that ascertainment bias undermines key parts of their case, though I'll grant that my priors have been shifting in that direction.)
Peter's point is that it can't be spreading too much more widely in early December because otherwise when testing becomes widespread you would expect a much higher rate of infection and death therefore we should not consider the HSM initial location to be biased (e.g. there is strong evidence that the first and most subsequent early Human-to-Human infections occurred at HSM). I think Peter (and the Judge's explanations) was extremely convincing that the probability of zoonotic origin is much more likely than an infected lab worker traveling to HSM and starting the pandemic. In addition Peter did a good job reducing my opinion of the likelihood that Sars-Cov-2 is caused by Gain of Function (or any other human manipulation).
Peter did say (something along the lines of) that if you want to assume that a WIV employee left work and went to the wet market then went home and quarantined for a week you are welcome to. It isn't as if he ignored the scenario. I think the point is, if you know you are infected, why go to the wet market at all and if you don't know you are infected then why are you quarantining...
To me, this argument makes no sense. The market was either: (a) a zoonosis plus a superspreader, or (b) only a superspreader. All of the confirmed cases at the market were from human-to-human transmission. None of the confirmed cases were wildlife sellers (unlike SARS). The confirmed cases were spread across the large area of the whole market.
Your b) is actually “a lab leak origin outside of the market, that only becomes a superspreader at the market”. If the origin is not in the market itself, you need to account for the probability that an infected superspreader from the lab (or infected by someone from the lab once or twice removed) infected a substantial number of people at the market, but not apparently anywhere else.
Genuinely asking: But do the subsequent case rates actually support that? Sure, initially all the focus was on the wet market, but if there were other simultaneous epicenters surely that would be ascertainable after the fact?
And lab leak proponents would be highly motivated to find such data. Instead they (at least Saar does) seem to mostly concede that the wet market was the source of the first significant outbreak, plus or minus some low level earlier circulation in weakly sourced or anecdotal data.
To clarify, my initial comment was just making the logical point that the market was a superspreader (and maybe also a site of zoonosis), not that it was necessarily the *only* superspreader.
I don't see why you think other epicentres would be ascertainable after the fact? There was a retrospective case search that ended in mid Feb 2020. There was a market bias in play for around half that time and the case count actually fell for part of the case search. In my other comment, I highlight evidence for ascertainment bias. for example, if you read this report about Zhongnan hospital, a hospital next to the WIV BSL-2 and BSL-3:
"The South China Seafood Market [link requirement] alone ruled out most of our suspected cases." Yang Jiong said that Zhongnan Hospital is 10 or 20 kilometers away from the South China Seafood Market. Most of the suspected cases received in the early stage did not have a history of exposure to the South China Seafood Market, but they had obvious clinical manifestations of unexplained viral pneumonia. However, the district experts refused to identify them as unexplained pneumonia after consultation. They could not be referred to Jinyintan Hospital, and samples could no longer be sent for genetic sequencing"
Because many cases in this area were not market-linked, they could not be confirmed at Jinyintan (the premier infectious disease hospital), so there was limited opportunity to identify clusters in this area.
Similarly, Mr Chen was only registered as a case because he happened to be transferred to a sentinel hospital north of the river. This suggests that other cases where Mr Chen lived (35km from the market) would just never have been found.
You have to assume not just early cases missed because of ascertainment bias, but also some explanation for why there aren't five times as many cases a couple weeks later if it had been spreading a week longer than hypothesized.
Before kvetching, I want to concede that this is a good argument, and it's been a big part of why I've updated away from assuming ascertainment bias.
However, I have never seen anyone substantiate its premises. I think that's important because there are two related areas of uncertainty underlying the point: first, that we got sufficiently accurate epidemiological data out of China during that early period and second, that we know what the R0 was at the time well enough to make the judgement.
Beyond that, the epidemiological dynamics of COVID were weird: more "punctuated equilibrium" than steady spread. Admittedly, my understanding of that is mainly rooted in subsequent history, when NPIs were in effect and people were on guard, but this is not an uncommon model for how diseases spread.
Maybe Peter addresses these issues in the debate, but I have yet to see anyone do so to my satisfaction, so my position remains something like "there's only a 50% chance we have information enough accurate to make these kinds of judgements."
(Edit - reading down thread, I see now that Peter linked to his analysis of the spread, which I'll read as soon as I have time.)
So, am I right in thinking you don't think there was ascertainment bias in the case search? I'm not sure I understand. In that quote above, you have >10 hospitalized cases at Zhongnan that don't get registered because they didn't have a market link. Assuming there are ~10x total cases, that's already 100 cases you don't see for market biased reasons, (in an area around the WIV)
Whether Patient Zero is a careless WIV lab tech who is also a wet market customer, or an unlucky wet market vendor, they both go home at the end of the day, probably a family, and very probably a normal life where they go out and do things at least until they get sick. So in either case, the pandemic starts with A: a superspreader event at the wet market and B: an infected guy going around spreading the disease the ordinary non-superspreader way. I'm not seeing how you can use observed doubling rates or anything like them to distinguish between the two.
If there were confirmed cases predating the market outbreak, those would give us great insight into what happened, but AFIK there aren't. The fact that there aren't, just tells us that Patient Zero infected very few people outside of the market and they all just thought it was a cold.
I feel like he's leaning very heavily on the average spread, but that's essentially his argument. That if early COVID spread like we know it did later, then the size of the spread would have been much larger if it was earlier, which Saar's argument requires.
Given small numbers at outset, I don't think he can actually rule out low and intermittent spread for a few days maybe weeks before catching in a more randomized way. If someone was infected early and then only spread to one or two other people and they only spread to one other person, it may have delayed the virus from spreading as we saw it later. Once it hits a certain threshold that seems very unlikely, but below that threshold we really can't say that.
Indeed. There is a paper (can dig it up if you want) showing that (iirc) the model number of people to infect (if you're infected) is 0 and the next most common is 1. Many many chains either die out completely or take several cycles to explode.
Yes, I’m very interested if you can find that paper. It could help shed light on the the question "on average, how much time do we expect to see between the first case and the first superspreader event ?"
One key point is that Peter insist on "very little time because exponentially doubling rate", which I believe is wrong.
Yes, I agree - I think the uncertainty about the start date is far larger than he's implying from "we know the doubling rate, so therefore it's tightly bounded". If you know the exact doubling rate you can talk about aggregate dynamics using this math when the number of infected is large, but when the outbreak is small in size the variance can be very large.
Now onto the specifics - COVID-19 transmission is characterised by a large degree of overdispersion. This means most people infect very few others, and the vast majority of transmission is driven by a small fraction of the infected. Multiple studies from 2020 found that over half of infected people never infected anyone themselves. Here are some links:
Sure, but why would the virus only “break out” at the market? I don’t think the claim is that these early cases were some less virulent strain (the A and B types are both pretty infectious). So if it was spreading at a low rate throughout Wuhan, there are a *lot* of potential breakout sites where it didn’t break out, or create multiple epicenters at roughly the same time as the wet market.
Which isn’t totally implausible! But it’s probably less likely than “no really, the first cases were spread in the wet market and not somewhere else”.
That's the other part of Peter's argument that I didn't like, the almost slavish following of "less likely = wrong" when at any point something unusual could very well have happened. That seems to be Scott's point at the end, where we know for sure that at least one of two very unusual things actually did happen.
I don't know if I have enough knowledge to say that it is "probably less likely" or not, but that's not the final determiner of truth.
This strikes me as a weird argument to defend someone whose whole project is rigidly following Bayesian probability. “Probably less likely” IS the final determiner of (as-close-as-you-can-get-to) truth.
Something as high as like 60-70% of people infected with Covid appear to not infect anyone (in a seronaive environment). It might even be more skewed than that once you account for under detection of asymptomatic cases to begin with (I'm guessing the never symptotic don't spread it much)!
This makes chains that die out, or chains that barely survive for a few generations and then explode completely plausible.
I understand “superspreader” to mean something like “an infected individual who sheds way more virus than usual in a confined environment creates a localized, discrete spike in infection rates”. Or in any case “a discrete event with a much higher rate of infection than the base rate expected by the observed growth rate of the epidemic”.
Peter is saying the wet market doesn’t look like that, it looks like a “normal” rate of infection and growth in cases.
Why does that matter? The two sides of the argument both need to explain why the epidemic was first detected and seemed to spread from the Wuhan market, rather than somewhere else.
Saar’s claim is that basically Covid leaked from the lab sometime prior to the wet market outbreak, and was circulating around at a low rate, but a “superspreader” event in the wet market made it the epicenter of subsequent infections. This theory takes a major hit if the spread in the wet market was not atypical, because in that case the market should just be one of many similar outbreak sites started by the various infected folks circulating around, rather than a clear epicenter.
Peter claims that the wet market was the actual origin of the zoonotic transmission. It looks like it all spread from there because it *did* all spread from there. The fact that there was nothing “super” about the spread at the market supports this theory - wet market infections don’t need to swamp other potential epicenters because there are no other epicenters.
Yes, this does make sense, which is the point. Such a scenario is a priori unlikely for lab leak, since it requires a lab worker to go to the market and infect someone there, but infect no one anywhere else, either in the lab or in between, since there are no other early clusters. To make this hypothesis seem more plausible, Rootclaim asserted that the market provided very good conditions for much faster growth of Covid, possibly even a superspreader event (i.e. a case where a single person infects a dozen or several dozen people all at once, greatly in excess of the average number of secondary infections, which for early covid is less than 2. This would explain why the market dominates the early cases, but as Peter points out, these claims have no evidence and are contradicted by available data. If the spreading rate is the same inside the market as everywhere else, then the strong clustering of early cases to the market is very good evidence in favor of the pandemic starting there (and thus, in favor of zoonosis).
Bu the upshot is that this argument is specifically a response to Rootclaim's argument, so if you don't believe the latter, then this part would be irrelevant for you.
This is a better summary than I was able to put together.
I think Rootclaim needs to defend their assertion of a superspreader event at the market, because otherwise Peter seems to have convincingly shown that it wasn’t, and if it wasn’t it’s pretty good evidence in favor of the wet market theory.
I don't think they're going to do that, and from Saar's point of view they don't really need to. This is his whole shtick with "both sides shouldn't have high BF" and Scott's analogy with the Muslim UFO spotter. Maybe it wasn't a superspreader event at the market; that was just one possible explanation. But this example (allegedly) shows that such explanations are not that rare or hard to find, and so this evidence must have a low Bayes Factor.
In something like this, you might end up with weirdly extreme probabilities simply because global pandemics are rare, and will necessarily be the result of a chain of individually unlikely events occurring. I know in theory all these probabilities are supposed to be “conditional on Covid actually happening” but that’s easier said than done.
I get the Muslim UFO Spotter analogy, but I still feel like Saar is a bit dodgy here.
Your model should account for all possibilities even if there is just a “what if this did happen, but not in one of the specific ways I already quantified” uncertainty term.
If your conclusion is correct, but for a reason your model didn’t even anticipate, you just got lucky.
As it is, it looks like Rootclaim simply didn’t consider the possibility that the wet market wasn’t a superspreader, which raises the obvious question “what else did they miss”.
> In something like this, you might end up with weirdly extreme probabilities simply because global pandemics are rare, and will necessarily be the result of a chain of individually unlikely events occurring.
I think this is likely to be true.
> I still feel like Saar is a bit dodgy here.
Oh, I agree that Saar's argument is extremely dodgy. His argument relies on the Bayes Factors that he produces being both high and extremely robust. If that is not the case, it totally falls apart. I was just explaining why I don't think Saar is likely to further support the superspreader claim, or care about it.
Which is a shame since Rootclaim kept repeating the claim in various contexts (based purely on Scott's summary. I haven't actually watched all the debate itself). Given that Peter indicated this as a week spot in their arguments, repeatedly referring to the superspreader event without justifying that claim is ... strange.
I'm confused about the "both sides shouldn't have high Bayes factor" claim. With plenty of ordinary claims, I can very easily come up with extremely high Bayes factor pieces of evidence on both sides.
Like consider the question of whether I was born in India or in North America. My last name is a very high Bayes factor piece of evidence in favor of the India hypothesis, and my first name is a very high Bayes factor piece of evidence in favor of the North America hypothesis. I haven't thought of any evidence I can provide that has a high Bayes factor in favor of the hypothesis that I was born in South America over the North America hypothesis (the fact that I currently live in Southern California is probably single digit Bayes factor evidence in favor of South America over India).
In this case, there's very good explanation available for the anomalous evidence (my father was born in India), and not every false hypothesis has an equal amount of high Bayes factor evidence.
But if you do a very serious search for anomalies related to something, you're likely to find a few anomalies in favor of some particular false hypothesis, even if most false hypotheses don't have any in their favor. There's a reason we're debating whether it was from the wet market or from the virology lab, and not considering whether it spread from pet dogs, or was planted by the CIA, or came from bats living in someone's attic.
I believe what Rootclaim argues is that if you have a high BF on conflicting hypotheses, then one of them is either wrong or p-hacked. Your name example is a good one, though, since each piece of evidence is so similar it's hard to allege p-hacking for either side. I wonder how strong each of those pieces of evidence actually is? Like "Kenny" being 1/1000 in favor of NA and "Easwaran" being 1/5000 in favor of India? I have no idea what these numbers actually are.
And of course, unlikely events can still happen. For most people, their first and last names probably don't provide such a high level of disagreement. I think this has to be the case, right? If lots of American-born people had Indian last names, then the Bayes Factor for "Indian last name" would by definition not be very large? I'm just thinking about this quickly and not doing math.
> But if you do a very serious search for anomalies related to something, you're likely to find a few anomalies in favor of some particular false hypothesis, even if most false hypotheses don't have any in their favor.
Yes, I think this is related to Rootclaim's assertion of p-hacking. But I think this is also an easy failure mode, which several people have pointed out is prevalent in the online lab leak community, of basically going on fishing expeditions with a high number of degrees of freedom and not accounting for the number of opportunities for an anomaly.
I also have no idea - I could imagine them being anywhere from a factor of 10 to a factor of 1,000,000!
In any case, I think that it's a bit awkward to talk about "p-hacking" when we're trying to do a Bayesian analysis. We know that most of the evidence people come up with is going to be gathered in a highly non-random way!
I'm a huge proponent of Bayesianism, but I don't see much prospect for explicit calculations of this sort to be that useful in this kind of format, at least in part for this reason.
It shows that the market outbreak can't just be an artifact of ascertainment bias because there weren't enough cases at that time in the outbreak, the size of the market outbreak in early-mid December is roughly what you'd expect to find in all of Wuhan.
So it can still be a lab leak, it just needs to go straight from the lab to the market and nowhere else (with both lineages). That's very unlikely and needs to be accounted for in any sensible bayesian analysis.
If you think that the exponential model is too simplistic to make that point, I've also done more detailed epidemic simulations to demonstrate a similar thing:
Universe 1: LL is true. Somebody at WIV contracts coronavirus and goes about their daily life, eventually effecting numerous people in Wuhan. If all the starting cases suddenly show up at some superspreader event (a football event? an orgy? idk), this makes sense. If they all suddenly show up at some specific, non-superspreader event, this doesn't make sense (they'd infect somebody at HSM, but also at the grocery store, at the brothel, at the school their kid goes to, etc).
Universe 2: Z is true. Somebody contracts coronavirus from the raccoon dog, and goes about their daily life. Since they are probably the raccoon dog stall vendor, they spend much of their potentially-infecting-people day at the HSM, so they cause most of their "doubling every 3.5 days" infections to happen there. They're also more likely to infect people at the HSM. This same logic could apply to a superspreader event, but it's less likely - imagine somebody at the football game (e.g. an athlete) being the cause of the superspreader event at the football game.
While it's possible either way that somebody got the infection really started at the HSM, it's much more likely for it to happen in the LL universe if HSM is a superspreader, and much less likely for it to happen in the LL universe if HSM isn't a superspreader.
Is it really such a coincidence that the outbreak occurred naturally nearby the lab? It's reasonable (I guess, I don't know) that the lab was strategically placed there in the first place due to the area's wildlife, which host many animals relevant to their virology research. If so, the alleged coincidence significantly fades.
Placed in the middle of a city of 10 million people to be near wildlife? No, the virologists in question were traveling over a thousand miles to random caves in the middle of nowhere to collect viruses which they were then bringing back to Wuhan because living in cities is more fun than living in primitive villages near bat caves.
My theory is that the local government of Wuhan knew they were at risk, because so many rural farmers in the area were bringing wildlife into the wet market. So they funded the WIV. As you say, a big research institute kind of has to be near a city. It's not going to be Bruce Wayne living in a bat cave.
I wish I knew more about the internal politics of China though.
It's somewhere between a pandemic started in San Diego and you discovered San Diego had a relevant lab (8th largest city in the country) and Los Angeles county has a pandemic start in downtown and there's a relevant lab in Pasadena (based on size of Wuhan vs American cities).
I don't think either one is really a fair comparison. Wuhan has *11 million people*, so it's roughly the size of NYC and LA combined. But it's also in an area where there are still a lot of desparately poor people living in the country and hunting whatever animals they can find as game meat, on a scale which just doesn't exist in the US. "Anything that walks, flies, swims, or craws..." (https://www.quora.com/Do-the-Chinese-eat-everything-that-walks-swims-flies-and-crawls)
I was using LA county (although thinking of metro area although the two are not super diffrerent in the case of LA) not LA City (LA metro area has 12 million people). My assumption is a pandemic starting in LA would most likely be caused by travelers so it wasn't meant to be exactly the same but rather it's not surprising for there to be a city/metro area with both a lab and a pandemic. When I was picking US cities (metro areas) my though it that many people had not heard of Wuhan until 2020 but it's not some podunk village.
This answer misses the point. Does it make sense to build a lab tasked with collecting samples for coronaviruses research in the London, in the same way it that it does in Wuhan?
From the post: "Peter: As mentioned earlier, the DEFUSE grant was rejected. Further, the grant said that the Wuhan Institute of Virology was responsible for finding the viruses, and the University of North Carolina would do all the gain-of-function research.".
I don't know what exactly "responsible for finding the viruses" operatively means, but it might mean that their location is indeed non-random, and that this was their "competitive edge" so to speak.
It's "non-random" in the same sense that Miami being a major banking hub for Latin America is non-random. It's not especially close to the places that it's serving, but it is a natural location to be an intermediary between the places that it's serving and the broader world of finance or biological research.
I don't guess, I've looked into it, and no. The lab wasn't placed in Wuhan because of the bats or pangolins or racoon dogs or whatever live near Wuhan. First off, the bats we care about live far away from Wuhan, and any pangolins or raccoon dogs we care about have to have lived near the bats. Second, it's the "Wuhan Institute of Virology", not the "Wuhan Institute of Coronavirology" or "Wuhan institute of Zoonotic Virology"; it was not initially meant to specialize in any particular sort of virus that would call for any particular location.
Probably the WIV is in Wuhan because when it came time to found a big Institute of Virology, the next guy on the CCP's "we owe this guy a favor, let's build a big infrastructure project in his district" list was from Wuhan.
I think the "lab leak" side of the debate focused too much on the probability that SARS-CoV-2 was engineered, and didn't properly account for the scenario that SARS-CoV-2 was a natural virus collected by virologists that accidentally leaked. In my opinion the latter scenario is more likely than the former.
Tell me more. My impression was that one of the strongest pieces of evidence for lab leak is the weird furin cleavage site. If you say it wasn't engineered, you lose that evidence - what do you gain in exchange?
Why? Couldnt they be stupid and reckless in multiple ways; a little bit of virus breeding, some humanized rats, some random crisper edits all in an attempt to make a bioweapon only to be deluded that they are making any progress. You mix it all together and you get a handful of engineered gene transfer events and a very unclear "zoonotic" pathway
No. CRISPR creates cuts and introductions at planned sites and with planned new sequences. Having something that no one would have planned is not particularly likely.
It is 2024, there is 0% doubt that the Furin cleavage site is natural and was not introduced by an engineer. Long list of reasons include (not comprehensive):
-All previous FCS created by researchers (for example here in 2006, or here in 2009, all published years before the "secret" DEFUSE proposal suggesting similar work got played up as a smoking gun) have avoided using an "insertion", and rather chose to "substitute" - rewrite - existing nucleotides in this position to change the encoded amino acids to create an FCS. This is necessary and basic biology because the addition of new amino acids to any spike protein will likely disrupt its structure and break the whole thing.
- Curiously, an additional, unnecessary proline residue is part of the insertion. Proline is not part of any FCS motif, however, the amino acid is considered a helix-breaker; something that will mess up your protein structure with a high probability. Why risk introducing that?
- Even more curious, the whole sequence insertion is "out-of-frame", meaning the nucleotides were not introduced neatly but rather staggered for no reason, again something that scientists would not do ordinarily (or at all)
- However, the FCS insertion is right next to an RNA sequence motif prone for copy-choice recombination, arguing for this natural mechanism bringing it about (and copy-choice recombination is reading-frame independent and favors out-of-frame insertions 2:1)
- The GC content of the insertion is 83%, very unusual for artificial introductions
- The coded amino acids themselves make an odd, non-canonical, and weak recognition motif (RRXR) for furin-like proteases; a motif that might ordinarily not work compared to strong motifs such as RRSRR (which would usually be used by virologists).
- The FCS cleavage at S1/S2 destabilizes the original (Wuhan-Hu-1) spike protein, which only got compensated partially when the D614G mutation emerged quickly in the early pandemic, arguing for a recent acquisition, not design or cell culture
- Furthermore, potentially because the FCS in the original SARS-COV-2 is not that stable, it is lost very quickly in Vero cells and other typical cell culture systems. Selection pressures seem to run against it in these research-associated amplification systems; so how exactly did an FCS-containing virus that first has to be amplified in cell culture before it can infect a researcher escape from a lab and do so with an FCS?
But these alone is not proof. Who knows what weird and arcane experiments those Chinese researchers would cook up in a lab, right? (Some lab leak truthers with no molecular biology experience would say. For actual genetic engineers, the situation is already clear as day) Fortunately, we do not need to rely on experience or speculations about any motivations or experimental procedures Chinese scientists might have followed.
This is because nature just has more secrets hidden in the FCS that humans could not have come up with in the first place:
After intensive mechanistic studies of the FCS in SARS-CoV-2, researchers discovered a hitherto unknown synergistic interplay between the odd, suboptimal FCS and the genetic backbone of the virus (which we by now know is natural because of recombination).
- We have learned that pre-cleavage of SARS-CoV-2 S1/S2 by furin is what boost TMPRSS2-mediated cell fusion and thus impacts cellular tropism and transmission dynamics (ergo infectivity)
- But it turns out, that was not the full mechanistic story. In nature, no genetic element acts alone
- The viral QTQTN amino acid motif is an uncommon natural sequence in several sarbecovirus spike proteins directly upstream of the FCS sequence
- In cell culture experiments, QTQTN is not stable and repeatedly lost, similar to the FCS. Loss results in impaired viral replication.
- QTQTN determines how tightly the loop harboring the FCS is bound to the spike protein, thus it regulates spike protein stability and shapes how well the Golgi-bound furin-like proteases have access to recognize the FCS motif and pre-cleave at S1/S2
- The QTQTN motif is also glycosylated and loss of that glycosylation impairs viral replication as well.
- Researchers showed beautifully that there is an intricate functional interplay between the FCS, the loop length, and glycosylation: These separate but co-dependent elements work together synergistically and all elements are ultimately required for efficient FCS pre-cleavage (and with it, impact viral replication and pathogenesis). In nature, no genetic element acts alone.
Co-dependency and synergy between genetic elements are hallmarks of evolutionary selection; these type of complex interactions are extremely hard to design for even with perfect knowledge of all structural components, which nobody had in 2019. In the case of SARS-CoV-2's FCS, there is no conceivable way how engineers could have designed the observed - but hitherto undiscovered - mechanistic synergy purposefully.
Take this together with all the other unexplainable design choices listed above; and you have your answer. The polybasic cleavage motif recognized by furin was not artificially inserted
Isn't the out of frame point relative to the backbone used? On the lab leak hypothesis, the backbone used is unknown, so we don't know the insert is out of frame
Nope. Out of frame with respect to the used reading frame for the whole spike, and that is based on the coding amino acids that barely differ between the various spike proteins because they are functionally conserved
I see, thanks. Is it possible that there might be differences between the spikes of some virus that they had that we don't know about that would make the insert in-frame?
Short answer no, longer answer: still no, but never say never; but we would need some really strong evidence to suggest something like that being possible
Yes, it's relative to whatever nucleotides are there already. Though I tried aligning it to every known similar virus and it was out of frame for all of them.
In theory, the leading Serine could have been spelled with different codons, in whatever virus the lab started from. There are 6 codon choices for serine, only one would work, so I called this a 1 in 6 chance at the debate. Probably a crude way to do it, but Rootclaim wanted to quantify everything.
So then isn't it wrong to say that the insert was definitely out of frame? It is correct to say that if there were an unknown virus that the WIV following the workflow outlined in DEFUSE, the insert could have been out of frame but could not have been. From a Bayesian point of view, one has to weigh this against the possibility that the furin cleavage site emerged naturally in a farm in Hubei. The 850 sars-like coronavirus don't have a furin cleavage site, suggesting it is selected against in nature.
One also has to think about the timing of the acquisition when weighing the probabilities. The furin cleavage site emerged the year after the WIV was on a proposal to insert the site. The mutations in the virus also suggest that the furin cleavage site emerged recently around Oct/Nov. Choe and Farzan (2021) argue that “The notable emergence of D614G suggests that the acquisition of a destabilizing furin site was a recent event”. So, zoonosis has to posit that the furin cleavage site emerged only once naturally, was never found in nature despite them looking, and caused approx 1-2 spillovers in Wuhan city.
Yes. I dealt with it in a probabilistic manner by assuming the original virus could have different codons.
I suppose what you really want is the ratio, i.e. 1/3rd of natural inserts would be in frame, but let's assume closer to 100% of lab inserts would be in frame, so you could change my numbers that way. But I was already being generous by assuming that the natural viruses have even priors for all codons at that spot, when it's actually out of frame for every one we know about.
The presence of the furin cleavage site itself is dealt with separately in the probabilities. To some degree that may just be included in your priors -- i.e. the judges both assumed a serious coronavirus pandemic on the order of 1 in 100 years, it wouldn't be serious without the furin cleavage site, so it's double counting to include that.
Saar used a factor of 2, that the lab pandemic is twice as likely as the natural pandemic to have such a site. That's probably double counting, but maybe not, since his pandemic prior was 1 in 20 years. I could see going up to a factor of 5 here since we don't know exactly how often the serious coronavirus pandemics are.
Then the last question is do you get separate odds for it being an "inserted FCS" vs just an FCS. This is the real thing Saar thinks is important, he gave it a factor of 50 at the debate but later said he privately thinks it's 1 in a billion, or something like that. He thinks that the lab would totally put a 12 NT insert into a coronavirus but nature could never do that.
Strangely, he actually has that backwards -- no lab in history has ever inserted 12 NT's to make a cleavage site. They typically mutate existing amino acids, rather than inserting. On the other hand, nature has shown that it repeatedly does 12 NT insertions at random spots, including sometimes at or near the FCS. So the odds here probably actually lean towards nature, if you understand how virologists normally work (Saar does not).
It's not the least bit clear that the DEFUSE grant suggests to "insert the site" -- that's a highly motivated reasoning that few virologists agree with.
Finally, the fact that this is a sarbecovirus with a furin cleavage site is kind of weird -- it's never been observed in such a bat virus. But those are enteric viruses, not respiratory viruses, so the FCS is selected against. In an intermediate host, it's quite likely that this feature would be selected for. We just don't know of many sarbecoviruses which have spilled over into something else... I think that's just SARS1, SARS2, and a couple pangolin viruses? Maybe there are others I don't know about.
So it's kind of a 0 divided by 0 situation, going off prior odds. 0 out of 5 labs have done anything like this. 0 out of 3 natural viruses were just like this.
Very hard to look at that situation and confidently assign some strong bayes factor for the lab.
Many technical reasons. Maybe the simplest: Viruses adapt to their host system. The original SARS-CoV-2 could not infect mice. Even with mice carrying a human copy of the ACE2 receptor, the virus shows none of the characteristic adaptations that it acquired once it learned to infect mice (alpha, delta varients) that one would have expected to see earlier if it has ever been passaged in such a system. The real problem is actually transmission, but that is more complex to quickly write out here.
The existence of the furin site means relatively little because no FCS implies no pandemic and we wouldn't be asking the question. The strongest evidence for engineering is how closely the virus matches the DEFUSE proposal.
Weissman does devote some space to what he calls the ZL hypothesis and finds that even if there was a purely zoonotic spillover, it probably made it to Wuhan via research.
Good point -- it's easy to double count in a bayesian analysis, so if you start with the odds for a serious natural pandemic and a serious lab leak pandemic, you probably can't go on to include things like "the virus has a furin cleavage site" or "the virus binds to human ACE2". (Rootclaim did both in their analysis)
I lean towards lab-leak and am skeptical early Covid would have developed in nature.
Beyond that, what I don't think is addressed, I think it's significantly more likely that a new pandemic coronavirus that pops up in the one city in East Asia with the biggest coronavirus labs, rather than any of the other hundreds or thousands of cities with wet markets, would be an entirely natural virus that made its way there due to improper safety precautions by the lab. The lab collected coronaviruses, was full of both resident and visiting researchers who work with coronaviruses, and was known for having lax safety standards.
Even if Covid wasn't engineered in the lab (I think it was), I think it's more likely that it would be a natural virus that appeared in the wet market by way of proximity to the lab than by way of an infected wild animal brought directly to the wet market from the wild.
This would be essentially a zoonotic lab leak but I would award it to team lab leak.
I don't think there are a billion people that live in cities as prone to wet lab spillovers as Wuhan is. Given that 11 million of those people live in Wuhan, this is at most a factor of 100.
Weissman is his blog post (linked by others - https://michaelweissman.substack.com/p/an-inconvenient-probability-v57) makes a pretty substantial seeming point that Wuhan actually had a very low share of the wild animal consumption trade. So Pr(Wuhan|Zoonotic, wet market) <<< 0.01.
Scott, I don’t understand understand your question. The only reason the lab leak is considered a possibility at all is the improbability of a coronavirus pandemic starting right next to one of very few global coronavirus research centers (indeed a facility which at that time had sampled and sequenced the closest-then-known relative to COVID-19).
Additional claims about furin cleavage sites may support certain scenarios within the lab leak umbrella, but their strength or weakness does not detract from that starting improbability.
What you gain in exchange is that you no longer need to assume that the Wuhan virologists were actually doing any engineering of their viruses.
And the furin cleavage site doesn't look engineered to me, someone who wanted to engineer a furin cleavage site would use different amino acids (as the debater mentioned).
I'm guessing that Wilf didn't argue this theory for two reasons:
1) He views it as much less likely than a leaked gain-of-function experiment. Rootclaim's website shows a probability of 89% of gain of function escape, but only 3.2% for a zoonosis collection accident.
2) According to their public discussion, it was explicitly out of scope of the debate:
Peter Miller:
>Would a zoonotic collection accident be considered a lab leak or zoonosis, for the judging of this debate?
Saar Wilf:
>The judges will have to decide which of the two hypotheses is more likely: GoF vs unassisted zoonosis. So it's possible they are both wrong, and there will still be a winner.
That's really interesting! I had convinced myself that zoonotic collection accident was the main plausible lab leak hypothesis, though gain-of-function leak is also relevant.
It is unfortunate that there's not much discussion of the three hypotheses explicitly - everyone seems to present it as a binary choice (though they sometimes differ between which two options are presented).
I don't know of any fourth hypothesis that is meaningfully supported by anyone.
To be honest, I find the zoonotic collection lab leak theory to be the least interesting variant of the lab leak theory. If all that a Wuhan researcher did was move an existing virus into a populated area, then they did nothing for the virus that could not have been done by a miner who stepped in bat shit.
Right. But the lab is continually going out and looking for new caves full of unsampled viruses, while miners look for a site with minerals and keep working there.
If the pandemic was caused by 'zoonotic collection', there is every probability the pandemic might otherwise have never happened. Scientists take their samples from very remote locations, and deliberately get close to bats in a way that most miners would be averse to.
This is intuitive, because no pandemics have been caused by miners, despite there being many more miners underground at any moment than virologists.
In other words, it would be a massive deal. And going forward it would have huge implications for what precautions scientists should take when sampling (if they're permitted to sample at all).
I don't think it's likely for bat virus sampling to produce a pandemic. In the whole history of the WIV, collecting 20,000+ samples, they only ever managed to culture 3 sarbecoviruses. Usually it's just enough RNA to get a sequence but not so much that you can get yourself sick.
There are also signs that the starting virus was already adapted to a respiratory system, whereas the bat viruses are enteric viruses. The furin cleavage site is the most notable such feature, but there are other smaller ones. So it was in some species other than a bat, before it jumped into humans.
I suppose if researchers went out sampling raccoon dogs or some other animals and a researcher got sick, that could do it. But I don't see why the hypothetical scenario of researchers going out and doing that should be any more likely than zoonosis at the market, where we know for sure they were importing lots of exotic animals.
The main thought I have is that while there’s a lot of people gathering wild animals for wet markets, the relatively few lab workers going out to gather wild animals are specifically looking for viruses, and will thus be picking up disproportionately many (whether it’s overall greater or equal or less is not obvious).
Another way to look at it is that the only known traffic of sarbecoviruses from the location of all of the closest known relatives to covid was via science, not via the animal trade.
Exactly. The reason to at least consider 'zoonotic collection' is that we know researchers were going and continue to go collect samples from remote locations where the closest known relatives to coronavirus exist. Only one cross-species jump is required.
In contrast, all animal trade scenarios currently involve multiple cross-species leaps between unknown animals with essentially no idea as to how, when, or where that occurred.
The issues with extreme odds, focusing on the big picture, and eating one's losses are all related to the same phenomenon: The high complexity of the underlying processes.
Reality is not generated by a simple "zoonosis" factor competing against a "lab factor", as one would think if one naively took the simplest Bayesian models seriously. Rather there was a wet market and a lab and a bunch of people all doing a bunch of different things, etc.. Way too complex to explain. (Though because of funky math where we condition on the fact that there was a pandemic, we also can't ignore that there is a competition between these two factors; they just aren't the generators of reality, and therefore not of the evidence either.)
It is extremely common to have pieces of evidence with extremely low priors, because these pieces of evidence contain a lot of information about the specifics of these underlying processes, and probability drops exponentially with information. Notably, contra focusing on the big picture, accurately integrating this evidence requires building a narrow, specific picture of what happened, because otherwise you don't really know how to evaluate the interactions between the probabilities when conditioning on multiple things.
As you build these complex, low-probability models, that means you end up eating a bunch of losses, in rough proportion to the evidence you add to your model to account for the complexity. If your resulting model of what happened doesn't have very low prior probability, you're not thinking in enough detail.
Yeah, I was very unclear what's going on with the claim that it should be unlikely to find high Bayes factor evidence on multiple sides of a debate.
It's certainly unlikely that any particular high Bayes factor evidence exists for a false hypothesis - but many high Bayes factor pieces of evidence for the *true* hypothesis are *also* very unlikely!
But if a lot of people are digging very hard, they're going to find a lot of unlikely things. And if there are some false hypotheses that are in some sense "close to true", you'll probably find a few things that strongly support them.
I'm somewhat unsure about the math for Bayes factors per se. Assuming you're doing your Bayesian updating correctly on a good model, you should be well-calibrated, so you shouldn't be updating outside the 0.1% to 99.9% range until you actually have a ~final conclusion. This sure seems like it means that strong Bayes factors should only line up for the true hypothesis.
I think the way one gets extreme Bayes factors is model misspecification. Like both hypotheses often need to eat the odds for the information contained in a piece of evidence, as they both need to be updated to give high probability to that information, so the low odds for the information doesn't directly give strong Bayes factors. But then the models should be ready to predict similar information in the future, and if they confidently disagree on what "similar information" means, that gives you the crux for what settles the debate.
But if one of the models get updated wrongly, for instance by not adding a factor that explains the information, then when similar information comes around, the model that was updated wrongly gets forced to eat this information too, which might turn into an extreme update against it.
Did either side get into the US intelligence assessments at all?
It seems to me that by far the strongest argument for lab leak is the US government conclusions to that effect. These agencies have an awful lot of raw data that no one else does, and the resources to put a lot of analytical effort into the problem. Obviously, they're fallible but they have access to the same scientific data as everyone else plus whatever kinds of useful classified raw data the US government has scooped up over the years and they employ experts on stuff. That's enough for a pretty decent presumption in their favor, especially when so much of the debate centers on data from Wuhan that could easily have been manipulated.
I'd find the zoonotic origins argument a lot stronger if someone from that side could explain what the FBI and DOE may have seen that led them to get it "wrong" based on some premise other than incompetence.
I also found it odd that official government positions were not discussed. Even if one side could dismiss them, you'd think they'd still talk about it?
Of course, independent researches without access can't talk about classified data they don't have, and that makes the argument an appeal to authority instead of something that can be modeled and debated.
You can absolutely talk about classified data you don't have -- historians and journalists do it all the time and learning to read through redactions is a fundamental skill in those fields. There are still valid inferences to be drawn even if you can't see the underlying material; you read around the redactions and see what could be under there and what it could mean.
In this case, for example, the fact that different intelligence agencies disagree with each other indicates there is not some kind of classified smoking gun. Like, if there were tapes of the scientists saying "oh shit we did this" intercepted by the NSA, then all the agencies would line up behind that.
We also know that the FBI and DOE both think a lab leak is the explanation, but according to the declassified report, they reached that conclusion for *different* reasons than one another. We also know that the DOE changed its position between 2021 and 2023 based on some kind of new intelligence received in that period, whereas the FBI thought it was a lab leak all along. Those are some very interesting data points.
You can also consider what the intelligence community is geared towards understanding and what kind of information they could plausibly have. These are not public health agencies. They're geared, more than anything, towards gathering and analyzing information about the Chinese government and military. So, they real comparative advantages in understanding the implications of things like Chinese cover-ups or disinformation but no comparative advantage in a debate about furin cleavage sites or whatnot. That's also important and gives us some sense of what could fill in the blanks.
I haven't done the work myself to work through this (and was hoping someone else would because I'd be very interested) but there are also plenty of bread crumbs in what has been publicly released that you can start following.
The most important exercise, though, is just to try to put yourself in their shoes and think "What could they know that would lead to conclusion X" and especially "What could they know that would *incorrectly* lead to conclusion X." If you can't give a compelling answer to that second one, then you have to rethink your own confidence level.
Those are fair points, but the problem is made more complicated because in addition to not having access to the evidence, we also don’t know in any detail what the conclusions were. The DNI says that the FBI and DOE both “assess that a laboratory-associated incident was the most likely cause of the first human infection with SARS-CoV-2.” That could mean they think that a researcher collected the SARS-CoV-2 in the field and become infected by it. It could also mean that they think SARS-CoV-2 was genetically engineered in a lab and then transmitted to humans. Those are rather different scenarios.
If you are the US president, you can perhaps trust what the intelligence services will tell you.
If you are the public, you can not trust them at all.
FFS, we are talking about groups who have worked hard to stage coups and went out of their way to torture prisoners in gitmo. Psyops are one of their tools. Do you really believe they would draw the line at lying to the public?
Before the NATO warning about Putin became true, my general stance was 'the tell for lying in public statements from the intelligence community is that their lips are moving'. I still refuse to update on them, even if they tell the truth, they don't want you to believe that because that because it is true (simulacrum level 1) but because it is advantageous for them if you believe it (level 2).
Yes, I am old enough to remember Powell trying to convince the UN security council with his 'intelligence' about Saddam's weapons of mass destruction.
Wouldn't one simple reason be that US is in a cold war with China, and in such a situation intelligence agencies might be predisposed to prod people towards explanations that place China in a worse light than other poential explanations?
Weren't there several intelligence agencies that reported their overall conclusion was moderate support one way, and several that reported their overall conclusion was moderate support the other way, and none that reported their overall conclusion was strong support for intelligent design?
While it seems reasonable to assume government agencies might have better evidence than the public, what's actually happened on this topic has been people in the government pushing stories that end up being false or exaggerated.
These have included claims such as scientists at the lab being the real first cluster, that SARS2 was part of a bioweapons program, that there's evidence a safety incident happened at the lab in late 2019, that a vaccine scientist was executed by being thrown off the roof of the lab, that the virus was already spreading at the Wuhan Military Games. All of these are from US government sources and all of them appear to be false, exaggerated or lacking evidence. Exactly what leads high-level national security officials to push things like this without good evidence isn't totally clear -- maybe it's the intersection of US-China geopolitical rivalry with a complex, unknown topic -- but all of this suggests the FBI and DOE's conclusions might need to be taken with a grain of salt as well.
We also know that neither of their assessments changed the views of any other agencies, which seems surprising if they really had strong evidence. IMO it's more likely their assessments simply represent the view of particular analysts at these agencies, who came up with their own theories about what happened based on limited evidence like everyone else.
Am I the only one who Peter lost at the end when he starts asserting that any evidence that disagrees with him comes from liars? And his "evidence" that people are liars is stuff like "it was reported in the Daily Mail" and "he was a drug user" and "his cat got it too" (by which Reed obviously meant tested positive, which can absolutely happen to animals, inanimate objects, and all kinds of other random things). Followed by, "and everyone else was also a liar".
I have a friend who is convinced he caught COVID in Jan 2020 before anyone knew what it was or what to watch out for. He had symptoms that were an exact match. Officially, COVID did not exist here at that time. Is he also a liar? I assert that he is not - he has no reason to lie about this, and is an honest and upstanding gentleman. I therefore assign it a 1:10,000 chance that he is being deceptive, and thus COVID clearly spread far earlier than officially stated.
I also have a heuristic about people who dismiss evidence on the basis that it appeared in a tabloid, or wasn't in a peer reviewed papers: naive and gullible. Here's some Bayes for you: what's the probability that a journal publishes an article full of photoshopped images and garbled machine-generated text? My prior for this happening in an academic peer-reviewed journal is high, because there keep being incidents where whole editions of journals get retracted due to every article being gibberish. My prior for it happening in the Daily Mail or any tabloid is zero because it's never happened. Clearly, the Mail is far more reliable than peer reviewed journals and Peter should accept that its evidence overrules anything he's read in journals.
Rephrased: most of Peter's evidence implicitly assumes the testimony of China, virologists and epidemiologists is reliable (e.g. the ridiculously precise stated doubling period), when we know it's not. Many times he makes statements to which a good response might be, "and how do you know they aren't simply wrong/lying", but Saar was too generous to do that. Peter did not return that generosity!
Ultimately, although it was interesting and Peter's performance was impressive, I don't know if this kind of debate helps. I remain in camp Lab Leak.
> most of Peter's evidence implicitly assumes the testimony of China, virologists and epidemiologists is reliable (e.g. the ridiculously precise stated doubling period), when we know it's not. Many times he makes statements to which a good response might be, "and how do you know they aren't simply wrong/lying”
They were lying for about a month. That’s why Li Wenliang was arrested for spreading malicious rumours. However he was the one who first mentioned the market.
Yes, I don't understand this (paraphrased) claim by Peter: (EDIT: this seems to be Scott making a mistake in paraphrasing Peter)
> He also told the Mail that his cat got the coronavirus too, which is impossible.
'Impossible', thus implying the man was lying? I was under the impression that, quite aside from cats having tons of coronaviruses in general (FCoV being a particularly serious threat to young cats, which also seems to be a remarkable case study of the harms of the FDA), that it was not just not 'impossible' for domestic pet cats to get the coronavirus too, it was routine for them to get COVID-19, and even other cat species in *zoos* have tested positive and this was true very early in the COVID-19 pandemic and quite well publicized and well known (eg April 2020 https://www.nationalgeographic.com/animals/article/tiger-coronavirus-covid19-positive-test-bronx-zoo ). This was a topic of interest to me at the time because I like cats and have a cat and was wondering what the implications of me being inevitably infected might be for my cat, and so I remember this quite well despite my general attempt to remain ignorant of as many COVID-19 matters as possible... And double-checking now to see if all of these reports were somehow false positives or faked, I continue to see everyone like the CDC stating that it is still totally possible and routine for cats in close contact with infected humans (you know, like a *pet* cat) to be infected with COVID-19: https://www.cdc.gov/healthypets/covid-19/pets.html
Given that Peter has supposedly spent years autistically researching every last detail and this detail in particular in order to discredit that British dude, I'm experiencing sudden Gell-Man Amnesia here about the rest of his claims, as well as the supposed experts evaluating Peter's claims if they didn't flag that (I have not checked).
Yeah, I found this comment from Peter on the Manifold Market a bit concerning:
> So what's my incentive to come up with the most reasonable values possible, if my opponent just says any data I present is fake/meaningless/actually supports his theory via some twisted argument?
> By the end of the debate, it was more like, "fuck it, if he's just going to make up numbers, I'm going to stick with largest numbers for my case".
The context for that is Saar asserting that the market data carries 0 weight as evidence for Zoonosis. Peter is saying that he hewed to an argument which was *clearly* unreasonable instead of contesting numbers which Peter thought were reasonably subject to debate.
>Saar is arguing honestly and presenting his actual views.
>Peter is dishonestly making the strongest case he think he can for his side, which includes deliberately saying incorrect and misleading things.
This makes sense when you remember that one of the debaters is just some guy and the other one is very wealthy. If 100k is a very meaningful part of your net worth, you can't afford to not play to win.
"making the strongest case he can for his side"- isn't that how a debate is supposed to work? It's certainly how the legal system works. It's a competitive process, you're not supposed to help out the other side. The *judges* are the ones who are supposed to be neutral and get at the truth of the matter.
During the debate, Peter also cited a paper showing that the first Beijing superspreader event (another wet market, what are the odds?) had originated in frozen fish.
"We know why there was an outbreak in the Xinfadi Market in Beijing: it was because the seafood stall got frozen fish from some non-Zero-COVID country, the fish had COVID particles on it, and the vendor got infected and spread it to everyone else. Something like this is true for the other Chinese wet market based outbreaks we know about it."
I think the dispute may arise because COVID is not a logically defined disease. Early on it was, back in 2019 before the PCR test for it was developed and diagnosis was handled by doctors, but once public health officials got involved the definition changed to become circular: COVID is defined as having a positive test, a positive test is defined as having COVID.
This leads to all kinds of stupid stuff like the way governments insisted that COVID testing had no false positives whatsoever, even when that clearly wasn't true. Except it kinda was true given their violent abuses of logic - if the test is positive then by definition you did have COVID, even if you never got sick.
Due to this it's possible for the cat to both "have COVID" and also "not have COVID" simultaneously. We might call it COVIdinger's Cat. The cat cannot actually be infected by the virus or get sick under any conventional understanding of disease, yet it can test positive anyway simply due to having some human-generated virus particles somewhere in its body. The PCR tests don't care and will detect tiny quantities of fully destroyed virus, which is why they could trigger in people who'd been sick and recovered weeks earlier. And because COVID is defined by the presence of a test, that's enough: the cat has COVID!
This doesn't justify Peter's claim. Both of my links confirm that cats do exhibit physical symptoms and disease traits like 'dry coughs' and 'loss of appetite', among many other symptoms like 'vomiting' or 'diarrhea', and in the case of the zoo cats, those symptoms were why they were tested at all for COVID-19 and the blood samples could then test positive for COVID-19 in addition to them being sick. Thus, it is clearly possible, and not 'impossible' for a cat to have COVID-19 beyond a circular test result, and it is established fact according to the CDC, Bronx Zoo and every other source I have checked that cats really can 'have COVID' under any 'conventional understanding of disease' and do not simply merely carry it while testing positive (which a lot of *other* animals do - just not domestic cats and a bunch of their felid relatives).
I think it's an error with Scott's paraphrasing -- Connor said that has cat died, not that it got sick. Here's his quote:
"Pictured: Connor and the cat that hangs around his apartment in Wuhan, China. Suddenly, I’m feeling better, physically at least. The flu has lifted. But the poor kitten has died. I don’t know whether it had what I’ve got, or whether cats can even get human flu. I feel miserable"
I am well aware that cat's can catch the coronavirus (though my cat never got visibly ill any of the 4 times covid's been in my household). But the prior on an pet cat dying from it should be low.
Another commenter above suggested that "maybe the cat just tested positive". This is late November or early December 2019, before covid was known to exist and before covid tests were invented. The only way that Connor himself would even know if he had it is if he went to a hospital and they saved a sample and then the hospital tested it later. Connor claims that the hospital called him weeks/months later and told him he tested positive, but there's no other record of that. Either way, they most definitely did not save a swab from his cat and PCR test that.
The prior on a 25 year old going to the hospital from covid should also be pretty low. This also can't be a sampling bias thing, given the predicted number of covid patients at that time, there aren't going to be many and even less are going to speak English -- even in the case growth model that Rootclaim tried to present, I calculated there would be at most 1 at that time)
The last interesting thing is that Connor says that he shopped at the market!
Specifically, he wrote:
"My local paper back in Llandudno, North Wales, has been in touch with me. Maybe I caught the coronavirus at the fish market.
It’s a great place to get food on a budget, a part of the real Wuhan that ordinary Chinese people use every day, and I regularly do my shopping there.
Since the outbreak became international news, I’ve seen hysterical reports (especially in the U.S. media) that exotic meats such as bat and even koala are on sale at the fish market. I’ve never seen that."
So, that's not 100% explicit as to referring to Huanan market, but Huanan market was the one in the news, and there was a specific rumor about Koalas being sold at Huanan market, so it seems pretty clear that's the market he's referring to.
So, either:
A. He's lying about the whole story and just fabricating it from stuff he's seen in the news.
B. He's a super early patient who got infected at Huanan market, and he's great evidence for the zoonosis theory.
C. His story is true except that he actually shopped at a different market other than Huanan, and that makes him the super secret early case that proves lab leak.
I interpret this as most likely A. Rootclaim sees C as most likely.
In the third debate, I even tried to lay out a quick bayesian analysis for how to decide between these options. I think that's sort of the problem with Rootclaim's method -- they decided from a high level view that they think that lab leak is true, then they went looking for evidence to confirm that view, with no good filter on what evidence should be trusted, i.e. they decided that the cases in the WHO report are 100% fraudulent but Connor Reed's case is 100% likely to be covid. It all just seems like motivated reasoning, to me.
> I think it's an error with Scott's paraphrasing -- Connor said that has cat died, not that it got sick. Here's his quote:
This isn't about Connor, this is about Peter. Did Peter claim what Scott paraphrases him as claiming, and apparently no one in the debate called him on it? (I'd just check myself, but I've now searched the YT auto-transcript of like 6 or 7 of these YT videos and haven't been able to find a quote matching it; and I'm definitely not watching the whole thing, so hopefully someone who *does* know where in this rats-nest of links Peter would have said it, can go rewatch it and tell us what he said.)
I'm Peter. I claimed that it was an untrustworthy account because the cat died, not because the cat got sick (as well as all the other issues I listed above).
IIRC Saar later made some counter claims that Connor shopped at a different market and that the cat died of some feline coronavirus. None of which is supported by the original articles, just a purely motivated reading, I think.
Off topic, but I find myself less interested in the epistemology behind COVID's origins (which is still quite fascinating) and more interested in what you're planning to do with your research and persuasion superpower. There are lot of people who capable of being intellectually rigorous but not persuasive verbally. And also vice versa. To have both is quite rare. Most traditional careers after a physics degree seem like a waste.
I certainly intend to research other things in the future, and I hope to be done with this topic soon.
I have never thought of myself as persuasive at all, have never debated before or given many presentations or speeches. It just seemed like it couldn't be that hard -- Rootclaim's case is so thoroughly wrong -- so I went for it. But mostly it was stressful and not very fun, so I'll probably stick to writing, coding, researching, and things like that.
> I claimed that it was an untrustworthy account because the cat died, not because the cat got sick (as well as all the other issues I listed above)...None of which is supported by the original articles, just a purely motivated reading, I think.
We are going off on a tangent here from my original question about the paraphrase, but now that I check... how is that not supported, and how is your description supported by the original articles, exactly? The source here seems to be the Daily Mail account https://www.dailymail.co.uk/news/article-8075633/First-British-victim-25-describes-coronavirus.html which includes photos of the cat, which is clearly a kitten (an important detail to omit), and all of the references to the cat are:
>> Day 9: Even the kitten hanging around my apartment seems to be feeling under the weather. It isn’t its usual lively self, and when I put down food it doesn’t want to eat. I don’t blame it – I’ve lost my appetite too...Day 11: Suddenly, I’m feeling better, physically at least. The flu has lifted. But the poor kitten has died. I don’t know whether it had what I’ve got, or whether cats can even get human flu. I feel miserable.
I do not understand how you go from his claim "I don’t know whether it had what I’ve got, or whether cats can even get human flu. " to your slide https://www.youtube.com/watch?v=Y1vaooTKHCM#t=1h10m30s saying "Claims his cat died from covid" (which you also repeat aloud to emphasize it). You then repeated it again in the slides & aloud (without any additional source so presumably still the Daily Mail source) at https://www.youtube.com/watch?v=CTK1hWOzwns&t=1510s as "He says his cat died from covid...this is what I read in the article". Are we looking at the same 'article'? Because I don't see how you could have read that in the Daily Mail article given the only references to the cat in it which I quote above.
Connor explicitly doesn't claim the kitten got what he got (much less got COVID-19), so a claim that he didn't make cannot make him 'untrustworthy'.
Nor is the cat dying anything to make him 'untrustworthy': it is entirely ordinary for kittens to die of respiratory illnesses like various coronaviruses, to which they are especially vulnerable. (Lots of cat diseases to go around, unfortunately.) This is a major source of kitten mortality in litters, pounds (so notorious as to have their own names like 'kennel flu'), and cat colonies, or cats in general, and there is nothing all that odd about it happening or a kitten dying of it; and it dying is that much more likely if not treated (which it apparently wasn't inasmuch as Connor was busy being sick and being an expat probably couldn't call on anyone for a local vet). So it should be no surprise that kittens die of COVID-19: https://bvajournals.onlinelibrary.wiley.com/doi/full/10.1002/vetr.247 Anyway, young kittens die like he describes of FCoV *all the time* (you may have heard about the Cyprus outbreak last year), so there's nothing odd about the death itself.
Both of those articles say Connor was hospitalized for two weeks, vs previous ones saying he was sick for 2 weeks and then went to a hospital briefly.
The first article also goes on to quote him saying, "I honestly don't think the virus is as serious as all that. The people who have died must have had underlying health problems", which is a weird description for hospitalized for 2 weeks.
I suppose you can't entirely determine whether Connor's story is inconsistent or this is bad reporting.
Same with "the feline coronavirus". How would he know what the cat had? He says he was housebound, wasn't going to the vet. I'm not sure the usual timeline for FIPV. In the Dailymail, he's asking if the cat got the "flu" he had. Then in the Wales online article, he's saying he's not sure if he caught it from the cat. But if he's certain it was "the feline coronavirus", that doesn't even infect people, right?
The path to this all being important requires both that "Connor is trustworthy" and also "Connor is lying about shopping at Huanan" (or else I misinterpreted which market he was talking about) and "We've settled on a specific growth rate for the market outbreak" for which this evidence would contradict and it's somewhat dependent on "Connor got covid on Nov 25th, not during his second wave of illness in early Dec".
I think I'm very likely to win either "Connor is untrustworthy", with "Connor shopped at Huanan" as a backup.
With a certain motivated reasoning, though, you can conclude that Connor is great evidence to contradict zoonosis at the market, and therefore "SMOKING GUN FOR LAB LEAK!"
I also think I have a strong enough model, between case growth rates, genetic diversity, Wuhan serology, etc, to simply put low prior odds on dubious covid infections in November as being likely to be true.
Likewise, if Connor really got sick in November, where are the interviews with 100+ Chinese people that also did, that you would statistically expect? There are interviews available with some of the earliest people sick at the market. Did China somehow retroactively censor every other similar account from before?
It's the same thing with e.g. positive covid tests in Italy in September. I did a laborious deep dive through all those papers just in case Saar brought them up, but you can also determine they're unlikely to be valid if you have a good enough understanding of the other evidence.
In a larger sense, I think there's a certain personality type that gets drawn to the patterns and another that gets fixated on outliers. This dynamic even created a new theory, at the debate:
I wrote some code to model and graph mutations that were reversions. There were 2 dots out of place in my graph. I sort of noticed when making it, but it didn't really register in my brain as that important.
Saar noticed the dots later and concluded they were "SMOKING GUN FOR LAB LEAK!" and they coincided with SMOKING GUN DELETED EARLY CASES and SMOKING GUN ANTARCTICA SOIL SAMPLES (which both also turn out to be misleading and unimportant claims under further investigation).
So I dug into the code and the data to figure out why my diagram didn't match Pekar's, and found it was a mundane explanation for the 2 dots and corrected the diagram.
Once you've been through this process dozens of times with the lab leak crew, you kind of develop an understanding that they're just throwing out one bad claim after another, and those rarely ever hold up.
I think this is actually a great heuristic, that's better than bayesian reasoning, for spotting conspiracy theories.
One of the first places I might have developed this approach was in 2020, when I was trying to figure out if the election was stolen, i.e. for curiosity and to know who to bet on in the political betting markets.
The 2020 election fraud people made a lot of claims that the election was stolen. People who doubted it asked "where is the evidence?"
So someone made a site called "here is the evidence"
It looks like it currently has 1,687 claims as to why the 2020 election was stolen.
My process, back in 2020, was to go through maybe 20-30 election fraud claims and evaluate them. Every single one I evaluated was either clearly false or just ambiguous and misleading. So I concluded that I did not have to read a list of 1,000 and I bet that Biden would still win (I forget the rules of the market, perhaps I was still taking a risk on a successful coup).
It's kind of similar with lab leak -- I've evaluated hundreds of claims now and found the vast majority either blatantly or subtly wrong. Had any of them been convincing, I would have dropped out prior to the debate and moderated my stance on covid origins, maybe even flipped it. Saar credulously relied on many of them in the debate, many more transparently false than Connor Reed's covid case.
From Saar's probabilistic perspective, it's more like: the details for Connor don't even matter. Even if Connor didn't have covid, someone else did, and if he dug hard enough, he's sure he could find that person. And because he thinks that, you can't apply high odds to the market outbreak. In fact, he thinks you can't apply any odds at all to it, he thinks all existing case data is bayes factor 1 between market and lab.
I think Saar has concluded that lab leak is almost certainly true based on the location in Wuhan and the furin cleavage site. Therefore, lots of non-market early cases must exist somewhere, and he doesn't even have to look.
If someone made "here is the lab leak evidence dot com", with 1,700 different examples of early cases, I would go through 100 and notice that 95 are clearly wrong and 5 are ambiguous and whoever curated the site is probably just not very careful.
Saar would look at the list and say, "wow, that's 1,700 different claims. If even one of them is correct, then it's surely a lab leak! What are the odds that 1,700 claims are all wrong?"
I could be wrong about all this, of course. If hospital records for Connor Reed ever turn up, I will apologize, change my mind, and evaluate whether that changes the origins picture based on those details.
At 57 minutes, Saar infers that it was not Huanan market.
It is ambiguous, Connor just says "the fish market". I'm inferring that it's Huanan market because he says it's in the news and the thing about koalas. Saar is inferring that it's some other market, but I don't see why.
On the next slides Saar says it was "the feline coronavirus". That appears to be completely made up, if you compare it to the actual quote I provided above.
Saar then claims that there's evidence Connor went to Zhongnan hospital. There's no evidence of that, as far as I know, other than Connor's claim. And I'm not sure if Connor ever even named which hospital. There's no official record of the hospital contacting him and telling him he got sick.
And then here's where we discussed it at the 3rd debate:
Thanks for the links. I did see those in the transcripts but I wasn't sure if they were the key ones or if the auto-transcripts had errors. I don't know what the 'fish market' or 'Zhonang hospital' have to do with anything I've asked, but moving on, I can't see where anyone in those 3 videos claims that "cats can't get COVID-19", and you do seem to acknowledge that they can in the third video when discussing how he could've tailored his story post hoc in 2020 (which would be a strange claim for you to bring up had you ever said anywhere in the debates something like "cats can't get COVID-19" like Scott's paraphrase).
So I think Scott Alexander needs to correct his paraphrase in OP: you did not say "cats can't get COVID-19", you seem to instead be saying that "cats never (ie. rarely) die of COVID-19" (which may sound similar but of course are quite different).
This doesn't make sense to me - how did Connor know what disease his cat had? He said he was housebound that whole period, so he couldn't have gone to the vet. Also, it seems to contradict his previous story about thinking it was COVID-19.
If nothing else this quote should throw Connor's credibility into question slightly: "I think to be fair the Chinese government have been quite transparent with what's going on."
You're correct that cats are susceptible to SARS-CoV-2 -- there's even a serological study of cats in Wuhan at the time of the outbreak that found they were infected. https://pubmed.ncbi.nlm.nih.gov/32867625/
However IMO there are a number of other reasons to question the significance of Connor Reed's story. For one thing, there's no actual verification that he was sick with COVID during that timeframe in the first place, which is why he was not included in the official cases in spite of all these media stories. Other people have claimed to be unconfirmed early cases such as athletes attending the Wuhan Military Games in October, but these are generally believed to be cases of misidentification.
In his story Connor Reed says he might have been infected at "the fish market" that was written about in the Western media, where bats and koalas were allegedly sold. This only describes one market in Wuhan, the Huanan Seafood Market, where news sites incorrectly reported the sale of koalas (it ended up being a mistranslation of another species).
So if we take him at his word, Connor Reed may actually be backing up the market origin, although it's strange that he seems to not know the name of the market itself. He apparently lives some distance from the Huanan Market. All the details in his story could be picked out of media reports.
There was a wave of influenza going through Wuhan at the time so it seems possible that he just had a bad case of flu and confused it with COVID-19 (although he claims to have been tested). It would be interesting to follow up with Reed's friends and contacts in Wuhan at the time and ask if they know exactly what happened.
Doesn’t a lot of Saar’s argument also rest on evidence that would be impacted by Chinese meddling?
The correct response to “Chinese meddling in the data” is to reduce the likelihood that we ever “know” the correct answer, rather than to just say “therefore Peter wrong”.
Recall that a wet market origin is also very embarrassing for China. The official CCP propaganda is that it came from some foreign source.
I think the primary evidence Saar has for the lab leak is independent of China, namely the DEFUSE grant, the furin cleavage site and the improbability of a coronavirus lab being just down the road from a global coronavirus outbreak by coincidence.
Which of Peter’s arguments on the Turin cleavage site rest on “China being truthful”? The participants largely seem to agree on what the site *is* but disagree on whether or not the site is more likely to be of natural evolution or a man-made insertion.
For DEFUSE, yes to some extent Peter is taking WIV at their word. But he also points out several reasons why it’s actually likely that they weren’t doing the right type of experiments on the right strain to cause the epidemic. And why if they were doing those experiments, it’s likely they would have said so (since before the epidemic they had no reason to cover up).
Meanwhile Saar’s theory *depends on* WIV / CCP lying about what they did after not getting the DEFUSE grant. And he doesn’t actually have any proof that they did, he’s just inferring that they did from his conclusion that the furin cleavage sites were probably man-made. His argument also relies on either WIV/CCP pre-emptively covering up their activities that eventually caused COVID (while being open about their other research and desire to be involved in gain-of-function studies) OR this research occurring at exactly the right time that it had gotten done but not published before the virus leaked and they went into cover-up mode.
Conditional on this being an unintentional lab leak, yes it’s very likely it would occur in Wuhan vs. somewhere else. Conditional on it being zoonotic, there’s not really any reason it would be in Wuhan vs some other Chinese city (but no particular reason it wouldn’t be in Wuhan). How much you weight you put on that “coincidence” will largely be swamped by how likely you consider zoonotic transmission and lab leaks in the first place (which is really hard to sus out, since both are rare).
We can be reasonably sure the WIV/CCP are lying about something, because they have stonewalled or misdirected all attempts at investigation, taken virus databases offline in the middle of the night, disappeared people involved and done various other highly suspicious things, some of which are known only via intel leaks from western intelligence agencies.
The debate here didn't touch on most of the evidence, but for example the US Government has claimed that they intercepted messages in November talking about a very severe situation at the WIV, and there is evidence they was a sharp uptick in activity there (hacked mobile phone records), and that they were rapidly acquiring "air incinerators" and other things. Air incinerators would imply they suspected or found a leak in the lab's air ducts or air handling equipment. It's been a while since I last looked at this but the weight of evidence for something having gone wrong at the lab is very high, unless you assume the leaks are all invalid (which would lead to a question of motive, as the establishment is very keen on zoonosis as a theory).
I do not buy the idea that it's unlikely they were doing such experiments:
1. It's very likely that the PLA were involved or at least monitoring the research going on there. All this stuff is highly tied up with military research, hence why DEFUSE was submitted to DARPA in the first place. Thus we can't assume everything would have been made public.
2. Peter argues the WIV would have been totally open and transparent before COVID, even though they became immediately 100% opaque afterwards and that has never changed since. If they're so innocent, why the sudden switch? More plausible: they have something to hide and thus were plausibly also previously hiding something.
3. I don't think WIV/CCP have ever commented on the DEFUSE grant or whether the WIV did or did not do the suggested research, did they?
4. The research occurring at the time of a leak and before it got published, assuming it was destined to be published at all, seems very plausible, no?
5. The "FCS is natural" argument makes some bold assumptions about the nature of whatever was being done, why and what could or could not have happened in a lab. I don't buy it. We know virologists love to play games with accelerating evolution, so what a geneticist might or might not have done to add an FCS doesn't seem very relevant to what WIV researchers would have done.
6. Lab leaks aren't particularly rare. They happen quite frequently, especially if you consider accidents that are bad but don't manage to start pandemics.
Again, your entire argument here is that there is a cover-up that is effective enough to make zoonotic origin plausible enough to convince a lot of smart people who’ve looked deeply into it, but not so good that Saar et al couldn’t prove it to be false. That’s a fine line! Meanwhile you expect Peter to prove a negative (CCP didn’t lie in precisely that way).
I would also reiterate: a wet market outbreak still looks very bad for China, and we can see that the Party line is that *neither* theory is true. Why would the CCP stage an elaborate cover up that still paints them in a bad light, vs. cooking up some evidence that it was from a Sri Lankan tourist or whatever?
I don’t have infinite time but I’ll address two points where I think you’re being uncharitable to Peter’s argument.
2. Peter is not saying the WIV is “perfectly transparent”. He is however saying they were open about their research, and until the leak of Covid would have had no reason to cover up specifically the line of research and exact strain that would later lead to an outbreak, because they would have had no way to know beforehand that that strain and that research was the one that would leak.
Unless you believe it was a deliberate bioweapon, in which case you’d sort of expect them to be completely opaque rather than “very open and then opaque only after a possible screw up”.
The WIV going full CYA mode strikes me as likely whether or not they were at fault. There was uncertainty, the people involved are paranoid, and they’d know they were obvious suspects.
5. I agree that it’s possible a human scientist *could have* made the FCS as we see it. However, Saar’s initial position was that it was *basically impossible* for the FCS to be natural. When Peter gave several reasons why that type of FCS could occur naturally, Saar shifted to “well okay maybe the researchers were deliberately trying to mimic or boost natural evolution”. That strikes me as dishonest argumentation and kind of conspiratorial - it’s the exact type of epicycle conspiracy theorists use to dismiss contrary evidence.
Has it convinced a lot of smart people? I have to admit, my biggest surprise when I started reading this article was that there was someone smart who was willing to argue against the lab leak. I know a lot of smart people and not one believes in it was a natural event. And as Scott shows at the start, the vast majority of all people don't believe that either. Peter is certainly being very counter-cultural. And if you want to argue that "smart people" are in the minority, well that's an appeal to authority that academics no longer have.
Putting that aside, I don't think Saar retreated at all. Let's assume both argue in good faith, and I think Peter did even though I think he's wrong. Remember that language is very imprecise. This imprecision has already created a ton of confusion on this topic for several years already. If the WIV do something to speed up the natural rate of evolution, is that "natural"? Is it "engineering"? That's entirely subjective and you could easily argue for both words. Peter here starts by assuming that not natural means genetically engineered and then starts arguing against it being genetically engineered, which isn't a correct assumption given the nature of the WIV's investigation. They were interested in what a coronavirus pandemic might look like. They wanted to look into the future (at least, officially). Just directly engineering an FCS into a pre-existing virus can't possibly tell them that, so why would they have done so? Rapid forced evolution in humanized organisms on the other hand are a much closer way to "speed up time", and could easily have led to the evolution of a strange FCS under lab conditions. If they didn't expect it to happen because it's never known to have happened in nature before then they would not have been taking the necessary precautions, as indeed we know they were not. And bang, that's all it takes to start a pandemic for real.
So Saar didn't shift at all, in my perspective. He corrected a misunderstanding by Peter of what was being actually argued, an honest misunderstanding coming from the vagueness of terms like "natural" when talking about serial passaging experiments. But it does invalidate most of Peter's arguments about the exact genetic code of the FCS.
You are missing some important context which is that many of the allegations you mentioned (about a "severe situation" at WIV for instance) seem to be exaggerated or made up.
The "severe situation" story by Katherine Eban and Propublica seems to be based on a bad translation, with China experts questioning it and Propublica eventually issuing an update on the article itself. And there are a number of stories about the lab like this that end up being the result of bad reporting, misinterpreted evidence etc.
When you look into who is behind these stories many of them seem to be coming from the Trump State Department under Mike Pompeo, which conducted an investigation into the lab back in 2020. "State Department investigators" is the source given for many of the key lab leak stories, such as the claims about sick scientists with COVID symptoms published by Michael Shellenberger and the Wall Street Journal or the Sunday Times story including the claim that a Chinese scientist might have been executed by being thrown off the roof of the lab.
There's actually an account by another State Department official who was involved in the investigation about what happened:
He says he was open to the lab leak but that certain officials like David Asher and Miles Yu became very fanatical about it. According to him they were already believing unhinged conspiracy theories like the virus being a racially targeted bioweapon and wanted to charge China under the Biological Weapons Convention, which he thought was a bad idea. Eventually they ended up accusing him of being part of the conspiracy and trying to cover up their work. The same people, along with Trump, Pompeo and Richard Dearlove, the former head of MI6, appeared in Sherri Markson's documentary "What Really Happened in Wuhan."
There's some interesting background to all this which is that both Miles Yu and Mike Pompeo have connections to the New Federal State of China, a weird regime change group that says it wants to overthrow the Chinese government. This group was the first to promote the lab leak theory in a major way starting all the way back in mid January 2020.
Mike Pompeo received an award from the group and Miles Yu has spoken glowingly of its founder Guo Wengui, who is now facing trial for defrauding the overseas Chinese community out of $1 billion, in speeches he's given. So it seems quite possible this group and others in the anti-CCP Chinese diaspora are influencing the lab leak narrative to some extent.
Only problem is that when the ODNI issued a report on the lab last year, it seemed to greatly downplay the significance of many of these allegations. Turns out there was no specific biosafety incident known about, the people at the lab who got sick probably just had normal seasonal illness and there was no indication they were the ones working on coronavirus research (discussed towards the end of this article).
All of this therefore needs to obviously be taken with a lot of skepticism. There are geopolitical motives, propaganda and BS stories on both sides unfortunately, not just from the Chinese government.
You have arbitrarily excluded the third possibility that your friend is simply mistaken. His conclusion remains extremely unlikely regardless of (your judgment of) his personal honesty.
I was being facetious. If Peter is allowed to decide large numbers of random people are all wrong and liars to boot, I'm allowed to assume my friend is correct. You might claim that no, the other evidence means I shouldn't but that's circular
As I've been saying since the whole debate started, whether or not it was a lab leak is really irrelevant. What we know for certain, with no actual controversy surrounding it, is far more interesting: China did not shut down international air travel out of Wuhan until significantly after they knew the virus was circulating.
While they were covering up the origins of the virus, and locking down Wuhan to try to keep it from spreading further internally, the CCP was knowingly, deliberately exporting this deadly plague to the rest of the world. And that is far more damning than any question of its origin. The origin of the virus is China. It deliberately inflicted it upon the rest of the world, which makes the pandemic a *de facto* bioweapon attack, whether or not the virus was developed in a lab with the intention of building a bioweapon.
Closing down airspace was considered a pretty radical thing to do back then - it’s only in retrospect that it seems obvious. Had China stopped Americans leaving, for instance, there would have been an outcry.
And of course flights from China could have been banned at entry, but the west was clowning around with either calling that racist or denying that that whole thing was fake - which I believe was the position of Fox News and Trump for months.
I have a memory of Trump barring travel from china fairly early, while also saying it's probably fake, while also calling it the China Virus, and while also saying that it'll blow over soon. This is consistent with my model of Trump saying lots of things for various reasons, none of which are primarily anything like "provide parsimoniously accurate information". I'm open to seeing a well-cited timeline of his actions and statements, but I don't think any of us care enough to create or find such a thing.
The Chinese travel ban occurred in late January of 2020 after an increasing number of cases were discovered in the US. In mid march San Franscisco and the Bay area imposed the first shelter in place order (although behaviors were changing in the bay area for the prior two weeks and on March 20th NY declared a stay at home order (effective march 22) https://en.wikipedia.org/wiki/Timeline_of_the_COVID-19_pandemic_in_the_United_States_(2020). One thing is that the behavior in early 2020 involved a series of relatively (within the lifetime of most people) unprecedented steps and while it is easy in hindsight to argue about what should have been done, neither China nor the US really understood the consequences and in particular were constantly make decisions that were days to weeks behind where they should have been.
Yeah I personally have no blame for any decision-makers. That itself seems an unpopular opinion, but it's just incentives playing out. What's the age of the electorate? How scared of dying are first-world elderly? Boom, tells you everything you need to know.
I agree that it's unimportant per se. I'd argue that the other specifics throughout were also unimportant: masks, vaccine efficacy, vaccine safety, lockdown efficacy, ivermectin efficacy, vitamin d efficacy, etc. None of the mechanisms seemed to matter. What mattered more was "are people lying to us?" and the answer to that was "yes, unambiguously" (on both sides of each and every one of these debates).
The consensus prior to 2020 was that shutting borders would only serve to delay respiratory disease outbreaks by ~2-4 weeks because, under exponential growth, blocking 99% or even 99.99% of travel still means its not long until covid hops the border. For China to have acted maliciously in not closing the border would require that 2021 values and beliefs travelled back in time to 2019.
Locking down internal travel is just something communist regimes do. In China it is part of the Hukou system. In these sorts of regimes, internal travel restrictions are historically the norm, and permitting internal travel is the aberration. Therefore China reinforcing internal travel restrictions is far less interesting than it sounds.
They should have locked down the whole country at that time but that would have been too extreme in 2019 so they did an extreme but less extreme thing of locking down the city but of course it had already spread to all of China by that point so the lockdown was only moderately effective. Again everyone was learning and reacting in early 2019/2020 and typically they were making decisions that were weeks late due to not thinking about exponential curves.
If SARS-Cov-2 did not come from WIV then presumably WIV and the Chinese government would have had a very powerful motivation to open their records to outside investigators and make all of their people available for interviews. Instead they clammed up. Why?
The problem is that the Chinese government is also unhappy with zoonosis, because it's still somehow embarrassing. The narrative pushed in China now seems to be that the virus did not originate in China but was brought in to Wuhan in frozen food. This actually happened several times after China's zero-covid strategy started so it's a convenient theory to push the blame on.
Ok, but that still leaves them with a strong incentive to open the WIV records if they are consistent with the argument that WIV was not the origin of the virus.
At this point in time, anyway, would anyone believe opened WIV records that apparently exonerated them? The CCP would have had ample time to cook the books. ( I doubt that I would trust _any_ government's claims in an analogous situation. )
What do you mean, "somehow"? People have constantly pointed out that wet markets were a blatant health risk, and yet the CCP never never took action against them. If the wet markets were entirely responsible for COVID, the CCP would indirectly be responsible for it through their inaction. That would be bad for the CCP.
All the usual reasons that most virology institutes don't generally open up their records to outside investigators. Plus possibly whatever not-totally-kosher things they might have been doing there.
These kinds of arguments were used to devastating effect in the run up to the 2003 Iraq War. Saddam Hussein refused to allow inspectors to see his chemical/biological/nuclear weapons facilities. The only plausible reason for this was that he had secret weapons that he wanted to hide from the West.
As it turned out, he did not have secret weapons.
Instead, he wanted to appear like he had secret weapons to scare off Iran, a large hostile neighbor which he had fought with repeatedly in the past. A totally rational reason to want to appear to have secret weapons of mass destruction which did not occur to the American government before they made the fateful decision to invade.
Trying to ascertain the intentions of your enemies is something humans are particularly bad at. I wouldn't use it as a basis for making decisions, if it can be avoided.
...if you could avoid ascerting the intentions of other people (not only your enemies) in your everyday life (rather than only believing you do not have to do this), you could not walk down any street without being on constant alert and armed to the teeth.
...Because then you could not ascertain if the passers-by are unlikely to have the intention to suddenly and unprovokedly attack you.
Ascerting the intentions of other people, at least in the limited sense of being reasonably sure they have not the intention to suddenly attack you, is nothing less than THE fundamental feature that makes in possible for humans to live in societies consisting also of non-kin. It is the founding stone of sociology.
The Chinese government's default reaction is to cover up anything remotely controversial, even for stupid or irrational reasons, so failing to open one of their high-profile labs to foreign investigators is not really that surprising. They did allow the WHO to go there and interview the scientists for several hours about their work, and Zhengli Shi gave an interview to Science where she answered questions about specific lab leak allegations. From China's POV this probably represents more openness than usual (by comparison, the WHO did not get the chance to talk to animal traders at the market).
But ultimately the CCP cares much more about domestic public opinion than international. And within China itself, the widespread belief is that the virus came from overseas, possibly from a US lab.
I hadn't heard of Rootclaim before this, but that first screenshot which puts the odds of widespread fraud in the 2020 election at 8.7% tells me all I need to know -- that is orders of magnitude too high as anyone who understands how us elections are run would understand. Seems like there's something about this method, at least as practiced by Saar, that produces excessive probabilities of out of consensus or malicious explanations for things.
This is a known bias we have. We have several safeguards against reaching extreme probabilities, and it often causes us to overestimate low probabilities. We have improved this mechanism since, and would probably get a lower number for election fraud today.
I'm sure this is not a new argument to you, but if you admittedly exaggerate a say .0005 chance to .05 why not .05 to .5? And a number of exaggerated small risks will add up to a very high probability quite quickly.
On the subject of this showing up baysian analysis, since viewers all tried to track it and got different results anyway:
Quant trading also works like this - it's basically using baysian inference for the stock market - and also ends up with a lot of big discrepancies and requiring a lot of Metis to get right. But it's still consistently significantly outperformed (in terms of real money earned in the real market) over traditional non-quanty trading. To the degree that baysian analysis does work in real life, this is what I'd expect it to look like.
One thing that I'm confused about is that I often see the following two claims being made:
1. Quant traders / financial people in general are one of the most calibrated and truth tracking professions.
2. Basically all traders, after accounting for fees, risk, and long run performance, underperform the index.
Which seems contradictory. However, there are obvious ways in which these claims can be both true.
1. We're talking about different populations
A. People talking online are highly selected by the competence of their peers, so all the anecdotes online are selecting for say, the top1% of traders.
B. There are two criteria for success, marketing and actual financial talent. Funds that are mostly good at selling a story to retail investors "drag down" the average, but remain solvent because their customers do not understand "underperforming the index". Calibrated people come from the latter, or produce models for the former that are then organizationally neutered in some way.
C. This is just talking about two time periods. Kahneman released his analysis something like 20 years ago, and most of what I read about this was from that timeframe. 20 years is a very long time for a bunch of intelligent people to incorporate new information and develop good norms around them.
2. The second situation is that one of the premises above is flawed.
A. Individual models do perform better or worse, but they do so for significantly short periods of time, or the models themselves are calibrated, but the financial EVs aren't, so you end up with something like a black schols situation, where most calibrated models do well, then someone accidentally has a programming error flipping a sign, or something which wipes out a lot of value, and these unlikely incidents form a balancing factor for performance. So calibration does not necessarily track financial performance.
D. The "fees" part of the "outperforming the index fund" is where all the magic happens: the individual trader or firm wins by having good models, but they aren't good enough for the retail investor. So calibration does track financial performance, just not enough.
That's not it, I don't think, since the index is a specific algorithm weighting everything and "all traders" here specifically mean "like 200 firms that Kahneman looked at in Wall Street" or something like that.
I'm not sure if that's definitionally true or false, it is the case that every trade has a buyer and a seller, but I have an intuition that at least some movement of the stock market reflects "real underlying wealth" whatever that means, and that it would be possible to have economic growth such that indexes in fact do not perform well. Tricky, tricky, I feel like my intuitive model is wrong in some way, but I don't understand what.
If Kahneman looked at a representative sample (and the guy was meticulous, so I expect him to had done that), then their performance should closely match that of "all traders". The sense in which "all traders are the markets" is kind of definitionally true - "the market" is just the sum of all trading transactions. So the "market going up" means transactions overtime happen at higher and higher prices - the reasons for the uplift are not important for this specific understanding. Once we know the traders make up the market it becomes clear that they - collectively - underperform by the amount of "friction", i.e., costs of transactions.
There's a subtle but important complication though: what we call "the market" usually refers to some kind of a market index, e.g., Wilshire 5000 or some such. These indexes attempt to represent the market, but are not "the market", and therefore have their own deviations, or "tracking errors". So in reality, it is possible for anomalies to occur where, for a period of time, it would look like "all traders" outperform "the market". When in reality what this means is "a representative sample of traders" outperformed "a market index".
No, Kahneman was specifically looking at the trading activity of Wall Street firms, which isn't representative of all trades. Even then I don't think this would be definitionally true: for example, you could have one super sucker firm, who trades in the maximally ignorant way, and has deep pockets such that they don't stop trading. You'd get the result that most firms do outperform the market. I think you have to do the hard work of actually understanding how both statements came to be, sorry.
Oh, sure, "most" is not "all", it's entirely possible for there to be outliers with outsized impact. In fact, we know this has happened - most famously in 1998 with the LTCM collapse.
Another piece is that hedge funds are generally more consistent than index funds (even if they match them on over), and for a lot of people uncorrelated returns are more valuable (see e.g. herre https://capitalgains.thediff.co/p/naivecomparisons )
Also note that Quant traders vs finance people in general aren't exactly the same groups (the quant heavy places have generally done better over the last two decades or so)
Yeah, this looks like the "Micaiah is unreasonably conflating two groups, quants and non quants into one, and people are predictably talking about quants when he hears about it online" and "outperforming index, isn't the only way to be truth seeking, turns out having a model that doesn't rely in default economic growth and still earns money has to be more accurate since you don't have that as a crutch".
There's one more potential point of explanation: exploitable mispricings are finite. Trading itself moves the market, so if you have a better model than the market consensus the absolute amount of money you can gain from trading upon it is still quite limited. Large funds – ones open to new investors and large enough to move the statistics of "beating the market" – are too large to meaningfully exploit these opportunities.
Hence, you see hedge funds and other large investment funds (pensions, for example) look at less liquid markets like private companies and privately-placed loans. In these markets, there's less public trading and thus less market consensus (thus more potential for mispricing) and without public quotes individual trading is less likely to move the average price for everyone else. If I'm a hedge fund and I know I can give some set of private companies loans at 0.1% less than my competitors, issuing that first loan isn't an immediate, obvious signal to my competitors that they have something wrong.
This doesn't roll off the tongue at all when I reply to "If you're so rational why ain't ya rich"!!
Thanks. That might also explain another flaw in my thinking about the study of below market index returns: if the firm's base rate isn't market returns but some different distribution, yeah, sure it's technically bad to give them your money, but that doesn't mean they're inaccurate. They're playing a potentially harder game.
"1/10,000 Wuhan citizens work at the wet market. So if a lab leak was going to show up somewhere random, the wet market was a 1/10,000 chance."
Why are we only counting the people who *worked* there? Wouldn't the relevant number be all the people who visited there (a much higher proportion of the population)?
The exact claim was about the index case being a market worker. That's generally OK. The main mistake is underestimating all the factors that would make that happen even if there was no spillover there.
First image: "The election was no different from previous elections" is not inconsistent with the claim that there was widespread fraud. As stated, none of the options are mutually exclusive.
(this is true regardless of the actual answer to the question)
I disagree with some of the arguments. I think COVID originally started spreading in rural villages. It didn't spread fast, because there wasn't that much unventilated space, and it wasn't well adapted to humans. And it wasn't detected, because until special tests were developed, COVID just looked like "atypical pneumonia". That is where it adapted to humans. Then it was carried into the wet market by a vendor. So it did spend some time learning how to spread in humans, but it wasn't by gain-of-function research.
It doesn't directly address it. But if I am correct, our knowledge of what viruses are circulating where is quite patchy and incomplete. I've heard that in some cases of xenonotic diseases it's taken multiple decades to track down the "original source"...with nobody trying to make excuses or prove that it isn't their fault.
Also, if the virus was originally a bat virus, bats fly, and some are migratory. So perhaps they just haven't been looking in the proper caves to find a closer example.
FWIW, I don't think the evidence is sufficient to invest ANY of the theories assigning blame with much belief. People can't even agree on what was the most probable non-human host. (Why not some weasel...we know it spread well among mink.)
The main challenge to this idea would probably be the fact that the pandemic seems to be descended from the strain of the virus that emerged in Wuhan, with scientists arguing it traces back in terms of timing to late November, right around the time the Huanan Market outbreak would have begun. This is why the scenario of spillover at the market seems to make sense, especially given there were multiple potential host species sold and slaughtered there. As it happens one of the stores at the market had a website that was archived on the Wayback Machine and you can get an idea of the range of species that were being sold:
Some of the animals were being imported from the part of China where similar viruses circulate in bats, while others are the same species found infected during SARS-CoV-1.
But all of that could be wrong of course. Apparently there was a wave of influenza spreading through Hubei province right before the pandemic began so you could speculate that this masked smaller-scale rural outbreaks. There's never been retrospective case-finding outside of Wuhan that's been publicly released. There's a report in a Hubei paper that markets in Enshi, where civets and raccoon dogs sold in Wuhan were raised, were shut down on December 23, 2019 -- a week before the Huanan Market in Wuhan was closed.
China has also failed to release testing of blood banks outside of Wuhan, and testing of the animal trade in general has been limited. Officially the government still claims that zero animals within their borders have tested positive at any point during the pandemic and there have been no outbreaks on fur farms in spite of tens of millions of susceptible species being raised there. So it's difficult to draw conclusions but the genetic history seems to point to the outbreak in Wuhan being caused by spillover to humans either at the market itself, or at a closely linked location such as a farm that supplied the market.
Is 1000 miles nearer or farther than you expect for a virus that is separated by several years or decades of evolution from the virus we're talking about?
According to Saar's argument, which Peter did not refute, the nearest example we've found was located 1,000 miles away. If there are no intermediaries between that sample and what we later saw in COVID, then that seems to increase the chances that this was a modified strain. If there are intermediaries, then we would need to know where those were located in order to determine how their descendants ended up in Wuhan.
Whichever debate participant is most well-versed is most likely to have thoroughly understood the topic and thus is most likely to have a correct intuition on the subject. Thus, we should just decide on these issues based on our subjective intuition about which debater seems to be most knowledgeable and rational, and not reason through the content of the debate except to fact-check the debaters and ensure their reasoning follows so they can't get away with intelligent-sounding bullshit. For example, instead of checking every single fact, just check a simple random sample of them. This approach may seem objectionable to some people, but it's probably what we're all subconsciously doing anyways, and it's probably approximately correct.
>Whichever debate participant is most well-versed is most likely to have thoroughly understood the topic and thus is most likely to have a correct intuition on the subject.
I just don't think this is true. "Crank that goes down an obsessive rabbit hole researching conspiracies" is an archetype that will tend to end up (a) having a lot of domain knowledge, and (b) being completely incorrect.
Like, I'm pretty sure if you took a flat-earth nut and put them in a debate against a random person chosen off the street, the flat-Earther is going to be far more well-versed. I've spoken to flat-Earthers! They love getting into minutia about discrepancies in NASA image processing, and flaws with Newtonian physics, and Ancient Greek historiography, and so on. The average person on the street is going to be a lot less knowledgeable on all of these things!
Being "well-versed" just selects for obsessiveness, not correctness.
That was exactly the example I was going to come up with. I do enjoy learning from flat-earthers, because they dig up interesting anomalies, even though I think they've fundamentally learned the wrong lesson from them.
(To be fair, the same is probably true for most real experts for at least some of the things that they have real expertise on.)
> this disease that came out of nowhere and ruined all of our lives for a few years
I wish we could have a debate on what the proper response to COVID should have been. In theory, if governments refused to do much other than developing vaccines, COVID would've had a significantly smaller impact on our lives. But also this would've also had significant costs to many people in the form of losing their elderly relatives and more people would've had long COVID.
I wish we could have a Rootclaim debate on what the Western countries should've done in late February/early March as a response to the virus. I think there's a large degree of disagreement about this question to this very day and it's definitely not been discussed enough in retrospective.
Asking the real questions! Lab leak vs zoonosis strikes me as mostly a culture war question more than anything meaningful, and serves a distraction from these brass tacks of: okay, whatever, what do we do *next time*? I fear we'll repeat the 2020 psychosis instead of having a better response.
I recognize some people are working on this- biosecurity has become a larger EA cause, as one example- but I suspect that the "people at the top" will be just as ineffectual, inattentive, and political as last time.
With regards to the culture war aspect, I collected a (possibly relevant) list of headlines on the lab leak hypothesis over time. My intent was to try to see how normal people might see the news over time on the lab leak hypothesis so I pretty much just focused on headlines and restricted it to fairly mainstream sources (I'm sure it would look different if I included, for example, Daily Mail).
If we see this sort of thing the "next time" then I expect we will see similar belief splits.
Part of the problem is that I don't think most people are very good at treating headlines as "this might be true ... but possibly not" instead seeing things as "the experts have changed their minds ... again"
There isn't any clear correlation between governments doing stuff and less covid happening. Sweden is the obvious go-to as a country that did little and got same or better results than it's peers, but Peru is also an example of a country that did a lot and got an unparalleled catastrophe for it.
Why exclusively compare Sweden with another country that had relatively lax restrictions? Regardless, middling short term results and brilliant long-term results (lowest or near-lowest excess deaths in Europe) for the one country that notably rejected all but the most modest of restrictions is hardly a ringing endorsement of restrictions doing anything.
Because they both share similar climate, culture, and density of population. I'd assume there's also no big difference in term of popularity as a touristic destination between the two. Norway had a stricter policy than Sweden, schools and several business were closed. This was not the case in Sweden.
How confident are we that climate, culture and population density are even relevant? Relevant enough that we should ignore Peru's terrible results with restrictions, and Sweden's good results with restrictions, because one neighbour of Sweden slightly outperformed it in the short term but not in the long term? Does the Scandinavian Taiga perform spooky action at a distance by affecting transmission in Stockholm?
Stringency of restrictions and covid deaths looks like a scatter plot with no correlation. This is relevant, and can't be excused away with culture.
Regardless, this is a post-hoc epicycle. When lockdowns were brought in, they were advertised as necessary everywhere to avert megadeaths. Not that some countries didn't have to do them because culture or weather. Not that rural areas didn't need to because population density. And not that the outcome would instead be going from Low Norway deaths to Middling Sweden deaths. If you believe that certain population densities, cultures and weathers mean you don't need to lock down, you're already 90% the way to opposing lockdowns, you just differ in the universality of opposing them.
I agree Norway had a stricter policy than Sweden. It was, however, still more lax than Europe.
Awesome analysis. And really good that you picked up on the personality factor at the end. Kinda odd that Rootclaim do deep Bayesian stuff on the arguments, but do no personality screening for willingness to be objective / fear of losing a debate. The adversarial system is super-Western, I guess, with two dudes fighting their corners,. But in this instance it seems to me like it could be huge factor to swing things.
The comment at the beginning suggesting that it was too bad Putin doesn't have cancer felt a little odd to me, BTW.
The snide comment about Putin also grated with me. But I remember Scott's thing about Bin Laden and Thatcher - if Putin was assassinated (for e.g. by the CIA) what would my reaction be?
You know, I looked at the "widespread fraud" one at the top, and it struck me that it didn't even consider that I would think was a much more likely hypothesis for it: That several big city political machines, acting independently and in parallel, jumped in an applied diverse methods of falsifying the outcomes, in just those places where such falsification was both (a) readily manageable and (b) critical to their respective states' electoral votes (giving motivation). It makes me suspect that you can get the results you want from Bayesian methods by carefully curating the initial hypotheses to compare. Not very different, perhaps, from C.S. Lewis's trichotomy about Jesus.
Hot take: Peter clearly failed to convince anyone.
The lab leak odds, in log10 (i.e. orders of magnitude are):
Peter -20.7
Saar 2.7
Eric -3.1
Will -2.5
Scott -1.2
Daniel -1.4
One of these numbers is clearly an outlier. Scott mentions it and calls it "trolling", I would argue that it is debating in bad faith. 2e-21 is a ratio which is just silly. For one thing, the gain of function at WiV pathway is not the only pathway towards a lab leak. The WIV could also have released a naturally occurring coronavirus at the wet market. At 2e-21 odds, we would probably have to consider the possibility that the WIV built a time machine and went back in time to infect the wet market.
Apart from the pangolin which got retconned out in favor of the raccoon dog, one thing I remember from the earlier discussions was that proponents of the lab leak pointed out that the bat with the virus likely did not come from Wuhan. Given that Saar has not made that argument, is that no longer the case? I imagine that by now we have a much better understanding of naturally occurring bat corona viruses in China.
thank you for mentioning my "preachy and annoying" blog. :P
Can't do much about my eyes rolling backward on this topic, but since you are all about evidence and probabilities; you might want to keep your eyes peeled for my next 15000-word sci comm article that comes out in two weeks and will bring people to the cutting edge of origin science, as well as put the "gain-of-function research" narrative to bed once and for all.
Maybe my goal is to accurately reflect reality, educate, and have fun writing, rather than catering to audience demand by constantly upholding the false equivalency spiel of influencers that care more about popularity than science :)
Scotts friendly bothsidism here towards a fanatic multimillionaire is what makes me cringe... guess that is what influencers do to be popular with the masses... "don't make enemies of powerful people, don't rock the boat, give your audience what it wants from you, everybody's opinion - no matter how unfounded - has value and must be taken seriously, isn't it that both sides have good points... maybe they might not be entirely correct, but those pesky Zoonati scientists certainly have their fair share of problems. too.." *yawn*
Seen it a hundred times, but a spade remains a spade no matter how many charitable hot takes are written about how it can also be seen as a fork
I think your writing is good and your videos are as well.
But I also think that the tone would put off anyone who's on the fence about the issue, because you lump together lab leak believers with other conspiracy theorists, and it makes it sounds like everyone who's taken in by the theory is just an idiot.
But support for lab leak runs around 60%, and plenty of smart people believe it. They aren't all anti-vaxxers or flat Earthers of UFO cultists.
I got taken in by it for a while, at least enough to say it was 50/50. You even did, yourself, for a briefer time (like 1 month in 2021, or something?)
I will admit that it's a struggle to write an article that simultaneously preserves the ego of the person who fell for a bad theory and also debunks that theory effectively. So you have to pick some approach there, and yours is just different than mine.
"People often frown when I call the currently advanced "man-made" speculations conspiracy myths, because "how can you stigmatize 2/3 people as conspiracy theorists?". So let me quickly clarify: Conspiracy myth is an accurate, functional term for these sets of ideas and attached worldviews. However, believing a lab leak is likely, or having fallen for one or the other false myth does not turn ordinary citizens into conspiracy theorists, nor does it make them gullible or stupid, just incorrect. We all hold wrong, uninformed, ignorant, or shortsighted beliefs on many issues, make mistakes, or just get unlucky with our information diet. Yet we are usually able to update our beliefs and correct course by ourselves when such opportunities present themselves, while conspiracy theorists usually can not. The term "conspiracy theorist" should - in my opinion - be reserved for the accurate description of participants and activists in closely-knit conspiratorial communities that are hyper-engaged and trapped in these all-encompassing worldviews, even victims of it. They need help and deprogramming; they do not deserve our attention, but also not our scorn. All my ire goes towards those who wield and amplify conspiracy myths as a tool for popularity, persuasion, profit, or power. Self-serving influencers, populist commentators, unethical journalists, calculating politicians, fanatic activists, power brokers, media manipulators, conflict entrepreneurs, contrarian grifters, and their ilk would probably fall into that group. What they have in common is that they are a type of social elite with an outsized impact on the discourse that they poison for everybody else."
Yeah, it's a hard balance to strike. I loathe a few of the people that professionally lie and manipulate the public on this issue and a few others (like vaccines or climate change).
But I have no ill will towards the average person that just thinks it's a lab leak because it's all really confusing and there are more newspaper articles about lab leak than about zoonosis. To them, it's more like:
"I've been there! I was there! I'm trying to save you a year of your life so you don't have to spend as much time as I did, sorting it all out."
In the expert poll, one fifth of respondents thought a lab leak more likely. The median expert puts it at 10%. This is, by definition, not a conspiracy theory
>However, believing a lab leak is likely, or having fallen for one or the other false myth does not turn ordinary citizens into conspiracy theorists, nor does it make them gullible or stupid, just incorrect.
I agree. So why take such a derisive and condescending tone? There is real work to be done to dispel the lab leak theory. You seem to be more interested in signaling than actually changing people's minds. Because your writing is laundry list of things not to do when trying to persuade.
I had to ctrl-f to find his mention of you after stumbling over your comment here, which stood out visually due to the annoying use of emoticons.
Does your blog contain a post about how you so fluently acquired journalism-speak? Did you really start out in 'science' without a journalism degree? Can you really write 15,000 words in that breathless, snarky tone on a regular basis? While you'll clearly never write for the NYT, this is more likely due to discrimination than lack of ability. You seem lke you'd be a great fit.
Mostly a compliment. You're in a comment section famously inimical to the NYT, but we'd all agree there's a hierarchy to journos. That is, while they're all bad for society, only most of them are bad at their jobs. Some are talented and competent. The dig is the implication that if it weren't for anti-white, anti-male bias you'd already be a big deal at the NYT. I have no idea if I'm joking when I say that part.
Separately, I'm bothering to write the compliment because I work hard at good mental hygiene. I hope never to lose the ability to give credit where it is due. The annoying emoticons and snarky tone probably activated some prey-drive-adjacent subroutine in my brain that looks for easy dunks, but you don't provide any.
I enjoyed your blog post and am grateful you chose to cover this very interesting event. I drew from your post when writing mine and hope I gave you enough credit.
I don't know if you're joking when you say you'll put the gain-of-function research narrative to bed once and for all, but I look forward to reading what you've got to say.
Tongue-in-cheek of course, my whole point is that the lab leak theory is dead, but the lableak myth will never die for factors that go beyond scientific uncertainty.
No matter what I write, media manipulators will not stop pushing a false narrative.
However, for those with an open mind and a real interest, yes, the next article will reduce scientific uncertainty to a point where no "gain-of-function" scenario remains possible.
Huh, for me the real "gain-of-function narrative" is that regardless of whether it's responsible for COVID in particular, it might very well become responsible for the next pandemic, possibly an even worse one, and whatever dubious benefits this research brings can't pass any remotely reasonable cost-benefit analysis. I'd be much more interested to see a discussion of that.
But tl:dr is: GoF is not a precise term, but has both risks as well as benefits that are outsized. Way forward is that cost:benefit needs to be assessed based on individual experiments and by experts who understand what they are talking about.
I was pretty sure about lab leak, now I'm 50/50. Normally I should be frustrated to end up 50/50 after reading all this but I'm filled with a feeling of satisfaction. Great job for all involved.
Similar for me. This nudged me from 85% lab leak to maybe 60%. I do wish the policy conclusions section included some reiteration of the lesson that clearly we need to clamp down on GoF research and generally need a ton of better, independent oversight of labs handling/creating dangerous pathogens.
I am WAY out over my skis here…but it just seems like there is no objective way to actually TEST any of your priors in a scenario like this. So you’re just giving a finite number (and “false” sense of precision) to what actually remains a messy and murky quandary.
Priors aren't the sort of thing that can be tested. Scott thinks that some priors are better than others, and that there's an objectively best prior out there, but I don't think anyone has made any better progress on this than "just assume everything is equally likely to begin with, and don't worry too much about what "everything" means". (Maybe that was a bit too snarky.)
But that’s what I mean. These guys calculated posterior probabilities to 6 or 7 decimal places. That’s an incredible amount of precision to be claiming when your starting point is an unquantified, unverified, and seemingly unverifiable prior.
So this debate brought out a lot of facts and information I wasn’t previously aware of. And it was presented by both sides incredibly well. This would be close to a best-case example of the concept, i would think. And it appears “zoonosis” “won”. Yet we are in fact still no closer to knowing who is right. That’s a fairly circuitous 15 + hr journey for that kind of payoff.
It depends on what you mean by "right". If you mean "who supports the theory that is true", I think we are closer to knowing that. If you mean "whose degree of belief is closer to the objectively correct degree of belief", well I don't think that exists.
I agree (to the minimal extent I understand this stuff). It seems, as you say, that there is no way to test how accurate your prior (or your belief) actually is.
And when I say “right”, I do mean “true” or “objectively correct”.
But this is what I’m referring to: without being able to say your prior (or starting belief) is correct (or how objectively relatively true your starting belief might be), how confident can you be that any reasoning predicated on that starting point will lead you to the correct place/conclusion? How confident can you be that your perfectly constructed and engineered thing will remain standing, when you’ve built it on sand?
If the conclusion is a probability as well, then there’s again nothing it means for that to be “right” or “correct”.
What we do have a guarantee of is that if something is true, then pieces of evidence that increase your probability for it will be more plentiful, and if it is false, then pieces of evidence that decrease your probability for it will be more plentiful. No matter what prior you start with, an investigation that continues finding new pieces of relevant evidence will in the limit converge to 100% if the claim is true and to 0% if it is false.
We can’t put bounds on that to guarantee that at any particular finite point in the investigation you will be particularly close to the truth, but we can guarantee that in the limit you will.
To ask for something much stronger than that is to ask for a guarantee of omniscience.
I agree that the objective is to get closer to knowing the true answer, without necessarily reaching a point of certainty. But this debate to me exposes some serious limitations to Bayesian probability.
Both sides were well researched and seemingly agreed to most of the facts. However, they had widely disparate priors that affected their interpretation of those facts…which then in turn affected their priors for other or subsequent facts. And they ultimately came to wildly and widely disparate results. And there’s no objective test of their priors. So it seems their conclusions are driven by their views of facts which in turn are driven by their prior biases. This does not give me confidence in the objectivity of the process or the accuracy of any result based on Bayes.
i "like" the Cosmic Joker hypothesis, meaning it's the most *interesting* one to me. i also like the idea of the Covid virus as a virus in a superposition, taking inspiration from Scott Aaronson's comment in a podcast that "There is a serious prospect that in our lifetimes people will be able to create superposition States of, let's say, a virus". (dunc tank podcast, July 11, 2020, with minor editing from his improv oral) in this krazy view, the two states of the virus are: (i) being from the Wuhan virology lab: and (ii) being from the wet market. the evidence never resolves decisively to one of them. to put it another way, a meta-version of the lab leak hypothesis is true: the Lab is some Schrodinger's Box, and the Covid virus escaped while still in its superpositional state. it moves though our world leaving a neat two-path evidential trace.
Very nice, thanks for watching, and reporting. I want to make an argument for agnosticism. It seems like we are all inclined to think we must have an opinion on almost everything. What is wrong with saying, "I don't know"? Being trained in physics, I just want more good data before I try and make any sort of decision. Eventually I hope we will hear news of some intermediate animal host, or news from some researcher that says, yeah we were doing GOF on those viruses. Until then I don't see how the odds (for me) can be anything other than 50/50. A sign of my almost complete ignorance*, and an unwillingness to give anymore credit to one sides 'experts' vs the others. I think the best path for humanity is to assume both things are true. And try to make both forms of possible virus mutation less likely to happen in the future.
Oh, and why are there no error bars in all of this analysis. I'm thinking the error bars should be huge.
*I read a lot of stuff about this in 2020, said then we need more good data, and stopped spending time thinking about it.
Because it was turned into a political issue (very early on) so you have had two choices since roughly February 2020: loudly proclaim the party line (of the tribe you belong to), or be othered. This includes loudly proclaiming masks don't work and being a derisive asshole to anyone who disagrees, then loudly proclaiming the very next week masks are absolutely required and being a derisive asshole to anyone who disagrees.
Saying "I don't know" is dog-whistling for the other side.
Yes, that does mean we on-the-spectrum get kinda fucked because we speak in terms of facts and objective truths, but this is nothing new. History is full of this.
You are saying it as a joke, but it is a very important point. Why do we care if the probability of someting is this or that? We often care because the probability we assign is a signal to others about who we regard ourselves to be, including who we regard as friends and who we regard as enemies.
...and even if we do not care, we overlook that others will interpret our assigned probability this way at our own peril.
Thanks for the write up Scott! This is definitely very interesting.
I find myself walking away with a couple of thoughts. First, about the debate itself:
1) I started out pretty strongly leaning towards but far from certain about Lab Leak being correct. A lot of this I think stems from finding out that the anti-lab-leak side was a literal conspiracy that had a lot of institutional support to try to shut down the debate. Even those shouting down the theory secretly thought it might be true (at least early on)!
2) Likely related, I could feel learned helplessness setting in while reading the back and forth descriptions of the debate. With no priors on the believability of either participant, and seeing that both were making very strong claims that could reasonably answer their opponents, I want to mentally shut down and hold my existing perspective.
3) I end the debate slightly closer to zoonosis/no opinion, but still leaning towards lab leak.
Meta-discussion I feel much worse about Bayesian reasoning, especially the hard version with math. Peter clearly has a number that's beyond implausible, but the error bars on all of the numbers seem unworkably large. You could adjust all of the factors up and down by relatively small amounts and come up with any interpretation.
Even more, sometimes really really implausible things happen. As Scott notes here, something really implausible definitely did happen! In fact, across the world, implausible things happen quite often. People die of incredibly unlikely events all the time - an average of about 7 kids per year in the US have drowned in buckets since 1984. If you had never heard that number, what probability would you put on the chances of a child drowning in a bucket? There are 22.4 million kids 0-5 in the US right now, with seven per year drowning in a bucket, which suggests about 0.0000003125 chance of a child drowning in a bucket in a given year. No matter how you do the math, the chances are very truly low. But it happens seven times a year.
How would we model COVID's origins if something incredibly unlikely happened? It's not impossible that it originated in neither the Wuhan Institute nor the wet market. In the realm of possible, it may not even have started in Wuhan. Saying that something is *more* or *less* likely is all well and good, but reality happens in exactly one way, which can be the very rare event. I think 1/10,000 is incorrect for the likelihood here, but sometimes 1/10,000 things happen. Just because we say something is unlikely doesn't mean it's untrue. Saar implicitly believes that the wet market theory is ~6% likely to be true. 6% is far from ruling something out, even if he's arguing as if it's closer to 99.99% the other way.
#2 My beliefs shifted from pretty confident in lab leak to 50-50
#3 Reading through your summary, I often got this horrific feeling of college debate and dropping points. College debate has this concept of dropped arguments, where the worst thing you can do is not respond to a specific argument your opponent makes, because they automatically win that argument. That's what this felt like, I'm not actually convinced by Peter's overall story, I'm not even sure there's a coherent story there, but Peter kept coming up with specific situations where Saar was wrong and that convinced me.
I'm reading back through the lineages thing and, just reading through, Peters positive argument for why the older lineage started spreading later is pretty handwavy. Compare that to the next section where he goes through why the intermediates are a programming error, he's on point. But I come back to the core question of why the older lineage started spreading later than the new lineage and I don't have a great answer to the main point, and I think Saar's is better here, but Peter dominates on the subpoint.
I'm just scared of focusing too much on the subpoint and irrationally weighing that too heavily. I dunno? For people who watched it, did you get "debate bro"/Destiny vibes from Peter? That's what this kinda smells like, I notice intense debates should work but I could also watch 50 debate streams and at the end of it be dumber than I started and I felt this way a lot in college debate as well.
This is probably not the important issue here, but as a former university debater I feel compelled to note that the 'dropped arguments' thing is specific to certain formats of competitive debating and far from universal.
This was a wonderful read and personally seems like a great experiment. A bunch of credit goes to Saar and Rootclaim for courtesy and making this public and educating all of us, whatever position you hold.
An insanely good debate in many ways, and congratulations to Scott on a most excellent summary. This reads like a communication from an alternate universe. Seriously, since when are there 15-hour highly structured debates at the object level with this level of intelligence brought to bear? That just does not happen in this world, except it just did.
Despite the outcome not being what Saar wanted, he has ably promoted Rootclaim as a useful tool. I want a lot more of whatever this is.
Despite Peter being one of the smartest people I've ever encountered online, arriving at an odds ratio of ~10^21 against lab leak is prima facie evidence that he has not correctly employed Bayesian reasoning. That sort of ratio is where you start getting "simultaneous miracles". I estimate that the chances of a *natural* zoonotic event, meaning one in which human action was not a causal factor, infecting its first patient *inside the virology lab*, as being *considerably* more likely than 10^21 to 1 against. I possibly could estimate the chances that I, personally, created COVID, had it released at the Huanan Seafood Market, and suppressed my memory of doing so, as 10^21 against. (By this I mean if I sat down and calculated how likely such an event was, 10^-21 is an answer I might obtain. In reality I would throw out any probability that low and group the event with "weird shit happens".) 10^21 is a sextillion. I can't count that high and neither can Peter.
Not getting answers like 10^21 touches on a real problem with practical Bayesian reasoning, which is that almost all independent events are only very nearly independent. Clinamina like the Self Memory Wiping Garage Biolab Operator Who Is Me exist in enough possible worlds to start having an effect if you multiply together enough large Bayes factors. This is far from disastrous to the project of being better Bayesian reasoners but there should be more effort going into developing and teaching the techniques that deal with it. Rootclaim looks like one of the places that might happen.
>Despite Peter being one of the smartest people I've ever encountered online, arriving at an odds ratio of ~10^21 against lab leak is prima facie evidence that he has not correctly employed Bayesian reasoning.
Agreed. 10^21 is an absurd number. There have only been 10^17 seconds since the Big Bang!
TBF I think this just is a matter of approach. Peter is reporting the nominal calculation on the assumption that the theory is correct, the evidence is what it's claimed to be etc etc..
I think it's clear Peter doesn't actually believe those are the betting odds that one should use on the subject.
So I think it's less of an error and more a notational difference.
"Seriously, since when are there 15-hour highly structured debates at the object level with this level of intelligence..."
I just wanted to suggest that actually, they happen all the time: in courts of law. And it's instructive to think how that comes about. It requires wigs and centuries of tradition; and really strict rules that get actually get enforced about what you can and can't say; and real consequences at the end.
And there is an obvious legal correlate to this error that Peter was making with his 10^21 claim: that poor woman who suffered successive cot deaths and was put in jail for them, because a statistics expert multiplied the probabilities of cot deaths together and worked out there was only a 1 in a million chance she hadn't murdered her own children. She was eventually released.
I was also thinking of that story related to the dangers of professionals tasked with making difficult decisions becoming (amateur) Bayesians. Could not find the proper reference, though. If you have it, I'd be interested.
One thing worth noting is that the odds Rootclaim gave at the debate are considerably higher, their central estimate was 1 in 180,000 for zoonosis and their high end estimate was 1 in 700 million.
To a large extent, I decided to just mock the process and make up even bigger numbers on the other-side. (more specifically, I gave fair justifications for many of the numbers, but the idea that you can just multiply them all together is not fair, because you might have picked the factors in a biased way, and in the end the argument should be bounded by whatever confidence you have in your own abilities)
For this write-up, Rootclaim asked Scott to change all their debate numbers to smaller numbers. The furin cleavage site that was 1 in 100 at the debate became 1 in 20, some other factors just got dropped entirely, etc.
Charitably, I could assume they asked Scott to make the change because they have since become less confident in their arguments.
But, cynically, I might think they did this to make it look like, "they were reasonable, I made up numbers that were too large, and the judges were just too stupid to tell the difference so they came to the wrong answer"
It is described as "excellent" in the post here but, despite being the most comprehensive and sophisticated analysis to date, its aggregated results aren't included.
He explains those choices in detail though. Many people would benefit from reading his explanation of why the market data doesn't support market spillover.
I know you know all this, but other readers here should note that of the three big papers arguing for market origin:
Worobey et al 2022's conclusion is under heavy fire in the statistics journals (Stoyan and Chiu 2024, Weissman 2024).
Pekar et al 2022 already had to correct their code ... and change their text to adopt a more lenient significance threshold. The same guy who found those problems has found further issues that invalidate the conclusion completely, though the authors have not had a chance to respond fully yet. Some ACX readers will be amused to hear that the four errors found in Pekar et al 2022 all had the same sign.
Crits-Christoph et al was debunked before it was even published.
I give A/B zero weight in order to be generous to the zoonotic case. Standing by itself, without the A/B data, without the missing RNA/hostDNA correlations, and without data on Wuhan markets as a fraction of China markets, then empirically the HSM cluster would get around a factor of 10 favoring zoo. Putting all those factors together tends to weigh against the market case, but a priori markets gave only a fraction of the zoo priors, so it's not worth penalizing zoo for the collapse of the market case.
OK, here's a table, for what little it's worth. Without the explanations, not much. Odds are for general zoo (ZW) vs. DEFUSE-style LL.
Priors: some pandemic start in 2019 1/70 ±x10
Sarbeco: 10/1
Wuhan: 80/1 (would be roughly 800/1 for market)
No-host: 3.3/1
Pre-adapt: 2.7/1
Has FCS: 3.3/1
CGGCGG: 7.3/1
Restriction site pattern 70/1
Each of the factors has already been discounted by a hierarchical inclusion of uncertain nuisance parameters. E.g. the FCS factor allows for selection by using only successful non-bat sarbecoviruses for comparison. I include a nuisance parameter for possible irrelevance and end up with only the 3.3 factor although there are 27 non-bat species with SC's detected and zero of them have an FCS.
The final odds are calculated by integrating over a fat-tailed version of the highly uncertain priors.
During the week of January 23rd, 2020, the day lockdowns were enforced, there was somewhere between 175-225 excess pneumonia deaths in the Wuhan DSP per 100K. This is eyeballable in figure 1 here.
They report 243 per 100K for the entire first quarter of 2020. That's an *enormous* number.
Smooth this out over a week and say that during the week of 1/23/2020, ~30 more people than expected were dying from pneumonia in the Wuhan DSP per 100K.
The cool thing about the beginning of pandemics is that there's no immunity and the mixing assumptions of SIR models probably work well in the beginning of a respiratory outbreak. You can play with this one:
I couldn't find good information on the size of the Wuhan DSP, but allow it to vary between 2M and 10M people. Assume the R0 was between 3-4, the CFR was .5%-1%, and a few other now fairly well known things about the course of the disease. And look at the fatality numbers.
Using these ranges and standard compartment model logic, I don't see how you get such an enormous increase in excess pneumonia deaths consistent with an early December outbreak in the HSM. It doesn't make any sense at all. The best case scenario from these simulations is sometime early in November 2019.
Several of Peter's arguments treat the doubling rate as a real property at low populations. That doesn't seem right.
If patent zero only infects one person, the pandemic is delayed by one doubling period. If they infect 4, it arrives one period sooner. What are the relative odds between someone infecting zero, one, two, four, or fifty? When there are enough infected this kind of effect averages out and we can just say it doubles at some rate. At small numbers you get swings in infection progress based on how each person rolls.
Because of that attempts to extrapolate infected populations backward using the doubling rate seem doomed once you get into single digits. The error bars on that kind of estimate should be large. It's noisy.
So I don't think you can confidently say things like this
> The COVID pandemic doubles every 3.5 days. So if the first infection was much earlier - let’s say November 11 - we would expect 256x as much COVID as we actually saw
Or this
> Although Lineage A is evolutionarily older, Lineage B started spreading in humans first.
> We know this because Lineage B is more common. Throughout the early pandemic, until the D614G variant drove all other strains extinct, a consistent 2/3 of the cases were B, compared to 1/3 A. Both strains spread at the same rate, so the best explanation is that B started earlier than A. Since COVID doubles every 3-4 days, probably Lineage B started 3-4 days earlier than Lineage A, which explains why it’s always been twice as many cases.
Yes - the Skagit chorale is a great example of the superspreader dynamics at play. R0 is a post-hoc measure for clowns when it's as k-dominated as COVID seems to be. The problem with "COVID spread in the wet market at exactly its normal spread rate, doubling about once every 3.5 days. Stop calling the wet market a super-spreader event." is that it's "normal spread rate" is an average that is lumping in **the fact that COVID has super-spreader events**.
I think it sits here unaddressed not because a rebuttal is easy and people are bored, but because a rebuttal is hard, and unless rebutted, it seems a fairly devastating point. To elaborate a bit - and to integrate over a few strands of ~independent evidence:
a) We have two lineages, Lineage A appears (with v. high probability) older, and the wet market cases are lineage B (and iirc, no original Lineage A cases at all).
b) We have other relatively early cases that don't have a known market link
c) None of the early wet market cases look remotely like Patient Zero smoking guns (none of the animal trade workers, no infected animals detected at the market etc).
d) We see superspreading events at other wet/seafood markets in other cities/countries.
These fit quite well with a story of "the outbreak starts at unknown location, the first detected cases we see at the wet market are several generations in [something that in the abstract is pretty likely - what are the odds of observing Patient Zero?], a superspreader event happens at the wet market". This story is bad for zoonotic proponents because it doesn't have a high odds ratio of P(.|zoonotic)/P(.|LL)
The response is "well, we can't be missing the first few generations because we know the doubling rate and if were missing the first few generations then by January there would have been far more deaths/cases". And then conclude:
(i) "well Covid can't have started earlier, so there can't be meaningful early non-market spread, so anyone who looks early and is unlinked must have had a really short incubation period", [even though ex ante that's less likely and should be penalised].
(ii) "well, Covid can't have started earlier, so even though Lineage A looks older, since Lineage B is more common, it must have started spreading first"
But these deterministic timing arguments are nonsense because not only is stochasticity potentially important anytime you have small n, the over dispersion of Covid spread is huge so the variance of spread early on is very high. The most common number of people for someone to infect is 0. The second most common is 1. The vast majority of spread is driven by a small minority of infected people.
So the argument dismissing the Lineage A vs Lineage B point dies immediately. (Suppose Lineage A has a 1-2 generation head start. What is the probability that a month hence it has more people infected than Lineage B, if both have identical latent spread potential? It won't be far off 50-50, even if you condition on both surviving, which we should). You don't even need to penalise Lineage B being smaller as "improbable bad luck, what a coincidence that the more predominant lineage is the one at the market", because while there is probably lots of room for random luck, the most likely reason Lineage B "got lucky" and outgrew A early on is that it (by happenstance) spread at the market while A did not. E.g. if I tell you "Lineage X has a 1-2 generation head start, but the first known Lineage Y case is at the wet market, which lineage has larger share a few generations hence", isn't the smart money on Y?
And the point of "the total outbreak would be too big by date T if it started a couple generations earlier" also dies immediately because the 95% CI of outbreak size after 1-2 months given a t0 probably covers at least an order of magnitude.
It seems really bad for Peter's arguments in general that he would consistently make this obviously wrong argument, to argue against multiple lines of evidence, seemingly without realizing that it is wrong.
I can't find the ref, but there was a paper that looked at changes in reporting criteria during the outbreak. In addition to any differences due to the small population, the reporting bias also changed over time.
The doubling rate argument is, indeed, oversimplified and invalid in the earliest stages when you have single/double digits of cases. I have also programmed epidemic simulations that take that into account. Here's a recent write-up:
I find that for the range of values that I find most plausible, it's hard to support any case that there was a large covid outbreak besides the market.
If I were really motivated to develop a lab leak case, I suppose the approach would either be to say that the outbreak in January is much larger than anyone thinks, or that the growth rate (R0) is slower than anyone thinks. I don't think that changing the dispersion/superspreading rate (k) would impact it as much.
Changing January size is hard because you can bound that by excess deaths and serology. And you can measure R0 in other cities besides Wuhan, so you can't make that so much smaller.
I am, of course, not sufficiently confident in this model to say, "there were exactly 70 cases in Wuhan, on December 10th". There's a range of values, both because the parameters aren't exact, because there's a statistical range even within given parameters, and because the model may not perfectly represent reality.
None of this absolutely disproves a lab leak, either. It just makes it very unlikely because the lab leak needs to cross town to the market within, say, the first 40 cases or less. And then both lineages need to show up there, and the clock needs to reverse by chance.
It's just some code I cooked up from scratch in C++. Some of the input parameters are based on Hao et al 2020.
I'll post it at some point -- one scientist wants me to publish it, maybe I'll do it that way, maybe I'll just share it otherwise. I'm too busy with some other stuff for the next few weeks.
The original motivation was just that I didn't want to trust Pekar's paper on faith, so I wanted to see if the same results just fell naturally out of any reasonable model. I tried several models and they got similar results.
I think the person suggesting you publish it thought that peer review would be a good learning experience for you, since you weren't accepting informal feedback. I'm afraid they weren't suggesting that your work was good enough that you should go public without anybody double-checking it. You didn't seem to understand why people were telling you to preregister either.
Pekar et al 2022 has serious errors despite having 29 authors and however many peer reviewers. @Nizzaneela has worked in the open so you can see how subtle the bugs are. And you think you can just hack up some C++ and settle the COVID origin question!
Have you thanked Rob Townley yet? If he hadn't taken pity and given you the Hao paper a couple days ago, you'd still be trying to reinvent SIR from scratch. You flat out insulted the expert who recommended the Epidemics on Networks book and python code for you.
The original model I built used a poisson distribution of secondary infections and it still showed a similar result for 2 lineages vs 1. I didn't test that one for outbreak timing.
I can do a write up at some point to show how this all varies with each parameter. But I need to get a smaller search space before doing that.
The first thing I was trying to do is get a better answer from a few scientists about how to set the mutation rate. The answers I got were highly technical and will take me days of study to sort through. I could just linearly scale it to some value that I think is right, but I want a better justification than that.
> This is a pointless too-clever-by-half “flourish” that there would be no reason for a human engineer to do.
Have you ever met an engineer? We do this kind of nonsense all the time! This particular argument cuts in exactly the opposite direction Peter intended.
I don't have any comment on the outcome, but this little detail drove me nuts.
I think it's different in biology (assuming you're speaking from a non-biological engineering point of view?). It's been a while since I did any myself, but making mutations in any organism is not something you mess around with. You insert your chosen mutation in the simplest possible well established way because even then most of the time it won't work. Messing around for the sake of it is best kept to other things, like giving genes and alleles funny names.
"As per last year, we will not be subco ntractin g any funds to the intuiti ons in these
countries. All efforts expended in these co untries will be from collaborating
partners and not funded by our award. PI, Co-Investigators or other team
members may con duct short field ttips to assess markets, identify wild life in
them, and anange for shipment of samp les of bats and other high-risk host species
in co untri es that nei ghbor China (Burma, Vietnam, Cambodia, Laos) and that
supply wildlife to the internationa l trad e to China (Tha iland , Malay sia,
Indon esja)."
i.e. they say in 2017, they did sampling in these countries, and all samples would be sent to the WIV. Weirdly, all of the closest known relatives to SARS-CoV-2 have been found in places the WIV was actively sampling for more than a decade.
This is not even considering the sampling that the WIV would have done without EcoHealth funding.
The DEFUSE proposal proposes extensive bat harvesting in Yunnan in 2018. This is also evidence against them having given up on bat harvesting in 2015.
My intuition is that Rootclaim is correct that you should be able to use Bayesian spreadsheets--especially after some adversarial collaboration--and get a format that consistently outperforms intuition and even professionals informally weighing the evidence. It's sort of like how as an analyst I use fermi-calculations to get much more reasonable estimates of unknowns than just trying to estimate the unknown directly. I think Rootclaim should focus more on testing that.
I'm with you (after reading two long summaries of their debate) that zoological is more probable but also that Rootclaim is a worthy project. I wonder if the change in format that Saar should focus on though is more emphasis on the spreadsheet and adversarial collaboration to get both participants using the same spreadsheet. I'd love to see a version of the spreadsheet that two smart opposing viewpoints people agree is a reasonable setup (without any multi-stage fallacies, etc.) so we can see which line items are driving most of their disagreement free of any logical reasoning errors (or at least free of errors that adversarial smart people could identify). And then judges can focus more on the individual ratios (and any remaining spreadsheet steps that debaters couldn't adversarially reconcile)
The problem is that it's us people don't come pre-equipped with a distinction between priors and evidence and we can't unlearn facts. Thus, even if in theory it's a good system you can't get reliable data for it because you can't go back and figure out what your prior was. As such it's very easy to double count or undercount evidence.
It works great in situations with some obvious prior because it's a common situation (do I have cancer...well I know what the base rate for my demographic factors is and can adjust that by test results) but most problems aren't like that.
Doesn't normal human reasoning have the same problem but just isn't explicit about it?
The hope here though is the adversarial collaboration aspect of it would point out when double counting or undercounting is occurring and judges could focus in on those areas as driving the crux of the disagreement.
Ofc it does, but the point is that we don't have a reason to think this mechanism solves those kind of failures as it's being promoted to do.
It's like saying: I'm going to solve the political mess we are in by first figuring out what the government should do and then persuading the voters to support it.
I mean maybe I will but breaking it up into those two steps didn't help, it just gave names to a way you could imagine making things better.
That was really interesting. My bayes argument would be something like "If covid WAS leaked from the Wuhan research lab, how plausible is that picture of events?" If so I'd expect an average amount of "well, we don't know" for scientific research. But each thing that needs to be explained away rapidly adds to "that doesn't sound right". Today I learned that they did do modifications like this. But I also learned that they were usually publishing their research like any academic institution, so you need coincidences or cover-up to suppress anything they'd done on COVID specifically.
This includes a lot of things that I don't know. But I still like my reasoning, because I know there's lots of biologists out there who have a better picture than me. And I haven't heard of any reason for biologists to avoid thinking about a lab leak, and don't think it's likely there could be a reason that wouldn't be well-known. (Whereas there ARE reasons for physicists to dislike many worlds, or doctors to dislike "chronic fatigue exists and easy weight loss doesn't") But what I usually hear from random biologists is "well, yeah, it could be, but I don't have any particular reason to think it is". And so I think that's probably right.
I'm interested in the ways of putting more specific numbers into reasoning but I think you can rapidly get into territory where it's hard to tell if the reasoning is sound. E.g. trying to describe how much of a coincidence each theory is raises questions like, if it was zoontic, how many other places would have sounded as suspicious as the wuhan research lab? Restricted secret areas? Other kinds of labs? How unique is the nearby lab, in advance? If it was a leak, how many places would have seemed as suspicious as a wet market? Nowhere else seems very suspicious, but "coming out of nowhere" would have seemed completely plausible for zoonotism. Or how much are the lab and the pandemic both there because there's convenient reservoirs of coronoviruses nearby, that could be, I've no idea. I think trying to make these precise leads to "things we didn't know we needed to account for" swamping the analysis.
I ought to read how the probability calculation is actually working. I would be really interested. Although I'm not sure it would help: e.g. I'm more persuaded by my argument of listening to biologists, so any attempt to quantify likelihoods for genetic arguments will be swamped by the chance of something I don't know mattering.
I should also add, I don't really know how plausible it is that COVID was circulating a few months earlier and just happened to take off at the time we identified. That seems pretty plausible from a general "we can't usually trace outbreaks back to an individual person" expectation. This debate showed some argument why it's plausible it started at the market, but I don't understand the biology well enough to really know. But it doesn't really affect how likely I think a lab leak is, that's more based on the default in the absence of evidence being spontaneous mutation, not the specific evidence for the market origin.
Perhaps this description of the debate wrongly magnifies this, but it struck me as really discrediting to the lab leak side that they kept relying on debunked claims. The LL side would bring up things lime the 90 missing cases, the daily mail guy, or the idea that the market was a superspreader event, and Peter would describe that evidence in greater depth and be able to pinpoint why that evidence was faulty, which was not rebutted by the LL side. One of the most annoying habits conspiracy-adjacent thinkers online have is to gish-gallop between weak evidence they're only lightly acquainted with, and then not change anything about their thinking when a piece of evidence is disproven. I generally find the meta-bayes "what about the odds your argument is secretly bad and mine is secretly good" mostly annoying, but how can you not apply a pretty heavy discount to the credibility to a side that is frequently trotting out rather flimsy evidence?
Agreed, and this is a significant part of what updated me heavily towards zoonosis. I didn't get the impression the lab leak folks updated at all on Peter's comprehensive dismantling of a large number of arguments.
The $100K prize introduces a huge signaling hazard, mainly because Saar is rich and Peter isn't. I'm sure Saar's lost that much playing poker. For Peter, this is a sizable portion of his net worth — so he plays to win (and, as other commenters have pointed out, may be playing a bit dirty).
That said, the biggest issue with this debate format is that it's about an *incredibly important topic* where several major players have *a lot* of skin in the game. As much as this is a battle of evidence, it's also a battle for narrative (see Peter's rationale for wanting to get one up on the conspiracy theorists). And the "Official" narrative, that all good, sensible, educated people should agree on (you don't want to be a conspiracy theorist, do you?) is zoonosis. There are probably better, less charged questions where the Rootclaim Bayesian approach would work.
Not that I'm entirely convinced by Lab Leak either - maybe 60% odds? But given that I read ACX while goofing off at work, I don't have time to watch 15 hours of debates. Gotta go with my gut at some point.
I don't think that's really a problem. Saar has an idea he wants people to take a look at so he's paying to get eyeballs. It's just like Pepsi advertising on TV with more social benefit.
I actually underestimated the signalling value of it since I'm so used to rich people (my pillow guy) announcing this kind of shit but then adding fine print that makes it impossible to win.
Maybe signaling hazard is the wrong way to put it. The $100K means there are very strong incentives for Peter to *win the debate*, rather than find the truth. Saar is rich af and it's marketing dollars to him.
Certainly, but that's not bad for the listener. Indeed, I'm well served if both sides are strongly incentivized to win (and Saar is motivated here by wanting his idea to look good).
Two big issues with trying to formalize Bayesian reasoning is to properly identify what is already part of your priors and what isn't. Second, using it at scale in a systematic way so that you actually benefit from the correct updating mechanism. After all, the more you just rely on the prior the less it makes a difference.
The problem is that we only pay attention to issues once something makes us think there is an interesting question there and often these issues are relatively unique so we have no obvious choice of base rate to use as our prior. And there is the very real danger of pulling in information that is really only a result of the update.
I think distinguishing the prior is one place where AI can sorta help. Maybe not yet, but a great feature with AI is that you can remove data from it's knowledge and see how it would judge things. I'm particularly hopeful for this in politics (this AI predicts your judgement in these political issues 90% of the time and when it doesn't know that Trump supports this policy it predicts you will too).
I'm more skeptical about the ability to do this systemically at scale in a way that uses enough interconnections to improve judgement.
Some people talk about mammal origin at the seafood market, some talk about accidental infection from a bat at the virology lab, some people talk about gain-of-function gone wrong - why does no one consider the hypothesis that it was actually the seafood at the seafood market?!
I think people get caught up about the furin cleavage site mutation because they're looking at the wrong base rates. To think about the furin cleavage motif, you need to ask (I'm phrasing these questions a bit imprecisely):
- How often are furin cleavage motifs studied in places other than Wuhan? How often are they studied in Wuhan?
- How often are furin cleavage site insertions the cause of viruses learning how to become pandemic? How often are other mutations the cause of viruses learning how to become pandemic?
I don't see people asking those questions. The fact that people don't ask those questions makes me suspicious of everyone involved, honestly. These appear to me to be the questions that arise out of applying Bayesian reasoning straightforwardly. (I'm also suspicious, as a sidenote, because the question overall is kind of irrelevant: everyone agrees lab leaks can cause pandemics, and everyone agrees zoonosis can cause pandemics, and everyone agrees zoonosis is more common than lab leaks, so we should just try to prevent both, focusing more on zoonosis than lab leaks.)
Also, WTF do people care so much if this particular disease originated from a lab leak.
We have a pretty good grip on the problem of biolab safety (eg the leaks in the UK) and the existence of gain of function research. So whether or not COVID itself happened to be a lab leak isn't reason to update too much on the future probability. Any forward looking policy question shouldn't really depend much on whether COVID was or wasn't a lab leak.
I get the debate is interesting but it's ironic that a claim to be more rational is being tested via a subject that it's irrational to be so concerned about.
But people aren't more morally blameworthy because they were unlucky. I don't think we have any reason to believe that the people in Wuhan virology were behaving particularly badly so why go looking to punish someone?
If you want to punish people who behave badly we have a great way to do that: look for the evidence of bad behavior and investigate those cases not the ones where things just happened to go badly.
Sure, for some types of behavior that's really the best we can do. The primary evidence we get that someone might have been shaking their baby is when it goes wrong and the baby ends up hurt/dead (tho still often wasn't shaking). But this is exactly the opposite situation.
We have piles of information about how biosafety labs are run and plenty of cases they behave incredibly dangerously. We know what research gets done because it usually gets published etc. So we have plenty of leads about who is behaving in unsafe ways if we are intent on punishment.
Focusing on the time shit went just seems like hurting someone to give us emotional satisfaction about a sucky thing we experienced and that's awful.
>I don't think we have any reason to believe that the people in Wuhan virology were behaving particularly badly so why go looking to punish someone?
Well, the talk about doing dangerous research in an insufficiently secure lab does sound pretty bad, and that the US were willing to collaborate with them also isn't a good look.
A general program to go after people doing what you just described seems reasonable but it doesn't seem we have particularly strong evidence it happened here.
If it shows up fine. But there isn't any reason to particularly focus on this case because bad shit happened. If this was the cause it would be crazy unlikely this was the first time someone was breaking the rules like this.
Most people who care probably have political or emotional reasons, whether it be a desire to blame China, or anger at the US health institutions and distrust of the mainstream media.
But ignoring that, if the risk of pandemic from zoonosis is high and the risk from lab leak is low, we probably want more gain-of-function research so we can prevent the former.
"Either a zoonotic virus crossed over to humans fifteen miles from the biggest coronavirus laboratory in the Eastern Hemisphere. Or a lab leak virus first rose to public attention right near a raccoon-dog stall in a wet market."
One thing that strikes me is that there is an asymmetry here. The first part of the above does indeed sound like a grand coincidence. It's not like the mutating virus would know that it happened to mutate near a lab. However, the second part is only a grand coincidence if we assume that if there was a lab leak, that it was a) fully accidental, and that b) no one acting in bad faith had the ability to distort data from early testing, etc..
I'm not sure I'm of the opinion that lab leak is thus more likely. But I can't help but think: it sure would be convenient if you were a lab researcher/administrator/related government official and you didn't want to be blamed for something you negligently made possible - then it sure is convenient that there is that nice big wet market nearby that perhaps you can shift the blame toward.
My current thoughts: though I didn't sit through the 15 hours - I am now of the opinion that the zoonotic hypothesis is stronger than I had thought before. I am updating to be less confident in the lab leak. But I'm probably now to about even odds, fwiw.
Exactly my thought. I don’t understand why this argument wasn’t presented or discussed. I know nothing about it, but why would we trust any data or research coming from China on this? My default prior would be that the ccp goes to extraordinary lengths to fabricate an alternate reality in many other places, and they have reasonably strong incentives to do so here, so why wouldn’t they? I’m assuming this presents as obviously not the case to the people on your ground otherwise it would have been brought up?
Well, the best argument here is that the CCP likely didn't know exactly what, if anything, they needed to fake in the earliest days, so there's no particular reason to suspect the initial data to be (competently, non-obviously) tampered with.
There are very few coronavirus laboratories, but there are many wet markets. We don't really get to condition on "raccoon-dog stall", but rather the set of all animal stalls such that they are plausible carriers such that a zoonotic origin could be posited - so:
Bats
Pangolins
Mustelids
Felines
Dogs
Deer
Rodents
Marsupials
And at this point I'm just tired of listing the animals that COVID-19 or sufficiently close diseases have been observed in that somebody could posit a zoonotic origin, given COVID-19.
But it's worse than that, because there's nothing preventing a zoonotic disease from appearing in places other than wet markets. That it appeared in a wet market near a particular stall feels like strong evidence because it's so specific, but -any- real world-case would be equally specific - what we actually need to consider is the full reference class of places which might -feel- weirdly specific. Like, it had first appeared in a zoo with an on-site veterinary clinic which regularly treats wild animals. My local zoo does this, although I can't find statistics on how common this is. Or maybe the zoo just got a shipment of wild animals, which isn't that unusual, particularly in China. (Sure, none of the shipment of animals tested positive, but did any of the raccoon-dogs test positive?) Or maybe it first showed up in a restaurant who had a staff member test positive, and who had recently been at a wet market.
That it first showed up where it did, then, doesn't actually give us very much evidence for or against a zoonotic origin - the reference class for places that intuitively feel weirdly coincidental/specific is huge, because the specificity is post-hoc.
However, that it showed up where it did relative to a coronavirus research laboratory researching not just the disease, but the specific thing that made the disease such a problem, is much more unusual; there are only a handful of places in the world which satisfy even a broadened criteria.
You'd have to have quite high priors against a laboratory leak happening at all to arrive at a conclusion, based purely on the location of first observation, to think that it is zoonotic based on that information. The location is strong evidence in favor of the lab leak hypothesis.
I think you are overstating the numbers here. 38 raccoon dogs were sold per month at four markets in Wuhan. So, maybe around 20 raccoon dogs were sold at Huanan in the month of November 2019. The zoonosis hypothesis outlined by Miller is the following specific claim: the 20 raccoon dogs sold at the Huanan market were infected with covid and caused 8 spillovers, 1-2 of which sustained an ongoing outbreak. This is an extremely small fraction of the relevant wildlife trade in China, <<<1%.
Despite this, there are various points of counter-evidence against the raccoon dogs:
Negative correlation with covid in the environmental sampling, but positive correlation for animal viruses the animals were actually infected with.
No raccoon dog ever found infected with the progenitor virus
No raccoon dog ever infected by humans in later outbreaks, despite this happening several times with other animals such as mink.
The raccoon dogs were wild caught in Hubei, not farmed
Michael Weissman's analysis is by far the most principled and persuasive thing I've read on the topic, it deserves much more discussion than just the link.
Some examples of things mentioned in Weissman's analysis that don't show up here:
* the initial outbreak at a market proves very little: the initial Beijing superspreader event was in a wet market, and the largest outbreaks in many countries were at markets. [EDIT: actually this is mentioned above]
* lab accidents are much more common than you think
* spatial statistics experts dismiss the methods of Worobey 2022
* the 2018 DEFUSE proposal really predicted some highly unlikely features of the virus
This was a debate between a smart autodidact and an unprepared multimillionaire, in a format which prevents easy annotation and rebuttal. It's maddening to watch the debate because Peter gets away with so much. I'm really surprised how much people are updating based on it.
Do you think someone reading Scott's summary and someone watching the debate video would come away with very different impressions? It sounds like you did?
I think I basically agree with you on all the substance - I was frustrated by some of Peter's questionable claims that were allowed through as fact, and the Weissman post you linked to was very helpful (and does a much better job of discussing what the conditioning set is, and how that affects what different bits of evidence mean). But beyond everything else, I agree with you on "why are people updating". If one person is much slicker, and is much more prepared, and has nice "just so stories" to explain away certain bits of evidence, then ***holding everything else constant*** this should reduce how much you update in their favour. A more prepared person is more likely to win a debate on points/style even when they are wrong. Verbal debates just aren't a very good mechanism of accurate determination of truth, they have a lot of noise/aesthetic factors seem to have great influence on perceptions! You just shouldn't update that much when a priori there's no reason to think the best arguments on both sides will have been presented / equal scrutiny been applied to claims on both sides. (This of course blows up 100x when you see the slicker person getting away with clearly erroneous claims).
The claim that you shouldn't have high Bayes factors on both sides of a debate is just false.
Every event with sufficient detail involves a ton of unlikely stuff happening. Every time you flip 30 coins you get an exceedingly unlikely list of heads and tails.
Sure, if you understand the data to be the **complete** specification of the even then yes you don't get high Bayes factors on both sides because you only get one. But that's not what's going on here. They are both cherry picking the aspects of the situation that most favor their hypothesis so of course you get high Bayes factors on both sides.
I don't have anything to add on the COVID origins topic, but I do want to register some (possible?) disagreement with this bit:
"For example, suppose I win the lottery, I’m told I win the lottery, the lottery company gives me a big check, I cash the check, and I become rich. Given that there were 1-in-100-million odds against me winning the lottery, the lottery company giving me the check and so on must be at least 100-million-to-1-level evidence - otherwise I should refuse to believe I really won the lottery, even as I enjoy my newfound wealth!"
The quip-level counterpoint: You *shouldn't* believe that you won the lottery, because which is more likely -- that I won the lottery and cashed the check and got all this claimed evidence, or that I'm *dreaming* that I won the lottery and cashed the check and all the rest? After all, I've had several dreams where I won the lottery before, and I can get just about any evidence you like in a dream (all made up by my brain, of course). Sure, a 10^8:1 update is possible in this case (er, maybe, I'm not sure because it sort of hinges on things I think I know about dreams but which might be false -- after all, I might be dreaming I know them), but the strongest counter hypothesis is this sort of out of band thing (dreaming, or maybe something like solipsism or a "living in a simulation" or "being in The Matrix" or similar roughly-equivalent thing) rather than something like an elaborate prank or a mistake.
Yes, on at least one occasion I have realized that I was dreaming by applying this logic.
This might seem academic or silly -- how do I **really know** that I'm not dreaming my whole life, being continually deceived by Descartes's demon, etc. -- but I'm pretty sure that similar caveats ought to apply to trying to reason about some things that people care about, like "Is there a God?". If you think you are more than 99% sure you know the answer based on Bayesian reasoning, you are being overconfident that you (a) understand the various conditional probabilities involved and/or (b) have properly considered all the relevant hypotheses.
Well, it's pretty obvious that I haven't just been dreaming my whole life because the human brain is simply not powerful enough to simulate the level of consistency that "reality" offers. Of course, that would be making a pretty big assumption that it's a human brain that's producing the dream, but there just isn't any rational incentive for any potential beings to produce such a simulation. ...If they wanted to torture me, they could certainly do far worse.
> Well, it's pretty obvious that I haven't just been dreaming my whole life because the human brain is simply not powerful enough to simulate the level of consistency that "reality" offers.
But within a dream (or within a dream-within-the-dream, if we accept the hypothesis) we generally don't take notice of the inconsistencies.*
So if you were dreaming your whole life, it may well be full of inconsistencies which you simply haven't noticed - but likely will once you wake up.
*Sometimes we perceive noticing the inconsistencies as causing us to wake up (as frequently depicted in movies and television - there's a House M.D. episode that immediately springs to mind, and I vaguely remember the Sopranos using the motif as well). In reality, though, it's the opposite - our conscious mind waking up causes us to notice the dream's inconsistencies.
I think that most people overestimate how confident 99% is, or even 99,999,999-to-1. Those are both big, but you can reverse the first by observing 7 coin flips all go the same way, and even the latter just needs 27 coin flips. You're unlikely to find this evidence if the thing you thought was likely is in fact true, but it's actually not all that much evidence you'd need to see to convince you of the a priori unlikely thing.
I think this is a general case of the principle that anyone in a weird enough situation should think their whole life is a dream or simulation. For example, I think all heads of state and billionaires should think this (though luckily most of them haven't reasoned it through, or else politics would get very weird).
I'm not sure if my situation (famous blogger) is enough to get me in this position; I'm about 50-50 on this one.
Isn't just being a human after the industrial revolution (and just before the singulairty maybe?) already a very strange situation, rather than being a trilobite during the Cambrian or something?
I don't want to divert attention from the debate, but since it seems appropriate, I'd like to share a post I wrote nearly three years ago proposing an hypothesis that is not exactly lab-leak, though not incompatible with it, and not zoonosis (at least as I've seen it stated):
The propagation of Covid-19 seems to point to certain areas of the world being less affected by the disease for no particular reason (no younger population structure, no less obesity, etc...). The main area being Southern China and Northern SEA. This is precisely the area where you would expect zoonosis for a bat coronavirus to happen. Other geographic areas correspond to places where infected people might have first, and in greater numbers, arrived if they had been infected by a previous virus in the aforementioned main area. This and other evidence (see full post) point to the possibility that a previous ancestor (to SARSCov2) virus emerged and propagated in that area, before mutating into SARSCoV2.
The competing 1/10,000 claims about the wet market are interesting, but feel too speculative without enough hard evidence for me to know what to do with. Far more salient and intriguing to me are:
Saar's claim that the 12-nucleotide furin site comes out of nowhere...
> COVID - which mostly just resembles BANAL-52 with a few scattered single-point mutations - has twelve completely new nucleotides in a row - a fully formed furin cleavage site that came out of nowhere.
vs. Peter's claim that the the furin site doesn't use the expected sequences for such a site and actually uses sequences that wouldn't be expected to work. Plus Peter's claim that WIV didn't have the technology or expertise to insert such a site anyway
> COVID’s furin cleavage site is admittedly unusual. But it’s unusual in a way that looks natural rather than man-made. Labs don’t usually add furin cleavage sites through nucleotide insertions (they usually mutate what’s already there).... COVID’s furin cleavage site is a mess. When humans are inserting furin cleavage sites into viruses for gain-of-function, the standard practice is RRKR, a very nice and simple furin cleavage site which works well. COVID uses PRRAR, a bizarre furin cleavage site which no human has ever used before, and which virologists expected to work poorly. They later found that an adjacent part of COVID’s genome twisted the protein in an unusual way that allowed PRRAR to be a viable furin cleavage site, but this discovery took a lot of computer power, and was only made after COVID became important. The Wuhan virologists supposedly doing gain-of-function research on COVID shouldn’t have known this would work. Why didn’t they just use the standard RRKR site, which would have worked better? Everyone thinks it works better! Even the virus eventually decided it worked better - sometime during the course of the pandemic, it mutated away from its weird PRRAR furin cleavage site towards a more normal form.
Peter's claims seem stronger to me, and I didn't see any response to these specific claims in the summary (although I may have missed), but I also can't explain Saar's claim, nor do I have the expertise to assign probabilities to any of this. But these claims seem to me to be the most important and what I would love to see further discussion about.
I'd put basic internet research on the subject as "very low reliability", so, take this with a grain of salt, but:
Some internet research suggests that the PRRAR furin cleavage site is better-adapted to animals (ferrets in particular) that were used as test subjects at the laboratory in question (but poorly adapted to bats and pangolins), and the RRKR cleavage site is better-adapted to humans.
Interesting, thanks! In which case, Peter's claim could possibly (accounting for salt grain) fall under the all-too-common "Simple Absolute Claim About How Humans Always Do A Thing Turns Out To Overlook Real-World Nuance"
It seems entirely plausible that the WIV was trying out a new type of furin cleavage site, to examine how it worked. It's evidence against SARS-CoV-2 being a deliberately-created bioweapon, not against it being part of a research project about how animal-borne diseases can jump to humans. In my own field I've often copied something new and interesting that I saw in the wild, just to see how it works and improve my own understanding, and I'm not even a research scientist.
I just wanted to respond to this point in the conclusion:
"Cool Machiavellian plot you have there, but maybe the fact that you’re losing 16%-66% should make you question whether you’re really as smart as you think you are...At some point you have to start debating!"
The reason that people in the field might be reluctant to accept this point is that they don't, and shouldn't, give a shit what 66% of the American public think. And debating just gives the American public the impression that it has some kind of a right to have a say in what the truth is. If you're a virus researcher now, working on how to predict and prevent the next zoonotic infectious epidemic, you really really don't want the public deciding to shut you down or order you to focus on lab leak prevention.
This is obviously very problematic - I'm kinda making the argument for important things to be decided in smoke-filled back rooms. But it's also unavoidable, I think. Science must by its nature be difficult and cutting-edge and unknowable to most people; and scientific advances are by definition decided by something other than human opinion - ideally, by what the universe is really like. So there is just an inherent limit on the level to which science can be democratised. Which means that noting which way public opinion blows is unlikely to be very relevant to the people who want to advance science.
That said, I do think that science is getting more democratic, and that that's a good thing. It's just a messy process.
"If you're a virus researcher now, working on how to predict and prevent the next zoonotic infectious epidemic, you really really don't want the public deciding to shut you down or order you to focus on lab leak prevention."
A question in good faith: Is "working on how to predict and prevent the next zoonotic infectious epidemic" really what this field is doing? We just had a major infectious epidemic. Did work in this field help predict or prevent it?
At UC Berkeley in the 1980s (and so presumably lots of other places) there were loads of scientists talking about the need for pandemic prevention research. The people who have been doing that kind of work for decades have been working without adequate funding. None of these people are surprised this happened. The fact it happened as they predicted it would is not evidence for their work being useless.
Millions of people were killed or disabled by polio.... until they weren't. We could make a list of many other things like this in the realm of scientific research.
Was there any research going on at WIV that was useful in predicting, preventing, or combating COVID-19? I'm sure there are lots of smart people spending lots of money studying viruses there and elsewhere. But what good have they done anyone.
I specifically ask, because it does seem like there's a >0% chance that the research itself is actually increasing the risk of pandemics.
I feel like the best answer we can give to that is: look at the historical record. Are there in fact more pandemics now than there used to be, or fewer? I think it's fewer, because many serious diseases like polio have been wiped out in the developed world, and flu has been controlled. You're right about that non-zero probability of science making things worse, but the facts seem to suggest we're moving in the right direction.
There is always a nonzero chance that scientific research will make some things worse. We have a ton of examples of this. Thalidomide given to pregnant women causing tons of children with birth defects. There are so many examples.
Do you think vaccines have been not worth it? Antiobiotics? Everyone studying infectious disease treatment is working with diseases in labs. Would you have them stop?
We have a risk of pandemics built into our ecology on this planet. So far as we know. all the pandemics we've had historically were from nature. If this one wasn't, it was very likely the first one like that. We know up to this point that there have been lab leaks that did cause deaths, you can go read the wikipedia page about it. Those deaths are dwarfed by almost any other cause of death you can think of.
Right, but specifically WIH and GoF though? Your argument seems to be "science is mostly awesome, even science with bad side effects, therefore this specific scientific endeavor with bad side effects is awesome".
Can you point to one concrete way in which having GoF research helped us significantly in handling COVID-19? If not, then you better be damn sure lab leak not only didn't happen but is extremely unlikely to happen (almost certainly false).
What would one like to see happen? The possibility that GoF viral studies are strongly net negative seriously evaluated and if not unlikely enough, GoF studies stopped or much more tightly controlled.
One would also like the people in those "smoke-filled rooms" with their superior expertise or whatever not to engage in mass misinformation. If they do because they think that accurate information might lead to just the results above - restrictions on their research - who am I to argue with their expert opinion?.. Let's consider having those restrictions on their research.
Because I agree with what you're saying, I think I potentially misread the intent of DavidK74 and heard him asking a broader question like what's the use of research that happens in labs that have diseases and infectious agents in them.
I also agree with the need for oversight and accountability for research, absolutely. I just don't want that to happen via general public opinion. I can't remember but I think the original thread of this discussion was off a side comment Scott made about public opinion.
I think it's good to have lots of arguing with expert opinion, including multi-disciplinarily (if that's a word).
I'm not in this field, so I can't claim any deep knowledge. But yes, clearly some parts of the field do do that. I think they did indeed predict more SARS-like outbreaks following SARS in 2003, and I'm sure did some work to try to work against them (those models that they kept wheeling out early in the Covid pandemic were created long before it struck). I think the work on vaccines made a big difference in the Covid epidemic.
I was really making a more general point, that science and technology don't advance through democratic processes. Even the most perfect democratic process couldn't actually help scientists uncover truth more quickly. And real democratic processes are not perfect; they are political. If science starts worrying too much about what the public thinks, then it just opens itself up to political meddling.
FWIW I asked people with expertise in domains somewhat relevant to mRNA vaccines and their opinion was that no, prior research in institutes such as NIH did not materially help in designing those vaccines. I'm not an expert - I could have misunderstood. They're not literally the team/ teams involved in various stages - they could be wrong. Other objections might exist.
But it's not obvious that the burden of proof is not on you with the claim that they "clearly" helped. Said burden could be met of course :)
For the more general point - 60% doesn't mean the dumbest 60%. If so many perfectly intelligent and qualified people strongly disagree with you - I do think it's your duty to take care of the messaging and convincing. As opposed to, say, deliberately and consciously lying.
I mean. The debate about usefulness of oversight over those "smoke-filled rooms" has to take consequences into account. If specifically GoF research has strictly and severely negative net value for society, would you expect this evaluation to be made, published and pushed by experts in the field? We know the answer to that. They explicitly lied about this.
If this is a line of research contaminated (har har) by conflict of interest, then public attention to and comprehension of key arguments seems essential.
Politics and meddling come in different forms. One is what you alluded to. Another is endemic to "smoke-filled rooms" - and against this one, just like for some viruses, sunlight is the best disinfectant.
I think there's two different issues here. My understanding is that a lot of gain of function research is basically about bioweapons. If that's right, then shining sunlight on virologists isn't going to make any difference at all. Virologists aren't the once who take decisions in bioweapons research. It's the army guys. And they don't accept public scrutiny.
My point was about what the field of virology should be doing right now. I assume that their research paths have been affected by Covid. Scott was suggesting that they should get out there and debate the public to make the case that it probably wasn't the lab leak. I think they have good reason not to do that: if they go out and start debating these points, they only risk increased political interference in their research decisions.
That increased political interference won't make any difference at all to gain-of-function research, which will or will not carry on under strict secrecy at the whim of the army guys. But political interference could make a difference to the work of mainstream virologists, particularly if public pressure ends up pushing grant money towards lab leak research projects. That seems to me by far the most likely political outcome. It wouldn't achieve the sunlight that you want (which is mostly achieved through open science publication, anyway); and it could send research up a bunch of blind alleys.
I think decisions about what research to fund needs to be made by people who are familiar with the domain of research in question. I don't think we need to think of this as smoke-filled backrooms. I don't want the public deciding which strand of breast cancer research shows most promise. I don't want decisions about research into the rise in colon cancer in young people being made by public opinion about whether genetics, environment, or having gotten the MMR vaccine is the more important line of research.
That this particular field of research has done much to lose its credibility to do what's right, probably doesn't have much in the way of practical benefits to show for it, and thus decisions regarding it should have a greater outside involvement. Which, in turn, requires a greater understanding of the underlying issues. Which, in turn, means at least that defending and explaining their point-of-view is the duty of scientists in the field, rather than shrugging at the deplorable ignorance of everyone who disagrees with their estimate of lab leak probability.
This is not remotely my field. I hear you speaking about arrogance in a sub-field that has questionable value. I can confirm I see this over in my corner of psychology.
This is not right. We can't assume *perfect* benevolence and *perfect* lack of conflict of interest. They're human, after all. But that isn't a minimal assumption. It's a strawman level assumption. Minimal benevolence and minimal lack of conflict do seem warranted - by, for example, all that paperwork from WIV that people have pored through. That paperwork exists because even in nasty old China, scientists who may or may not be interested in creating biological weapons still observe basic safety protocols. That shows minimal benevolence to the rest of us.
My whole theme here is that information looks different inside a specialist field and outside a specialist field. I'm a teacher, and I "lie" to the children I teach all the time. I do it because they're not yet ready to process the full complexity of English grammar yet. I would expect virology experts to say different things to people inside the field than they do to people outside the field. So when we observe them doing that, I don't really see it as a "lie". For example, it would seem perfectly reasonable to me if they chose to communicate to the media the *most* likely origin story as understood by their field, i.e. wild animal origin; while internally discussing the less likely but more dramatic possibility of a lab leak. That doesn't sound mendacious, it sounds like normal professional behaviour.
But I could be wrong about that. Perhaps a significant number of virologists (not all, remember they're not a monolith) behaved badly in the immediate aftermath of the outbreak. Even if they did, that absolutely does not invalidate the whole field of virology, or suggest that virologists need to be put under greater public scrutiny. You'll recall that the president of the USA at that time suggested drinking bleach. This does not imply that all presidents should henceforth be subject to stronger rules on what they can and can't say.
The lying itself doesn't invalidate the field. The possibility that the field is responsible for COVID-19 does - unless there's proof of some benefits to outweigh this possibility. Which is why I brought up the field's questionable contribution.
The lying invalidates the a-priori trust in the decision-making that the field enjoyed beforehand. If there's a real chance the field is a net negative, and I know for sure that such possibilities are exactly what you might be lying about, you don't get to tell me "oh trust me I'm a scientist" without so much as the courtesy to even show up to debate something that tons of intelligent people disagree about.
Two more general points.
Firstly, I think you're under-playing the role and duty of science in education and enlightenment of society. If the origins of such a world-changing event are of interest to the public, and the opinions of the public (including the generally well-educated public) strongly disagree with scientific consensus - it is a failure.
Secondly, I think you're overestimating scientists. I did a PhD (in an entirely unrelated field) and am married to a post-doctoral researcher in a biology-flavored field (involving RNA, as it happens) and I had many opportunities to see how the sausage is made. It's not pretty. And it's not pretty exactly in a way pertinent to our discussion. Ego, desperate need to publish or to secure a grant at the expense of quality, integrity, basic decency. I absolutely do not believe that their internal and opaque communications are more likely to be truth-seeking and productive than anything exposed to broader scrutiny.
Some aspects of modern science are profoundly broken, and there's every reason to believe that viral GoF is not one of the healthier ones. So when there's a reasonable debate about its usefulness, they do not get to go "please play quietly in the corner while we brilliant grown-ups decide matters". They lost this right, if ever they had it. They get to show up and at least explain.
I'm against using a text format for future long-form debates like this.
I watched this debate an a second monitor while doing other non-cognitively-demanding things like working on design projects and playing video games- and I enjoyed it a lot more than the podcasts and audiobooks I usually use for that purpose. In contrast to those, nothing about the debate felt like dull filler, and yet, unlike most really demanding books, it was possible to miss bits here and there without losing the overall thread of the argument. It's almost a perfect format for content to listen to while slightly distracted, and I think that for that reason, debates like this can provide a ton of value to people beyond just shifting their opinions on the topic.
Granted, a text format would probably result in slightly better arguments- though I don't buy that Peter ambushing Saar with new claims was a major factor in his loss; the lab leak side had plenty of time to research claims between debates, and plenty of time to respond later. A book-length written debate, however, is something I expect almost nobody would be willing to read, since it doesn't provide that extra value as good content to consume while doing other things. It would also require a lot more effort on the part of the debators, so I'd expect there to be both fewer long-form debates and much less engagement- and I don't think a slight bump in quality would at all make up for that loss of value.
Oh interesting argument, I like that. You're saying it's worth losing some potential accuracy in outcome on the issue in question in exchange for the public educational value of it?
I don't think very many people are ever going to watch this kind of very long form debate and so I have a preference for thinking of it as a kind of science project on its own and that maximum accuracy is better for that. And then the NYT or Nature or bloggers or whoever can talk about how this very sophisticated well-run debate with very knowledgeable people happened and they concluded X instead of Y, and they basically do a book review of it.
Education, definitely, but also entertainment- which is a less high-status sort of value, but still real value.
Well-produced podcasts and hours-long video essays will often be seen by very large numbers of people, even when they dive pretty deep into technical questions. Dwarkesh Patel's two-hour-long interview with the Anthropic CEO, for instance, has over two hundred thousand views, and wasn't much less technical than this debate. I also found it a bit less engaging- while a good interviewer will challenge an guest's ideas to some degree, they'll never do so as thoroughly as an opponent in a debate. Ideas are not only more relevant, but also more interesting when you have good reason to trust that they're not being presented misleadingly- and while in a polite conversation, everything has to be taken with a grain of salt for that reason, misleading statements in a long debate tend not to survive the response.
I have a feeling that the low view count of this debate has more to do with people not being familiar enough with the format to think of adding it to their usual rotation of long-form audio content- plus maybe the fact that it's usually been described as a single 18-hour debate rather than a series of 1-3 hour debates. If this sort of thing was a lot more common, I think we'd see some series with far higher view counts.
Agree and I think Saar should take credit for coming up with a compelling format for debating complex issues like the origin of COVID. True he lost on this particular issue but the outcome was consistent with the majority view of experts in the field based on recent polling.
If there were a series of debates like this I'd expect Rootclaim to win some and lose others but either way it's much more compelling and will reach a wider audience when you get to actually watch the two sides debate.
Fantastic post - I've shifted my opinion quite a bit in response.
One thing that sticks for me is this: it didn't take long to find the civets responsible for the earlier SARS outbreaks. Why has this proved so challenging here?
The troll in me wants the answer to be that it was zoonosis that occurred in a lab, and then it leaked out of the lab.
"The troll in me wants the answer to be that it was zoonosis that occurred in a lab, and then it leaked out of the lab."
I've thought about that, too. If that happened, what would we call it - zoonotic or lab leak?
And what if the virus evolved through inadvertent natural passaging in the lab? Lots of weird animals in cages, lots of humans around with not much in the way of safety precautions, plus thousands virus samples collected from all over the place being handled. Sounds like a real witches brew.
Thus, a natural emergence, in the sense that it wasn't the result of any deliberate human manipulation. But also something that would not have happened if the lab hadn't been there doing the types of things they did.
The Chinese government prevented all research by outside investigators for a year, which made it very hard either way.
I'm a bit surprised that they didn't consider the zoonosis-in-a-lab hypothesis, because that's the one that I had thought was most probable before I went back to the wet market hypothesis.
Thank you for this excellent post. It's one of the best things I've read on the internet ever.
I remain in the Lab Leak camp, where I started. I'm of course not 100% certain, and I don't think it's possible evidence would ever emerge that would make me 100% certain. I really don't like people who claim Lab Leak is a "conspiracy theory" and am inclined to discount everything they say.
A question:
Should we accept claims by the Chinese Government and technical/academic/policy experts in this field at face value? Aren't they all hopelessly conflicted? Even if what they say is logically sound, aren't there myriad ways in which they could be fudging the analysis a bit, even unknowingly? What value have researchers doing gain-of-function work on viruses provided to the world? What is the likelihood someone in the field is going to publish findings saying the risk reward in their field just isn't worth it and all their skills are worthless because the work should be shut down?
Another question:
What do those maps of the outbreak in the wet market purport to show? The location where someone was exposed to COVID (how can that be known)? Some point that a person who got infected passed through (what about all the other locations they visited)? Whatever they claim to be showing, how accurate could their measure be? Based on interviews? How hard would it be to fudge the data a bit - for example, an infected person tells a researcher all the places they've been in the last week, and of the dozens of locations the researcher chooses the wet market to record.
I agree with you about this: " I really don't like people who claim Lab Leak is a "conspiracy theory" and am inclined to discount everything they say."
First of all- awesome post! I loved all of this. Saar for setting up this challenge, Peter for taking him up on it, the judges for being fair, and you for summarizing the results- all great stuff.
As for the weird coincidence of the Wuhan Institute of Virology being located so close to the epicenter of the breakout- is it? Everyone knew that wet markets were a potential risk center for spreading diseases. See for example: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7141584/ about the 2003 Avian Flue outbreak. That's the same year that China approved construction of a biosafety level 4 lab at WIV.
They built the lab there because it was a good place to gather research data, and maybe also because the local government was strongly interested in that sort of research to protect themselves. It's the same reason that the "Center for the Study of Active Volcanoes" is in Hawaii, the "Earthquake Engineering Research Institute" is in California, and the "Hurricane Resilience Research Institute" is in Houston. Of *course* they are there where the threat is greatest. And gain-of-function research was just the most effective means of researching that kind of virus, if you want to research it before it actually starts spreading.
You might ask, if they knew the threat of wet markets, why didn't they just shut them down? My understanding is that China doesn't have that kind of state power. There are too many poor, rural people who really want to eat meat, and still can't afford ranch-raised animals. If you try and shut down the wet markets they'll just start selling them in secret someplace even worse.
Nitpick on that poll asking Americans if they think lab leak is true. (Before this) I'd have had to answer this as "probably true" but my probably was that I was at like 60% lab leak and 40% zoonosis. 40% chance of zoonosis is still pretty dang high and doesn't really seem that indicative of acceptance of the bad epistemic edifice that accreted around lab leak. I mean 30%ish saying it's "definitely true" is a problem but "probably true" seems to be doing a lot of work.
Also, good post. Quite markedly shifted my opinion.
*Edit* Also, both these posts and that parliamentary procedure kid that's been mentioned so much recently has made me think that the best model for thinking we have isn't heuristics or Bayesian reasoning or the wisdom of crowds. It's probably just smart guy who does like a ludicrously unreasonable amount of homework.
> Zero COVID era Chinese outbreaks were concentrated in wet markets because they received infected animal products. We know why there was an outbreak in the Xinfadi Market in Beijing: it was because the seafood stall got frozen fish from some non-Zero-COVID country, the fish had COVID particles on it, and the vendor got infected and spread it to everyone else. Something like this is true for the other Chinese wet market based outbreaks we know about it. So this makes the opposite point you think it does: wet markets start outbreaks because there are infected goods being sold there. Then the virus spreads through the wet market at a completely normal rate.
I definitely know that China ascribed COVID outbreaks to things like "the person in the plane on the previous flight left covid residue". China was desperate to explain every outbreak so they could continue attempting to achieve a horrible and impossible goal. I don't actually believe that is how the virus was spreading? In particular, surface contact spread and the idea that you should wet wipe your boxes quickly fell out of style. I think the ubiquity of spreads getting first observed at wet markets during covid zero is strong evidence that even if covid is zoonotic but first entered the population by like, bats during a vacation, you would still have first observed it in the wet market.
I filled this out with others' hidden, and ended up weighing that even less than Saar did, which probably means I'm overly dismissive.
I think that surface contact spread and the idea that you should wipe your boxes very naturally fell out of style, because it's probably several orders of magnitude less likely a mode of transmission than airborne directly from a person. But when you're in covid-zero conditions, and the rest of the world is in full pandemic, then suddenly "several orders of magnitude less likely" is still the only way for anyone to get infected.
Sort of like how, to a first approximation, you can think of a gas stove, or a candle, as basically having zero respiratory issues compared to a wood burning indoor stove. But once you've eliminated the wood burning indoor stove, those last few indoor flames, even if they're very clean, may be your main respiratory risk.
"But when you're in covid-zero conditions... then suddenly "several orders of magnitude less likely" is still the only way for anyone to get infected."
Are there any clearly documented examples by a trustworthy source (e.g. not the CCP, i.e. in another country) of infection due to residue on frozen package after international transport? Just off the top of my head, Australia, New Zealand and Taiwan all had "zero-covid" regimes for an extended period and had multiple outbreaks. None afaik was due to frozen packages. They were all "failures in the quarantine process" of one sort or another. It's pretty suspicious that China would be the main source for "spread driven by imported frozen packages" claims given tha is reasonable to t this is the sort of thing that perfectly aligns with their desire to shift blame for Covid onto someone else (even more narrowly, ignoring that motive, "our quarantine of incoming travellers failed" is much more egg on face than "we were tricked by those dastardly frozen food packages coming from those imperialist bastards in XYZ").
And thus, like Robert Kennedy, I think it's reasonable to infer from this that large public markets are probably decent locations for superspreading. (Indoor location with a lot of foot traffic is a prime potential site for a virus outbreak - doesn't seem shocking).
Good point! The Australia and New Zealand comparisons seem important. If their outbreaks were all traceable to quarantine failures, then that’s pretty good evidence.
Thanks. To be clear / maintain epistemic humility,
(a) I may be misremembering - every outbreak I know of is either believed to be due to / was traced to some quarantine infection/escape, but my memory could be faulty and my recollection of outbreaks is not exhaustive. (e.g. worker in quarantine hotel caught it from infected quarantined traveller, then took it home with them etc).
(b) logically there should be outbreaks that happened that the authorities never noticed (e.g. one person is infected by some source, they never infect anyone else - which happens >50% of the time - so the outbreak dies before it is noticed), about which we naturally know little
(c) there was an outbreak in New Zealand that initially had some speculation about imported frozen food transmission (if I understand correctly, justified on "well, the Chinese are claiming this can happen), but this was ruled out within a few days.
I hope this isn't a derail: In one of the previous ACX posts' comments, someone posted a link to a page explaining the case for gain-of-function research. I'm embarrassed that I didn't either follow the link or bookmark it. Does anyone have that link? Many Thanks!
Way back in 2020 there was an article https://www.independentsciencenews.org/commentaries/a-proposed-origin-for-sars-cov-2-and-the-covid-19-pandemic/ , which I read after George Church tweeted it (!) (without comment or explanation). Their proposal (they call it "Mojiang Miner Passage" theory) in brief was that it WAS a lab leak but NOT gain-of-function. Rather, in April 2012, six workers in a "Mojiang mine fell ill from a mystery illness while removing bat faeces. Three of the six subsequently died." Their symptoms were a perfect match to COVID, and two were very sick for more than four months.
The proposal is that the virus spent those four months adapting to life in human lungs, including (presumably) evolving the furin cleavage site. And then (this is also well-documented) samples from these miners were sent to WIV. The proposed theory is that those samples were put in a freezer at WIV for a few years while WIV was constructing some new lab facilities, and then in 2019 researchers pulled out those samples for study and infected themselves.
I like that theory! I’ve like it ever since 2020! It seems to explain many of the contradictions brought up by both sides of this debate—it’s compaible with Saar’s claim that the furin cleavage site is very different from what’s in nature and seems specifically adapted to humans, but it’s also compatible with Peter’s claim that the furin cleavage site looks weird and evolved. It’s compatible with Saar’s claim that WIV is suspiciously close to the source of the outbreak, but it’s also compatible with Peter’s claim that WIV might not have been set up to do serious GoF studies. It’s compatible with the data on differences between COVID and other previously-known viruses. Etc. (As I understand it.)
Old as this theory is, the authors are still pushing it and they claim that it’s consistent with all the evidence that’s come out since then (see https://jonathanlatham.net/ ). But I’m sure not remotely an expert, and would be interested if anyone has opinions about this. I’m still confused why it’s never been much discussed.
Update: I read the talk page! My understanding is that the miners got sick from Disease Agent A, and then later someone took a sample from the same cave (NOT from the miners) and found Disease Agent B. The Wikipedia article you found is about B, not A.
It’s POSSIBLE that A=B, but neither the Wikipedia article nor anyone else is claiming that A=B. There’s actually some evidence that A≠B, but that evidence didn’t make it into the main wiki article, partly because the WIV master’s thesis discussing the miners was declared not to be a “reliable source” per Wikipedia standards, and partly because the anonymous person who created the article was very aggressively opposed to providing fodder to what they called “baseless sinophobic paranoia” “conspiracy theories” 🙄
Thanks! That link isn’t much evidence though, right? Like if we’re gonna do the Bayes thing, I dunno, likelihood ratio 1.1 maybe? Like, if a biologist is accused of stealing a car, and then the biologist announces that he did his own DNA analysis in his own lab and found results that exonerate himself … well, we haven’t learned very much from that. :-P
I’m much more swayed by e.g. the fact that RaTG13 is from the Mojiang mine.
Let me know if I’m misunderstanding or if you find anything else :)
Yes, they could easily lie here. You have to decide for yourself what weight to put to those words. Johnathan Latham explicitly says they’re lying in https://jonathanlatham.net/faucis-covid-origin-swat-team-versus-the-mojiang-miner-passage-theory/ ("Other criticisms of the MMP theory"). I’m somewhat sympathetic to that : "they died of fungus" is clearly nonsense and can easily interpreted as the researchers saying to us "we have to lie to you but we’re not happy about that so we are going to pick the least believable lie".
The big gap in the MMP theory is BANAL-*. I hope you’ll forgive me if I can't find the exact quote, but in the initial MMP theory Johnathan explictly said something in the line of "one thing that will be a big problem for the MMP theory is if the virus hunt yields a genetically closer ancestor to SARS-CoV-2 than RaTG13". And we found one : BANAL-52. Johnathan handwave it by saying "we found it in north Laos, and north Laos is geographically the same thing as Mojiang".
But we don’t know where exactly the mine is in Mojiang county (eyeballing google maps, it’s 300 km from north Mojiang to south Mojiang). More problematic, the paper that announced the discovery of BANAL-52 never said it was found in north Laos ; indeed they say that all BANAL clades were found in the same site, "site 1", which was in east Laos, close to (again, eyeballing the map) Pak Lay, which, again by eyeballing Google Maps, is close to 800 km to Mojiang county.
This is a big problem for MMP theory, I haven’t seen a good answer to that.
But to be fair, it’s a big problem for the Wuhan wet market theory too. It’s mostly a point for the Chinese official theory, "it comes out of China".
This was incredibly informative. Thanks to the participants for doing it, and to Scott for summarizing and publicizing it. I hadn't thought about COVID origins in a while, but I would have told you it was in the tossup zone, and digging into this a bit puts me closer to Scott's 90-10.
I wish this was a highly replicable method for illuminating tricky topics, but I kind of suspect that the conditions here (two reasonable and sincere people who believed an impartial jury of smart, curious neutral observers would be won over to their side, and who were able to agree on who constituted a smart curious neutral observer) are tough to arrive at. I hope I'm wrong!
Not to be too facetious, but perhaps Rootclaim should have calculated the probability of the existence of a debate savant before building an attractor for one.
The big problem is that we still have very few details about all this. How is that possible after all these years?
The Chinese authorities haven't been very transparent. Other governments haven't demanded transparency from China. Why not? They just accept that China doesn't want to give any details.
The pandemic was a global catastrophe. It killed a ton of people and caused enormous economic losses.
The people deserve to know how it all happened. Secrecy is unacceptable.
The main reason why I'm leaning towards the lab leak is the severe censorship about the topic. That came from the individuals who strongly supported the zoonosis hypothesis. Usually people who desperately want to hide something have something to hide.
I think most leaders know it was a lab leak. They just don’t want to deal with the political shitshow that will come out of making it official. It has the potential to disrupt the fickle relations with China and as a result, significant amount of global trade
Whether or not it's zoonosis or a lab leak doesn't actually matter. Either way the CCP is indirectly responsible. The CCP will never admit wrongdoing as it has zero incentive to do so. Every other country is obviously upset about this, but not enough to stop doing buisness with China, as that would put them at a major disadvantage. While COVID did cause damage, and it damaged the entire world pretty much equally, and thus most countries are not in a worse position than they were before, relatively speaking. Economies can be restored, and human lives can be replaced.
As for the gain of function research... do you really think that even if there was conclusive evidence in favor of a lab leak, it would change anything? Bioresearch is just one of those things where the positive consequences are localized and the negative consequences are not, and thus there is no incentive for self-regulation. Any attempts at globally regulating it would be just as futile as nuclear disarmament.
Nobody really demanded anything. They just asked nicely and China said no. And that was it. There was no real pressure against China to provide information.
Everyone else was happy that China didn't share any details. That is a bit suspicious.
No one was happy that China didn’t share any details. I’m not sure where you were getting the idea that they were. The WHO did try to put pressure on China, but China has more power over the WHO than vice versa, so they just couldn’t put all that much pressure on them.
Nobody put any REAL pressure towards China or any other suspicious party like EcoHealth Alliance.
WHO put extremely weak pressure on China, it's like they didn't even try.
I don't remeber seeing any politician being genuinely enraged for China hiding so much important information. Given that the negative effect of the pandemic was enormous, I would have expected to see lots of powerful people saying "you must show us everything you have on this topic, or else...". But nobody did that.
Of course it's not only WHO but all the governments.
The pandemic had an enormous negative effect on the population of the world. People deserve to know what happened. The governments who represent the people should find out what happened but for some reason they don't even try.
If it's not enough to ask nicely, it's not hard to come up with harder ways to create pressure.
For example, setting some sanctions would definitely help. Similar to which are in effect against Russia now. They would be even more effective against China because it's so depending on imports from other countries. A few months of blocked fuel imports would give a nice incentive to start being more open.
“Hey, this bat we just collected on one of periodic field trips to yunnan had an interesting viral strain with some really interesting structure around how it’s furin site works and implications for protein folding”
“We should probably research it, is there any grant money we can use?”
“I think the Americans just announced some new initiatives?”
“Great, let’s apply. Write a proposal that says we will investigate Furin site placement and it’s impact on virology.”
….
“Did we get the funding?”
“No”
“But the grant reviewers didn’t know about the bat, right? Just that we proposed doing the furin site stuff?”
“Yea, we’ve changed the names of the ones we found and generally obfuscated that we have this specific one in house”
“Should we just do the research anyway?”
“… … … yea, probably, could be interesting”
“We haven’t really done this before, and maybe we don’t have the right safety procedures and protocols, specifically to make sure it doesn’t escape?”
“No. But this particular strain can’t even last that long in the lab on its own. So what’s the worst that could happen”
I have no strong opinion about the Lab Leak hypothesis. I have a very strong opinion that Bayesian reasoning does not, in practice, help with deciding such matters.
Early in the post, Scott offers the following. After reading the entire long post, I believe that it is, in fact, the correct overall summary:
> But the joke goes that you do Bayesian reasoning by doing normal reasoning while muttering “Bayes, Bayes, Bayes” under your breath. Nobody - not the statisticians, not Nate Silver, certainly not me - tries to do full Bayesian reasoning on fuzzy real-world problems. They’d be too hard to model. You’d make some philosophical mistake converting the situation into numbers, then end up much worse off than if you’d tried normal human intuition.
Disclaimer: I have not watched the full videos. Apologies if this point is addressed at length.
To what extent are both parties relying on CCP-curated information for their data on the early cases? Was there a discussion of how (un) trustworthy that information might be?
I think at least some of the consideration is that whatever curation was going on in December 2019 is unlikely to have any evidence about the actual origin, since the CCP had no reason to suspect it was related to the virology institute.
If you assign 1/10,000 probability of that particular wet market being the location of the first major outbreak conditional on the lab leak hypothesis, you should also calculate the probability of that particular location under the zoonotic origin hypothesis. The population of Wuhan is <1% of the population of China (maybe the whole world is a better reference class, though rural areas and developed countries are going to have lower risk), for example, and that’s before you account for the fact that zoonotic origin could also cause major outbreaks at other locations within each city. All in all I’m not sure “covid began to break out at this particular wet market near WIV” is terribly strong evidence one way or the other.
Another piece of evidence I don’t see discussed here is the lack of openness from China, and relatedly the lack of evidence of a zoonotic source. If China had information against the lab leak hypothesis they would likely release it, so the relative lack of information we have is more likely given a lab leak imo.
For all the talk of Furin cleavage sites, there's one thing that confuses me. Could we have had a (far less deadly, probably not even noticed) coronavirus pandemic if there was no Furin cleavage site? My understanding of gain of function here is that there are multiple proteases that can do the cleaving even without Furin, but Furin does the job faster, and therefore sometimes outruns the immune system in a way COVID wouldn't do without that extra cleavage site. But that raises the question of why we got an outbreak of a virus with that Furin cleavage site, instead of a hypothetical precursor that lacks it. The existence of a Furin cleavage site is considered unusual, so for Zoonosis would you expect a precursor virus without it to be far more likely to cross over into humans first, and therefore prevent COVID from being a major issue in the first place?
Shi's team, in 2018, published a paper reporting serological evidence of bat-human coronavirus spillover around the caves in Yunnan province. Doesn't seem like this strain/these strains would've been infective enough to kick off a pandemic, but if this sort of transmission was happening regularly enough for long enough, then the people around these caves could've definitely served as a natural "gain-of-function" experiment.
The notorious fact is that there were only 3 labs in the world at that time doing gain of function technology in the world...2 of those were im US and 3rd one in Wuhan..GOF was banned by Obama administration so the only destination to outsource technology was wuhan lab...🤗🤭 even famous defuse grant for the almost same virus creation by GOF was proposed and refused for wuhan lab by darpa im 2018...immpossible coincidencenes happen😅🤣
"Still, it’s awkward to use “conspiracy theory” as an insult when the conspiracies were real. Maybe a better question is whether lab leak is “pseudoscience”.
The argument against: lots of smart people and experts believed it was a lab leak. There were all those virologists giving 50-50 odds in their internal conversations. Even Peter says he started out leaning lab leak, back in 2021 when everyone was talking about it.
The argument in favor: since 2021, experts (and Peter) have shifted pretty far in favor of zoonosis. They’ve been convinced by new work - the identification of early cases, the wet market surveys, the genetic analysis. "
I have admittedly not looked into the details, and I cannot claim to truly know how to handle filtered evidence, but:
If the experts (mostly) all agreed to have a conspiracy pushing for A, and then all the evidence that turned up after that pointed to A, this really doesn't make me feel like I have to raise my probability of A all that much. For all *I* know for every published result supporting A there are 2 unpublished results supporting B, or claims that have not been investigated because experts (correctly) believe they would turn out to support B.
One issue I have with the opening presentation in the first video (I have watched 38 minutes so far) is that he lists, for example, about a dozen different theories about how the virus could have been made, and his slide says, "it's not possible for all these theories to be correct." So what? If there's one true explanation for something, the fact that 11 other people came up with a false explanation for it doesn't diminish the probability of the first explanation being true. Otherwise, I could make the probability of any true idea go down just by creating a bunch of false theories.
What you should do it reserve some probability mass for the theories you have never heard of, or that no one has even thought of. The probabilities of the set of theories that happen to exist at time T don't have to sum to 1, in general.
Skipped the whole lot and am working purely from Scott's post-transcript summary.
>So the real p(wet market|lab leak) isn’t the 1/10,000 chance a pandemic arising in a random place hits the wet market, but the (higher?) probability that there’s something wrong with Peter’s argument.
This looks like an excuse to assign whatever numbers you like to any point. If you're readjusting part of your stats, surely you should be readjusting all the other parts with the same readjustment metric, and thus reach a conclusion indistinguishable from the original formula.
It does sound like Saar lost mostly because he got caught flat-footed, but considering he set up the system it still looks bad for his methodology. Surely he used it to calculate a proper debate setup beforehand, and it underperformed.
I learned more about this debate and the arguments in favor of each side in the hour and a half reading this post than I did in any previous discussion of these arguments. Thank you for the rich summary, and while I agree with your arguments about why this kind of thing probably can't happen very often, I really wish we'd see more of these kinds of rich debates
A systemized version of this plus community notes is my basic model for how all news should work. But with reputation scoring in there to smooth out some of these transaction costs for actors who have proven themselves credible.
“Dumb” question that I did not see answered anywhere, and not mentioned much at all even in these comments to date: does “lab leak” also incorporate the possibility that it was created in a lab and *deliberately* circulated in the Chinese public?
And if done deliberately is an option within “lab leak”, doesn’t that throw off the odds by several OOMs, given that for someone deliberately wanting the virus out there, but wanting to cover their tracks a bit, “releasing” it in a wet market that had all those animals would be a fairly high probability thing to do (in addition to being a very good way to ensure spread)?
Is deliberately put out into the world its own separate possibility, or is it incorporated into “lab leak”? And if the latter, what is faulty about my reasoning that it throws off the supposed unlikelihood of the outbreaks being centered in the wet market by many OOMs?
...Because that would be completely insane? Not the possibility of it happening, but... There is no potential rational justification for doing that, ever. It would just be plain and simple bioterrorism. Except it wouldn't even accomplish anything of value except kill a lot of people.
Like, I'm sure the probability isn't zero, but... people generally assume that the people involved in the situation aren't completely insane. And also it would be incorporated into the "lab leak" hypothesis, since the existence of the lab would still be responsible for the event.
Really? Seems to me all it would take would be a single Chinese bureaucrat - high, middle or even low-ranking (the last of which seems admittedly less likely) - who decided the benefits of the knowledge gleaned from human trials would outweigh the risks. And merely not understanding how high the risks were. The Nazis had “rational justification” for things we people with decent western liberal morals thought were far more ‘insane’.
And if this case is in fact covered within “lab leak” then it changes completely the logic that the wet market would be a 1 in 10,000 chance to something more on the order of 1 in 10…
...That would possibly be the stupidest, most psychotic way of doing human trials possible. And if the Chinese government was responsible, why the fuck would they spread it in their own country?? Even if they were completely stupid, spreading it in another country would at least have a non-zero chance of avoiding the virus spreading back to China. And all of this is irrelevant regardless, because if they just wanted to have human trials, I'm sure they have plenty of political dissidents they would work with.
I loved this whole thing, thank you for writing it and I'm glad you enjoyed it!
I have no dog in this hunt and don't feel like any important policy-making we have to do hinges on this outcome. We know we need to keep labs safe and we know we need to deal with zoonotic risks. And I still found it riveting to read, as well as all the comments here. I think this lands in the genre of competency porn and I'm here for it.
A meta-observation: I started today 50-50 on this question. Right after reading it, I was briefly at 90-10 zoonotic. An hour or two later, I'd say it slipped to 75-25, and now I'm landing maybe at 60-40. That's an interesting experience just to watch over the course of a day. I don't know, but I think if I read it all over again, it might goose me back up to 80-20, temporarily.
When I ask myself why looking at the debate as presented swings me a lot but then that regresses over hours, I would summarize that as "mistrust". Mistrust of the institutions that produce information, mistrust of the arguments presented, mistrust of the biasing potential of the debate format. That's my emotional/psychological bias and so the further I get from direct contact with the elements of the debate (which I found persuasive), then that default bias starts to take over more.
While my intuition listens along to the content of the debate, it also has some deeply-held opinions about the trustworthiness (or lack thereof) of experts. I have my hobbyhorse example of this around guidance to parents for what position to put their infants down to sleep -- which has changed three times since my adulthood. It's not that anyone was misleading anyone, but that things that look absolutely true to science in one era turn out to be absolutely wrong in the next era and that's how science proceeds. We don't know what we don't know, and that makes all my reactions to these kinds of competing stories pretty uncertain. And it also makes me marvel at other people's certainty.
I don't know, but I think this might have some relevance for questions people ask about how to persuade people about the importance of large scale risks based on a mish-mash of reasoning and data. Where and how mistrust shows up is a huge issue.
> Fourth, for the first time it made me see the coronavirus as one of God’s biggest and funniest jokes. Think about it. Either a zoonotic virus crossed over to humans fifteen miles from the biggest coronavirus laboratory in the Eastern Hemisphere. Or a lab leak virus first rose to public attention right near a raccoon-dog stall in a wet market. Either way is one of the century’s biggest coincidences, designed by some cosmic joker who wanted to keep the debate stayed acrimonious for years to come.
Scott messed up. The lab was built near wet markets because wet markets were a threat, and GOF research probably was conducted there because leaks could be blamed on wet markets.
The lab is fifteen miles of dense cityscape away from the wet market. There are thousands of buildings between the lab and the wet market. There's no way "sit around, wait for something to leak, surely it will surface at the wet market" would be a reasonable strategy.
The new notes USRTK obtained through FOIA in January 2024 aren't trivial. I had been pretty much on the fence about the lab leak hypothesis, but the new evidence is so overwhelmingly damning it literally made me nauseous reading it. The notes explicitly propose generating a consensus sequence with an inserted furin site. They propose using six BsmBI sites to assemble the infectious recombinant - confirming a 2023 hypotheses that the spacing of the six BsmBI sites in SARS-CoV-2 had a suspicious spacing. The notes proposed passaging the recombinant virus in human cell lines under BSL2 conditions.
It's a mistake to casually dismiss new smoking guns, all caps or otherwise.
"Why did the older strain start spreading later? Probably the virus crossed from bats into raccoon-dogs on some raccoon-dog farm out in the country. It spread in the raccoon-dogs for a while, racking up mutations, including the (less mutated) Lineage A strain and the (slightly more mutated) Lineage B strain. Then several raccoon-dogs were taken to Wuhan for sale, including one with Lineage A and another with Lineage B. The one with Lineage B passed its virus to humans earlier. Then 3-4 days later, the Lineage A one passed its virus to humans."
So apparently the zoonosis story requires *two separate* zoonotic transmissions, from the same location and within the same week, but from separate viral lineages? Am I understanding this right? And if so, then can someone please explain why both sides are just passing over this claim as a minor aside rather than a devastating hole in the theory?
I am not an expert in the field, at all, but this intuitively strikes me as *really* implausible.
The lab leak hypothesis seems like it either needs two leaks of its own, or a long enough period of spread to produce this mutation while somehow not leading to the explosion of hospitals until late January.
B is only two mutations away from A, and one of those mutations is silent. It could have happened in the same person.
Zhang et al 2024 has new early samples from Shanghai with a bunch of intermediates between A and B, consistent with a single spillover. I think their proposed phylogeny roots at an A0 strain which is one mutation closer than A to the bat viruses.
Out of curiosity, how widespread is the software bug Peter alluded to? If this is not fixed/ improved and the software in question is still widely used, and further if this bug might make such findings more likely, should we believe this latest finding?
I'm genuinely perfectly clueless and ignorant on the topic, so truly don't know.
I haven't seen anything about the bug he refers to, I would be interested to see a reference supporting his claim. I'm sure he's referring to something real but his interpretation seems overconfident.
Ok, I'm guessing he's talking about this line in Pekar et al 2022: "We identified numerous instances of C/C and T/T genomes sharing rare mutations with lineage A or lineage B viruses, often sequenced in the same laboratory, indicating that these intermediate genomes are likely artifacts of contamination or bioinformatics".
This is somewhat more cautious than Peter's "we can dismiss both" so maybe he was referring to something else?
The new paper (it's Lv et al 2024 contrary to my earlier ref, Zhang is the senior author) mentions this in a couple places but don't seem to think it affects their result.
That seems most plausible to me but it blows a huge hole in everybody’s timeline since it means cov2 was spreading around in humans before any of the cases flagged by the authorities. For a disease with a 10+ day incubation period and a relatively low rate of hospitalization though it could have easily been spreading for a month or so around young healthy people before anyone started looking for it.
Side note about the use of "conspiracy theory" in this context. It's got these two meanings, one being "nutty, unfounded belief" and another being "a theory about how people coordinated actions secretly behind the scenes." There's a Venn diagram where flat earthers believe an endless stream of government officials are conspiring to keep the truth from us.
It doesn't seem accurate to me to call the lab leak idea conspiracy theorizing in the "nutty, unfounded belief" sense. Smart people in very recent years, including in this debate, disagree. When you accuse the other side in that situation of "conspiracy theorizing", you're just engaging in ad hominem attacks which add no useful information.
Some lab leak people do believe that there was behind-the-scenes conspiring that happened, but so do some zoonotic supporters. There are more people on the lab leak side who do seem to be more of the nutty, unfounded type of believers. But it's a distraction to look at them, because the evidence in favor of lab leak is not like the evidence in favor of the earth being flat, etc.
Sometimes too the people we call conspiracy theorists are just people who have been slower to catch up to the preponderance of evidence, given that science is a process that unfolds over years. I don't think it's helpful to call people in this category conspiracy theorists.
Strong arguments were made on both sides of this. Peter's arguments look stronger. But as a person who doesn't have their hands on all the raw data, I wouldn't say that he "debunked" the other side. He just presented some other information that was more persuasive.
Anyway, I guess this is a wish for not calling people "conspiracy theorists" as it relates to things scientists themselves are still debating.
The whole Session 2 issue seems irrelevant to me. The simplest lab leak scenario is a virus being collected in the wild, brought to the lab, then escaping in guano on a tech's scrubs or something similar. No need for genetic engineering of the virus at all.
The argument seems to have been "natural virus in wet market" versus "engineered virus from lab" while neglecting other scenarios.
In a sensible world, the uncertainty about whether the virus came through the door or the window would be seen as a good reason to close both the door and the window; in the real political world we live in it's being used as an excuse to leave both the window and the door open.
In an even more ideal world the rest of the world would demand trillions in reparations from the Chinese government.
Aside from the obvious questions about whether we need to improve lab safety, and whether gain-of-function research caused a problem for us (yet)...
And aside from Saar's underlying motive, which appears to be an attempt to implement practical Bayesian reasoning...
Mostly because the discussion picked up political heat. Which warped the "official narrative", in ways that are still relevant. It's important to know which of our official scientific sources will follow along with tribal groupthink, and which will base their decisions on data. (Or rather, where on the spectrum each individual source falls.) And so this leads to a question of what the data actually are: what did we know, and when did we know it, and what kind of mistakes were made along the way, and why were those mistakes made, and what if anything can we do about it in the future.
Yes, the irony on my part is that I actually do want to "trust the science", but I also am slightly familiar with how the sausage gets made. And when we have scientists who are also public figures, making claims with far too much certainty, that makes me skeptical of everything they say: it's being a public figure and engaging in public relations*, not being a scientist and engaging in science. Give me something with error bars, and explicit limits on our certainty, and I'm much more likely to trust it. It's not a perfect heuristic, but I came to it the hard way. And it does give me a bit of moral high ground, in that if people would just stop trying to score political points and be honest, I'd be on their side.
"And when we have scientists who are also public figures, making claims with far too much certainty"
...It's a selection effect, at least in my country.
Journalists call around till they find some scientist that says what the journalists want to say themselves but do not have the academic cred to say. And/or till they find some scientist who is willing to say something dramatic or controversial about the topic. Those scientists receive new calls. The ones who offer error bars do not.
I do not blame journalists. Most live and die by clicks these days. Those who hate the system leave journalism, creating a new selection effect where the remaining are those who like to live by clicks.
In the old days, newspapers could print a few long-winded serious articles in each issue, since no-one knew for sure how many - or few - that read them. (The Germans for some strange cultural reason still have some newspapers like that. It also helps to have a potential pool of 82 million readers.) But after online publishing, editors and owners have accurate data on how often such pieces are read. Which has been a death knell to serious science journalism.
...admittedly, I am talking here of the social sciences (broadly defined). The natural sciences is somewhat better placed.
Instead, we have substack and a few other places where the global audience of people who are actually interested in nuanced stuff (with error bars) is large enough that it can fund at least some writers who have nurished a "the truth, whatever it is, shall set you free" attitude to their subject matter.
Yeah, it's weird to me that they didn't mention the third hypothesis, which I thought was considered as plausible as either of the other two initially. But I think that Session 2 deals with a lot of evidence that had been taken by some to be extremely strong, and it only supported the gain-of-function version, not the lab-collection-accident version.
"As mentioned earlier, the DEFUSE grant was rejected."
I do not know how many people here are familiar with how the federal grant system (whether NIH or DoD) for biomedicine works, but if one claims that the only way experiments get done is if 1) they are proposed in a grant and 2) that grant is funded, they are either unfamiliar with how the system works (like Peter) or else engaging in intentional misdirection (like the many virologists who have made this claim and certainly know better). As a general rule, the proposed experiments in a biomedical grant fall (often roughly equally) into 4 categories: 1) work you have already done but not published and inot included as preliminary data (you have to include just the right amount of data to justify the proposed experiments/money) ; 2) work you never intend to actually do but which looks good on the application; 3) work you do during the 1-2 years the grant is under review but no decision has been rendered; 4) work you actually will end up doing during the funding period.
That the DEFUSE grant itself was not funded is drastically less important (frankly not important at all) from the perspective of a Bayesian trying to figure out the origin than the fact they proposed making the precise modification that is so striking about Covid and has not been seen in any close relatives. David Baltimore gets this, Nick Patterson gets this, and I think anyone with knowledge of both how these grants work as well as the molecular biology gets this.
I agree, several of my friends are in grant-seeking areas and universally this has been the reported experience. Most times your current work is being funded by your last grant, your future work is being funded by your current grant, and you current grant was based on your past work.
Related, a saying by an old colleague: "Collecting the data takes 90 per cent of the available time. Reporting the findings takes the other 90 per cent".
>I do not know how many people here are familiar with how the federal grant system (whether NIH or DoD) for biomedicine works
I am _not_ familiar with it, and I appreciate the comment!
>1) work you have already done but not published and inot included as preliminary data (you have to include just the right amount of data to justify the proposed experiments/money) ;
Now that you mention it, that sounds _very_ familiar from other areas of research as well!
Saar raised the issue of ascertainment bias, but undersold it. There's no question imo: the case search *was* heavily focused on HSM, and the HSM sample search *was* heavily focused on the southwest corner. The Chinese authorities who did the search have been clear on this (Bahry, 2023, Table 1 and Fig. 1).
E.g. we now know that the raccoon dog stall was sampled 10 times, with 5/10 positive; while the pool of blood outside a beef stall in the east wing was only sampled once, with 1/1 positive (Liu et al., 2023; cf. Bahry, 2023, Fig. 1). That, not the racoon dog stall, was the highest positivity ratio. Peter's heatmap was based on a paper by some Western scientists who had no idea how many times each stall had been sampled, and had instead speculated that every sampled stall had been sampled equally (Worobey et al., 2022).* After Jesse Bloom also showed there's no positive association between raccoon dog and SARS-CoV-2 genetic material, nothing implicates them, not even in a God's-joke way (Bloom, 2023).
China CDC director George Gao has also emphasized, when asked by the host of a BBC podcast, that this overwhelming ascertainment bias means you can't assume genuine case clustering around HSM (https://www.bbc.co.uk/sounds/play/m001ng7c, 24:00–25:12).
Bloom, JD. (2023). Association between SARS-CoV-2 and metagenomic content of samples from the Huanan Seafood Market. Virus Evolution 9: vead050. https://doi.org/10.1093/ve/vead050
Worobey, M. et al. (2022). The Huanan Seafood Wholesale Market in Wuhan was the early epicenter of the COVID-19 pandemic. Science 377: 951–959. https://doi.org/10.1126/science.abp8715
*This zoonosis-theory paper is one of many by a prominent coauthor network centered on "Proximal origin" authors Andersen, Holmes, Rambaut, and Garry.
*edit: Peter's figure is the version from that co-author network's preprint (Crits-Christoph et al., 2023, https://zenodo.org/doi/10.5281/zenodo.7754298), not the paper (Worobey et al., 2022). (It also includes the partial metagenomic data that they scooped from Liu, Gao et al., getting in some trouble with GISAID for.) It's still obfuscatory, acting as though the SC2-negative samples weren't *also* concentrated in the sw corner.
I do not believe that the pool of cow's blood in the east side of the market is the source of the covid pandemic based on 100% of its samples (1 out of 1) coming back positive.
Otherwise, I think I adequately answered all the questions about ascertainment bias at the debate.
Nobody's saying it came from the cow's blood (just like Bloom wasn't saying it came from fish). The point is that the Worobey et al. / Crits-Christoph et al. heatmap is an obfuscation, based on their making a baseless assumption that was already known to be baseless and is now known to be false. At present nothing genuinely points to raccoon dogs.
The only adequate answer to the clear, known, overwhelming ascertainment bias in the early case search is to say "Wow, you're right—the ascertainment bias *is* clear, known, and overwhelming, so we can't appeal to apparent clustering as evidence."
(For the HSM samples, now that we know how many samples were taken from each sampled stall, it probably *is* possible to properly redo the Worobey et al. / Crist-Christoph et al. heat map, modifying their code to use the positivity ratio. I haven't done it myself because I'm not good at coding. It could either be done with the measured positivity ratios, e.g. 5/10=50% and 1/1=100%; or using pseudocount smoothing like in my figure, e.g. [5+0.077]/[10+1]≈46% and [1+0.077]/[1+1]≈54%.
Note: Worobey et al. / Crits-Christoph et al. haven't properly redone the heat map either; instead, they've since chosen a new way to automatically downplay the less-sampled stalls away from the southwest corner, by mapping p-values instead of positivity [https://www.biorxiv.org/content/10.1101/2023.09.13.557637v1#comment-6279754825].)
Can you just say why George Gao said they focused too much on the market. If we're looking for strong evidence - the guy in charge of the case search said there was ascertainment bias both within environmental sampling in the market and in the city of Wuhan. Is he just very confused? What is going on here?
Question for Peter: don't you think you're being a little credulous with respect to the early case map, given an obvious and known effort from WHO/China to support anything but LL?
Separately--great job. To go into this with nothing but your own research and bet a good portion of your net worth against some rich guy, and win, is something from a movie. You should be proud for sure.
Peter claims the Worobey study's map is from "before ascertainment bias", but I just don't understand how. If the first known case was from the market, and with the knowledge of SARS zoonotic origin, you don't even need to attribute any bad faith to get ascertainment bias around the market. Once the LL hypothesis existed (and I don't really know when it was first proposed, or when the CCP thought it might be a possibility, you also have bad faith to contend with, even if their official story was propaganda about an American lab.
There are lots of studies vaguely like this. I think we should be skeptical, because COVID has doubled every 3.5 days in every population it's been exposed to.
Suppose there was a person with COVID in Brazil on November 27th. The first confirmed case in the Americas was January 21st. By January 21st, the one COVID case in Brazil should have grown into ... 55 days divided by 3.5 = 16 doubling times, 2^16 = 65,000 cases. There's no way to miss 65,000 cases of COVID, especially when the next week that's quadrupled to 260,000.
So you would have to argue that COVID fizzled out sometime after November 27 - eg one person had it, and didn't pass it to anyone else. That's possible for one case - maybe that one virus just had bad luck, or the person quarantined. But it quickly because implausible once there's more than 2 or 3 cases.
So the question you have to ask is whether there was some way that there were only three cases of COVID ever in Brazil (until it officially arrived from China), and yet some scientist still got very lucky and found one of those three cases in the wastewater.
I would rather believe they made some kind of error, which also fits with the fact that nobody has any explanation for why a Chinese bat virus should be in Brazil before it's in China.
That's a good guess, and sound reasoning. If covid was in Brazil in November 2019, why isn't it also in the wastewater in other cities around the world? Why just there? Where are all the cases? All the deaths?
I had 20 more slides prepared, in case Saar brought up any of these outside of Wuhan issues. Since he never contested them, I ended up turning them into a blog post:
The specific issue with the Brazilian study is probably lab contamination -- they tried sequencing some of the covid in the wastewater and it shared mutations with B.1 strains that weren't otherwise seen until later in 2020. The authors theorized that meant that B.1 was circulating worldwide before Wuhan (one lab in Italy also found something similar, probably for the same reason). But the B.1 mutations made the virus better adapted to humans, so it would need to then travel to Wuhan and revert back to something less well adapted and then mutate into B.1 again.
A far more likely theory is that the 1 sample from Brazil and 1 sample from Italy were processed in labs where someone was a little bit sick.
I think one of the Italy studies tested samples from a range of dates, with positivity showing up maybe in mid-December? It's harder to explain that as contamination.
Not trying to Texas Sharpshoot, but I can think of some plausible explanations:
1. The stored samples are physically sorted chronologically, because this is a research lab and research labs like to be organized and alphabetical order wouldn't make sense in this case. Material from an "actual, for-real positive" sample contaminates its neighbors.
2. The work of studying the samples is passed to multiple people. The work is divided into chronological batches. By a stroke of luck, a positive-but-asymptomatic employee happens to handle and contaminate the batch of samples immediately preceding the "actual, for-real positive" batch.
3. The samples are stored "bottom to top", chronological order, and processed "top to bottom". Residue from an actual, for-real positive sample is left behind in lab equipment.
4. (comedy option) All the samples are processed by one person in chronological order. None of the samples originally contained covid, but the one person processing them caught covid at some point and began contaminating the batches as they worked on them. All the positives are false.
Basically, jokes aside, it becomes a lot more likely (or at least less-unlikely) when you consider the on-the-ground human aspects of lab work and sample storage.
I mentioned this elsewhere, but how is an early December market outbreak consistent with a 3.5 day doubling time? Fig 1 of this paper reports ~200 excess deaths per 100K per week from pneumonia in Wuhan on 1/23/20, the day of the lockdowns.
A 3.5 day doubling time for an outbreak starting on 12/1 combined with a .05% CFR predicts (2 ^ 54 days / 3.5) * .05 CFR * 2 three and a half day periods in a week = ~4400 deaths the week of the lockdowns. So unless Wuhan has a population of 2M people, this is completely inconsistent with the excess death reports.
which is directionally consistent with the simpler doubling time argument.
Even so, we're being asked to believe that under a late November market emergence date, the first confirmed cases were third or maybe even second(!) generation cases for a respiratory disease in a crowded city that no one was looking for and no one knew anything about? I am skeptical .
Maybe this has been answered elsewhere, but isn't it awkward to be relying on Chinese death reports? I had understood that was one of the areas that was most suspect about their records - they consistently downplayed the amount of covid in their population to give the impression that the governmental response was good and/or that there was no problem that even needed a response. Basically deny that there's a problem until it can't be covered up, then cover it up as much as possible after that. I thought that was common even outside this particular debate, that nobody really believes that the Chinese govt handled covid so well that their population death rates were a tiny fraction of everywhere else.
I think (and may be wrong here) that there would be some lag time before the Chinese authorities noticed there was a problem to cover up. The first week(s) of increased death counts would look like statistical noise, or just the usual annual fluctuations between the relatively-less-morbid summer season and relatively-more-morbid winter season. So while the "deaths from Covid" statistics can be (and probably are) fudged downwards, we would still expect to see a "hiccup" in the "overall deaths" reports released prior to January that, with hindsight, isn't simply seasonal fluctuations.
The idea that all these data involved in this debate are suspect is a fair point. My personal view is that excess deaths is possibly the most robust way of looking at the early pandemic in Wuhan, but I might be mistaken about that.
I'm surprised that you have that kind of trust in a continuous epidemiological model with only a couple parameters, surely things are noisy at small population sizes? As an example from the low end, Pekar et al 2022's modeling suggested 8 spillovers of which 6 died out. There is clearly some point at which COVID does not double every 3.5 days.
There is also documented bias from the case reporting criteria changing over time. Wouldn't cases of a previously unknown disease be undercounted at first?
The experts seem to be using Monte Carlo for this rather than fitting an exponential.
I agree at small population sizes things are more iffy, which is why I'm saying the data don't rule out a very small COVID epidemic, but they do rule out a sizeable one. I guessed "not more than 2 or 3 cases", but you could convince me the real number was 10 cases or something. I wouldn't believe 100.
I would bet those 6/8 failed spillovers didn't reach the 3 case threshold, certainly not the 10 or 100.
Sure. Complex models get you published in a fancy journal. They don't necessarily give you correct answers. I mean, 8 spillovers? In Wuhan, but no where else? Does any one seriously believe that?
All of the many errors uncovered in Pekar et al. were in the direction favoring zoonosis, which is noteworthy.
If you don't trust a doubling time model, see if you can plug numbers into the above compartment model predicting a such an enormous excess death figure by 1/23. I couldn't do it in a back half of november time frame without unrealistic assumptions about the CFR or R0.
A priori there is no continuous model that I would trust enough to extrapolate backwards in time, through all the discrete behavior around the spillover, and come up with a credible start time down to a week.
I'm not any kind of expert in the field, but isn't that a very suspect use of modelling?
Well, all of this just confirmed my priors: most debate is utterly worthless. They can talk about likelihoods and logical fallacies all they want, but the problem is that no one has any hard evidence. Peter's obviously the smartest one in the room here, but just being intelligent is of no use if you don't have anything to work with. Sure, you can take a bunch of circumstantial evidence and then claim that they make your theory more likely by pulling numbers out of your ass, but none of that gets you any closer to verifying the actual truth. And at this point, actually verifying the cause of COVID is likely impossible-- any hard evidence would have been burned by the CCP, since even they don't have anything to gain by knowing the truth.
...So what's even the point of all of this? As others have pointed out, the actual dangers of wet markets or GoF research is completely independent of whether or not COVID in particular was caused by either of them. One event doesn't change the reality of the systems at play. We already agree on the information that matters: that the situation was potentially preventable by the CCP, and that they failed to do so. But even that wasn't enough for countries to stop doing buisness with them, so how would figuring out the origin of COVID change anything? This entire discussion is utterly pointless.
The (applied) point of the discussion is the hope that the CCP will react to this (for them) embarassing and apparently never-ending discussion by cleaning up both the wet markets and the research laboratories studying stuff like this.
...and from that angle, the ideal outcome of the discussion is 50-50 for the wet market versus the laboratory leak - hypothesis.
More generally, there are two motivations for caring about the probabilities for a lab leak versus zoonosis:
(1) You portray yourself as an observer of the world, and simply are intellectually curious about what goes on in it. You care about the probabilities only because of the off-chance that during your lifetime CCP will fall from power or something equally dramatic will happen in China, which makes the authorities open the archives so it is possible to check if what you thought was probably the cause, was actually the cause.
(2) You portray yourself as a participant in the world, and want to influence policy-makers and your fellow man to limit threatening risks. The probabilities you and/or others assign to possible causes serve as calls to action to do something to change/reduce the probability of similar pandemics in the future.
These motivations are very different. They are likely to influence your own stated probabilities, as well as how you interpret why others offer the probabilities they do.
This is a highly interesting topic. Unfortunately I don't have the energy tonight to get deep into all the great points brought up in the post. But I want to share a few scattered thoughts in babble mode:
(1) Taking a 5% chance of dying on a climb is pure insanity. Maaaybe taking a 5% *lifetime* risk is worth it if you really love climbing. Hardly an endorsement for risking $100k at 4%.
(2) It's a terrible idea to offer a standing bet on your true beliefs. You're literally doing adverse selection and end up with a sneaky rules lawyer as your debate "partner".
(3) It seems that Peter is just a good debater, of the "debate club" variety. Laser focused on winning the debate, rather than finding truth. I was repulsed by the dirty arguments made, for example the quibbling about whether to call the first outbreak a "superspreader event" or not. Regardless of the completely irrelevant doubling times Peter brings up multiple times, a place like the Wuhan market clearly is a mechanism for turning a handful of cases into 80k cases. I've only skimmed the *summary* of the debate, so please point me to where Peter tackles the real underlying issue head on.
(4) Re: orders of magnitude. There is no limit to how badly you can fuck up a bayesian estimate. For example, Peter's 1 in 10'000 probability of the first infected person working at the wet market. That's just a uniform prior! It's so, so bad. The easiest way to debunk this is to take these odds, and start listing conclusions that would need to follow if they were actually correct: "viruses are equally likely to spread to everyone regardless of their profession and their position in the social graph"... um no, everyone with an inch of common sense (or kids) knows that viruses spread where people gather: schools, workplaces, and yes, markets. The vendor at that market didn't need to be the first, merely among the first *few*, and then keep coming to work while contagious.
(4a) It strikes me this might be a valuable method to flesh out. Let's call it the *reverse inference test". I don't know why, I just feel like it latches onto my intuitions more cleanly than trying to refute a claim in the same direction it's made.
(5) Probability cascades are bullshit. To me, this is the classic example of "science as attire"; you pretend to be scientific by multiplying numbers together, but what you're really doing is spinning a narrative. It's extremely hard to come up with an exhaustive enumeration of posibilities. MECE, as it's sometimes called (mutually exclusive, collectively encompassing). Or for the mathy people, an equivalence relation on the possible states of reality. If you doubt this, go take an advanced algorithms class and report back. I'm not joking... repeated encounters with impartial yet implacable entities (compilers, test suites, theorem provers, etc) will drill this into you like nothing else. Peter's argument style is that of a college freshman, and any competent math mentor would beat it out of him before it rots his brain. To the extent Peter is aware of this, it just reinforces (3).
(5a) to clarify the above point, the typical mistakes made is multiplying together probabilities that aren't independent, and to implicitly assume that the only way to reach state X is the specific path towards X you've laid out. These can make a 50% event seem like a 1% event, or the other way around.
(6) Explicit bayesian calculations in the absence of a *formal model* are a horrible idea. It just takes confusion and multiplies it by orders of magnitude. What I've learned in my time as a scientist is that the relationship between the real world and the model you're working on, and how to transport assumptions and conclusions back and forth between them, is one of the hardest things to get right. And we don't really have a systematic way of teaching this skill.
(7) If you wanted to blackpill me on the applicability of bayesian calculations to everyday life, you've succeeded. In particular, I see no reason to update my beliefs regarding the lab leak hypothesis in response to this debate. I'm still 50%+ in favor.
Wow, that was a lot more than I thought. Thanks for reading, and I hope it was helpful, even in this low-effort format.
Actually, I have a question, what specifically in advanced algorithms are you thinking about when you say that it'll help with models? I'd like to implement a couple of those things as an exercise you see.
(Personal guess it's some kind of probabilistic algorithm, or just seeing a bunch of greedy algorithms that looks optimal... but actually doesn't even converge. I would guess that I'm not thinking expansive enough but seeing the gap in my prediction and your answer would be illustrative of what I'm missing)
For one, combinatorial algorithms. Stuff like sat solving, computational geometry and so on. The state space boggles the mind. Graph algorithms, including stuff like reductions of various problems to max flow.
One of the most interesting classes I took in my master's was models of computation. We had to program algorithms on Turing machines, cellular automata, tile completion systems, and so on. The interesting part about it is that we could have our solutions automatically checked. Not sure if there's equivalent stuff out there on the internet.
There's so much more. Stuff like computer graphics or physics simulations, where the operations are very elementary, but you can't just debug by looking at a single code path. You kind of have to think on a systematic basis and judge the output as a whole.
My brain skipped a beat when I read the summary that Peter dismissed one of the later market spreads as "that didn't come from a human source, there were COVID particles on the fish". That feels eerily lab leak mechanics adjacent and I found myself slightly confused why he thought that was likely but COVID from the lab coming to the wet market and then spreading from there unlikely. Probably if I watched the videos, I would be less confused about this, and I'm not trying to make an argument, just registering my confusion.
Generally, I have no strong position on this question at all and a low prior on that the things I think about it are true. I thought this article was a fantastic deep-dive! I realise I could get even more information if I watched the videos, but I'm not really a video person. Thank you so much for this. :)
Something about that whole line of reasoning felt a bit off to me. If the explanation is due to non-COVID-zero imported goods, shouldn't we expect to see a similar occurrence of outbreaks with other goods importers. Why did it specifically happen in a wet market?
Frozen fish is the leading "anywhere but here" theory, whose popularity is almost entirely restricted to the government of China. Peter doesn't actually believe in it.
Regarding the 1 in 10000 chance of being discovered in a wet market if lab leak:
That is flawed. You might as well assign 1 in 365 that it was discovered December 30th. You can tell it is flawed because if it had been discovered in a school, Saar would assign 1/10000. Or in a restaurant, 1/10000. Or a bakery, 1/10000. He gets a free 1/10000 wherever because any public place can be differentiated by at least 1/10000. “Only 1/10000 Wuhans work at Wuhan IHOPs!”
That was discussed in the 3rd debate. You can't assign high numbers unless there's some importance. If someone gets murdered in one building in Wuhan, and there are 10 million buildings, that's not a 1 in 10 million coincidence.
On the other hand, there are only 4 markets in Wuhan selling wild animals (out of a giant city of 10 million plus people). Huanan market is the one selling the most animals (about 40% of all the wild animals coming into Wuhan), but it's otherwise not the biggest market or anywhere near the most crowded place in Wuhan.
So if you're trying to weigh the odds between "wild animals caused it" and "the lab caused it", then it is of significance that it first showed up a the place where the wild animals were sold, as opposed to 10,000 other places it could have first showed up in Wuhan.
That's not the complete reference class, however. Had the virus first showed up in a zoo which treated wild animals, or which took wild animals for its exhibits, and the chain of custody involved intersected with the region in question, that would also be treated as equally valid evidence for zoonosis. If it first showed up in a restaurant which used products from the wet market, or somebody who made purchases there, that might also count as part of the reference class. Anybody involved in the logistics of moving animals and anybody who visited a farm (there's no reason to limit ourselves to raccoon dogs). The workplace or school of anybody who visited a farm. The workplace or school of anybody who traveled to the province in question. The raccoon dog farm already has two degrees of separation, so any plausible path with two degrees of separation is part of the reference class - and that's an enormous reference class.
Note that the reference class for zoonosis isn't limited to Wuhan, and, indeed, we wouldn't expect it to first show up there given a zoonotic origin.
Once you start critically examining the actual reference class to which the wet market belongs, without arbitrarily limiting it to particular observed qualities, it is pretty clear it is actually enormous.
Whereas the laboratory is almost a reference class in and of itself; there are a few partial matches, so it's not entirely its own reference class, but it is something like unique in a way that the wet market absolutely is not.
You nod at this briefly in the conclusion, but what I find fascinating about this debate is that the what truly happened doesn't seem to matter at all! If an oracle arrived and told us the source, I don't think it would (or should) change our planning at all. The investigation of COVID has revealed that there are huge risks from *both* lab leaks and zoonosis that we need to invest significant effort in reducing, and which one actually happened in this case doesn't change that.
It seems to me that it matters. When you are estimating whether the next pandemic will come from nature or a lab, the result here should affect your prior.
What matters more immediately is whether to even continue investigating the origin of COVID. The most visible advocates for a zoonotic spillover have been hopefully declaring the case closed for some time now. But if COVID really was related to the DEFUSE research, there is a huge amount of evidence within the reach of US investigators and it would be a shame not to look at it.
Maybe I'm missing something, but it seems like Peter's argument hinges on two separate cases of zoonosis within days of each other in the exact same location and the same mutation? Or maybe I misunderstand that point.
Hopefully Peter will show up and explain, but my impression of his case is something like:
- Raccoon-dog on a farm catches COVID from bats
- Lots of raccoon-dogs on that farm get infected
- Several raccoon-dogs from that farm are brought to Wuhan for sale
- This version of COVID is already good at infecting humans
- The virus spills from two of the raccoon dogs to humans, for the same reason that if a bunch of infected Chinese visited America, it might spill from two of the Chinese over to Americans - because the virus spreads quickly in a naive population.
So the story is that both strains arose on a farm, and then it crossed to zero humans at the farm, but swiftly infected a whole bunch at the wet market? Seems pretty tenuous.
It's entirely possible there were spillovers elsewhere that died out. Before the pandemic the CCP had encouraged rural people to get involved in farming animals to make money. It was part of Xi Jinping's drive to "end absolute poverty" (in reality the threshold of what "absolute poverty" meant was a shifting target, but China did give many people a way to make a living this way). So you had people in very remote regions raising animals, in some cases near bat caves, then shipping them into big cities to be sold.
Pandemics generally don't start in rural areas, because the average person has many fewer daily contacts, lowering R to the point where they're unsustainable. I think there was a study showing that a bunch of animal handlers had SARS antibodies (suggesting they'd gotten the disease) when all known SARS outbreaks were in cities.
1. Everyone *on both sides* including the Worobey group and the WIV, agrees that the ancestral virus is in Yunnan or SE Asia. Miller keeps denying this but it is so totally at odds with all of the evidence, that you have to ask questions of him here.
2. The claim therefore has to be that the raccoon dog was infected in Yunnan and taken to a farm in Hubei where it acquired the furin cleavage site in Oct or Nov 2019. There is no direct evidence of this, and raccoon dogs roam wild in Hubei, so it seems weird for them to ship them from Yunnan. There is no evidence that raccoon dogs can sustain such intense transmission to cause acquisition of the furin site as happens with avian influenza, and lab experiments suggest that they cannot do this. Freuling et al found an R0 of 1 in their experimental infection of raccoon dogs.
3. Raccoon dogs have never been found infected outside of a lab, unlike eg mink.
4. If lots of raccoon dogs on a farm get infected, why were there no outbreaks anywhere except Wuhan city where the farm was supplying? If you are envisioning thousands of infected raccoon dogs on a farm, this is a big BF in favor of the lab. It only really works if you are envisioning about 100-200 raccoon dogs on a farm, but then you are conditioning on a very specific farm size, and we have evidence that the typical farm is an order of magnitude larger.
5. The raccoon dogs were wild caught in Hubei. Wang et al 2022 tested 15 wild caught raccoon dogs at the suppliers of the market in Jan 2020 and they were found negative. Around 38 raccoon dogs were sold at four markets per month. So, this is half of the inventory. The raccoon dogs photographed at the market were agreed by the Worobey group to be wild caught in Hubei. Maybe you think half of the raccoon dogs were topped up from farms in Hubei. There is no evidence for this. If they were topped up from fur farms, they would have been in eastern China, even further from the ancestral bat viruses.
6. So, the claim is now that these ~10 raccoon dogs are responsible for the outbreak. This is an extremely small fraction of the total wildlife trade in China, and conditions the zoonosis hypothesis on an extremely specific proposition. To justify this, one would need extremely strong countervailing evidence of raccoon dogs being the intermediate host. But what we do have is extremely small traces of covid in their cages and a negative correlation of raccoon dog genetic material with covid. Whereas for the viruses that the animals in the market were actually infected by, like bamboo rats covs or canine covs, there is a strong correlation. Miller's claim that 'we have down to the stall' is just wrong.
The collective BFs in favor of this causal chain are extremely weak, and in fact rule out raccoon dogs.
2 of those infections go on to infect significant numbers of other people.
The odds of there being 2 spillovers as opposed to 1 are about the same, once you have the right components in place. Eric asked a very similar question in the first week debate, if you want to take some time to watch.
>The six estimates span twenty-three orders of magnitude. Even if we remove Peter (who’s kind of trolling), the remaining estimates span a range of ~7 OOMs. And even if we remove Saar (limiting the analysis to neutral non-participants), we’re still left with a factor-of-50 difference.
In a way, the scatter in estimated here is even more disturbing, because AI risk unavoidably involves estimates of future events, many with only weak analogies to known past events. _This_ estimate is all about events which _do_ have strong analogs to known past events.
I don't know what to suggest. One large class of worries is indeed treating some probabilities as uncorrelated when one could indeed expect correlations. E.g. as you said, researchers modifying viruses in ways intended to be less researcher-like and more nature-like in order to anticipate natural changes.
One other thing that generally bothers me is the blanks in a lot of the probability calculations where one side estimated an effect and the other side just didn't discuss it. Maybe it would have been helpful to have a final pass where all sides gave estimates for _all_ of the factors?
I truly hate to be a meta-commentator refusing to engage with the meat and bones of the debate, but I'm still puzzled by why this matters enough for the debate to be worth at least 100000 dollars and a few dozen hours of a few intelligent people's time (probably a lot more time and money, actually, given the extent to which all of this was discussed in other sources, of varying levels of sophistication, that I don't follow).
Would it help anyone in particular, even people who lost loved ones or whose businesses did not survive the economic consequences of the pandemic, to know whether the virus was zoonotical or originated from a WIV leak?
I doubt there would be meaningful political or economic consequences either way. I'm moderately sure virological research is still being funded about the same way it did before the pandemic, I'm a bit more sure wet markets are still as unsanitary as they were before the pandemic, I'm quite sure trade/political relations/economic dependence of the world on Chinese economy and industry is about the same as it was before the pandemic, and I'm very sure human beings still die the same way they did before (and during) the pandemic.
How much time time and energy has been spent debating whether the Holocaust happened as claimed, or what were the root causes of the American Civil War, or what caused the Great Depression, or how the pyramids at Giza were constructed? People like to discuss and understand important events in world history, even if there are no direct political or economic consequences hinging on the outcomes of these discussions.
Regardless of whether or not this pandemic originated in a lab, it's still the case that GoF research is potentially quite dangerous, its claimed benefits quite dubious, and we certainly should be having a serious debate about whether or not it should be more strictly regulated than it is. If COVID hadn't happened, we could make that a dispassionate theoretical debate. But here and now, we're stuck with a debate involving people who remember and care about COVID. And the way real people actually work, if one of the things they "remember" about COVID is that a bunch of conspiracy theorists claimed that it came from GoF research, but they were later shown to be lying liars who all voted for Donald Trump, then we are not going to get a fair discussion about how GoF research ought to be regulated going forward.
The same may be true of wet markets, but the case is I think weaker there because the animal reservoirs will still exist and will still interact with local human populations, and it's not clear how much benefit there would be to shutting down the wet markets if it just delays the remote village -> cosmopolitan city step a bit.
But so long as both hypotheses are credible, we need to keep making it clear that both hypotheses are credible and arguing against anyone who claims to have "proven" that it was 99.99...% certain to have been one or the other. And if both are credible but one is genuinely more probable than the other, then we want to know that so we better know how to allocate our political capital in a world where we may not be able to shut down both the wet markets and the GoF labs,
...since the main applied point of the debate is to shame CCP into making sure neither wet markets nor laboratory slip-ups cause the next Big Pandemic, a conclusion that the wet market and the lab leak hypothesis are equally likely explanations of the Covid pandemic would be the ideal (applied) outcome of the debate.
...that is, if your main motivation for caring about these probabilities in the first place, is as a call to action to limit major risks of similar pandemics in the future.
Great write up, and incredibly intellectually stimulating. One thing that struck me is how much harder it is to understand the past than to predict the future! If Rootclaim is so good, it should prove itself on the prediction markets, instead of relying on judges. This also solves the problem of financial imbalance, challenging people to high stakes debates is like bullying someone at the poker table when you have a pile of chips.
Importantly, Miller incorrectly claimed the N501Y mutation would result from passage in hACE2 mice (mixed them up with BALB/c mice). The major papers Miller relied on have been seriously challenged since the debate. See Stoyan and Chiu (2024), Weissman (2024), Bloom (2023) and Lv et al (2024). Overall the circumstantial evidence makes lab v plausible:
1. Chinese researchers Botao & Lei Xiao observed lab origin was likely given the nearest known relatives to SARS-CoV-2 were far from Wuhan. Wuhan Institute of Virology (WIV) sampled SARS-related bat coronaviruses where the nearest relatives are found in Yunnan, Laos and Vietnam ~1500km away. They refuse to share their records.
3.Molecular data: Only sarbecovirus with a furin cleavage site. Well adapted to human ACE2 cells. Low genetic diversity indicating a lack of prior circulation (Berche 2023). The CGG-CGG arginine codon usage is particularly unusual but used in synthetic biology.
4. DEFUSE full proposal: virus 20% different from SARS1, consensus seq assembled with 6 segments, without disrupting coding seq, BsmBI order, FCS. SARS2: 20% different than SARS1, 6 evenly spaced fragments w BsmBI and BsaI restriction sites, FCS.
Jesse Bloom, Jack Nunberg, Robert Townley, Alexandre Hassanin have observed this workflow could have lead to SARS-CoV-2. Work often begins before funding sought or goes ahead anyway.
5. Market cases were all lineage B. Lv et al (2024) indicates there was a single point of emergence and A came before B. So market cases not the primary cases. See also Bloom (2021), Kumar et al (2022). Peter Ben Embarek said there were likely already thousands of cases in Wuhan in December 2019.https://t.co/50kFV9zSb6
8. Lineage A and B only two mutations apart. François Ballox, Bloom and Virginie Courtier-Orgogozo note this is unlikely to reflect two separate animal spillovers as opposed to incomplete case ascertainment of human to human transmission (Bloom 2021).
9. Sampling bias. George Gao, Chinese CDC head at the time, acknowledged to the BBC stating they may have focused too much on and around the market and missed cases on the other side of the city. David Bahry outlines the documented bias. Michael Weissman has shown this mathematically.
Miller incorrectly claimed the N501Y mutation would result from passage in hACE2 mice (he mixed them up with BALB/c mice). Since the debate the key papers that Miller relied on have been badly undermined:
2. Lv et al (2024) found new intermediate genomes so the multiple spillover theory is unlikely (it was anyway given lineage A and B are only two mutations apart). Single point of emergence is more likely with lineage A coming first. So market cases are not the primary cases (all lineage B). Their findings are consistent with Caraballo-Ortiz (2022), Bloom (2021).
3. Jesse Bloom showed again the market samples don't support market origin. t.co/rorquFs1wm
4. Michael Weissman shows mathematically ascertainment bias in the early case data the judges relied on (George Gao, Chinese CDC head at the time, acknowledged this to the BBC last year - they focused too much on and around the market and may have missed cases on the other side of the city).
5. Account that identified errors in Pekar et. al. leading to an erratum last year has found another significant error. Single spillover looks more likely. t.co/GAPihZu51P
> I have a weird urge to visit Wuhan as a tourist, see the Wuhan Institute of Virology, stroll through the Huanan Central Seafood Market (unfortunately closed), maybe eat a raccoon-dog.
Scott N Alexander N plans N to N eat N Raphtalia N CONFIRMED
Thank you for an incredible write-up and analysis.
I am actually shocked at how poorly Saar fares here. The vast majority of his evidence is basically "if things were different in a way we can't know, then our conclusions would also be different, therefore lab leak." E.g. "if COVID spread rate was lower in the beginning", "if an infected WIV worker spread to the wet market then isolated", and so on. That's it, really? I was expecting so much more actual evidence. And then what data points they do provide (e.g. "Mr Chen", 90 cases, intermediates, etc) have apparently valid reasons why they're not evidence. I started out 60/40 zoonotic, but based on this I feel I have no choice but to update to 99/1.
I feel the real problem with this setup/framework is that a specific position is taken, with overly-precise numbers, and the goal of the debate is to simply justify that final percentage. You see this with Saar's "but the judges just did the math wrong" sour grapes complaint.
In the IT and infosec world, we have the FAIR method--a similar method for building quantified models of risks. The real value of these isn't that you convinced yourself you have only a 1-in-10000 chance of being compromised by a state actor in the next 10 years... it's the shared understanding of the risks your business thinks are important, how you think those risks are correlated, how you think they're evolving over time, etc. That enables your organization to act from shared goals, and not get bogged down in interminable debates day-to-day. The final numbers are fine, especially if you use confidence intervals, but you know it's only a snapshot in time using your best data and estimates as of the last time you underwent the exercise; by the time you update in a year/quarter, you will have more data and different estimates. Sometimes, risks that weren't even on your model last cycle, are now the top priority because of what's happening in the world (or because you gained new information that you missed previously). It's a fact of life.
With that in mind, I think Saar fares better if his arguments are framed as reducing absolute confidence in zootonic origin, and not the reason why LL is the only possible explanation (which it obviously isn't, even from Saar's own arguments). So yeah, I was only joking about updating to 99% in favor of Z--I think there's enough doubt thrown in from Saar's evidence alone that it could not possibly be more than 80%. But Peter/Z still wins this version of the debate, and Saar's over-confidence in "no way it's anything other than LL, here's the math to prove it, and if you disagree then you're doing the math wrong" does border on pseudoscience here.
To that end, I think Scott's final takeaways are the most important. If we have any shared goals anymore as a society, then we should devote resources to better preparing for pandemics from both origins, and in also improving government/organizational transparency so it's easier next time to understand what went wrong. As it is, either a natural disaster can be ignored because "teh government conspiracy," or else somebody got away with murder; either way, the debates that focus solely on convincing everybody that one's favored position is the only explanation means we're not more prepared for future disasters than we were with this one, we're just more divided and easier to manipulate.
"The final numbers are fine, especially if you use confidence intervals, but you know it's only a snapshot in time using your best data and estimates as of the last time you underwent the exercise; by the time you update in a year/quarter, you will have more data and different estimates. Sometimes, risks that weren't even on your model last cycle, are now the top priority because of what's happening in the world (or because you gained new information that you missed previously). It's a fact of life."
I found this debate to be very valuable. I started with 70:30 odds on lab leak's side (I didn't do any extensive research on COVID origins) but after this debate that shifted to 5:95 (i.e., favoring zoonosis). I wish there are more such debates on contentious topics.
The Rootclaim model seems pretty sensitive to modest differences in probabilities; I'm guessing a factor of having too many variables so small differences across all of them add up to a big effect? I tried to average the six guesses giving Saar the least weight and add some of my own intuition and got (I think) a modest pro-lab leak outcome. https://docs.google.com/spreadsheets/d/14OkrvJ21qCqQcXINQnMO4-a1Xi-npidF/edit#gid=2057332411
> About half of cancer patients lose their hair, and Putin hasn’t, so we’ll divide by two.
This leapt out as methodologically unsound. How do you know whether Putin's lost his hair? Here are some possible scenarios:
1. Putin has never had cancer or anything cancer-like. His hair is fine.
2. Putin has had chemotherapy. He wears, or has worn, prosthetic hair to hide this fact.
Note that these two scenarios are perceptually identical. You have no way to tell one from the other.
If the question you're evaluating is "Is Putin covering up a case of cancer?", his visible hair provides no information at all, because all scenarios predict visible hair. In the "no" case, the hair is irrelevant, and in the "yes" case, the hair is deceptive.
It's kind of like making an adjustment to your evaluation of "Is this a coral snake, or is it a king snake?" for the observation that the snake in question has red and yellow stripes. That observation was already part of the premise of the question; you're double counting (among other weird mistakes) if you adjust for it again.
Haven't you ever watched a movie? You know how the same actor will show up in different movies with different hairstyles? (Or... the same movie, with different hairstyles?)
Generally those are wigs. It isn't possible to tell the difference between a wig and real hair; often they are real hair, just not real hair that the person to whom they are attached grew.
If you had the opportunity to physically manipulate Putin's hair, you might be able to demonstrate that something was off. But you don't.
It reminds me of a Taleb quip directed against only-thinks-within-the-box-rationalists, in one of his books. Someone rolls a dice nine times. It comes up with sice (six) each time. He asks a statistics professor how likely it will come up with six also in the tenth roll. The professor answers "1/6 chance". He then asks Fat Tony, a street-smart mafioso. Fat Tony answers: 100 percent chance. Why? "Obviously, the dice is loaded."
> Another of Saar’s concerns with the verdict was that Peter was an extraordinary debater, to the point where it could have overwhelmed the signal from the evidence.
I've mentioned this before in the ACX comments, but this is such a great lead-in that I'll say again that this is why I didn't make a habit of reading Overcoming Bias. Eliezer Yudkowsky used that blog to make a lot of excellent points in an entertaining and well-written fashion.
And I got the sense, reading his work, that sounding persuasive was a bigger part of his strategy than being right, meaning that reading him was dangerous. So I enjoyed a number of pieces, but I didn't follow him or seek out his work.
My suspicion here is that using estimated probabilities instead of probability distributions leads to the same issue as for the Fermi paradox, in that your result is extremely sensitive to your assumptions and you end up with extreme outcomes on either side like we have here: 10^-25 and 534 are both clearly wrong.
This is a very cogent observation. Like Eric Drexler et al. in 'Dissolving the Fermi Paradox' they realized using point estimates instead of entire distributions lead to weird & paradoxical results. Good job!
My overwhelming feeling while reading this was a sense that a lot of the important evidence is itself subject to long debates and explanations that are being glossed over. Is a certain length of protein sequence likely or unlikely to be done by a mutation in the wild? What is the right way to construct priors about wet market and virus lab locations? Everyone was just stating certain things confidently, while providing detailed documentation for other things.
In the end, “evidence for my claims” seems to have a fractal quality sometimes. The closer you look, the more evidence you need to back up your use of evidence. It’s turtles all the way down.
While I'm terribly sympathetic to most of Wilf's points, one thing I absolutely agree with him on is the superiority of written to real-time/spoken debate, because the things it rewards correlate much more strongly with being right.
It seems to me that Saar's evidence-weighing analysis is basically "outside view" with extra steps.
As I see it, Saar's analysis hinges on the point that the zoonosis hypothesis contains a coincidence that is very unlikely to be explained by some out-of-model error (which I agree with), whereas the lab leak hypothesis only contains a coincidence that could very plausibly be an out-of-model error (I agree with this as well). This sets up a barrier that is almost impossible for any amount of mere digging into the details to overcome, as that details-digging is all itself quite vulnerable to out-of-model errors. And so the end result is an outside-view judgement that no amount of inside-view analysis can outweigh.
This surely seems like an error to me, though those with different positions on the outside-versus-inside-view debate might disagree.
I genuinely do not understand why there is not more talk about the third option: that one of the many bat cave spelunkers from Wuhan caught COVID-19 and brought it back to Wuhan.
The WIV and the market were fairly far apart from each other, as noted in the debate. The Wuhan CDC is *right next to the market*. And while it doesn't seem that the Wuhan CDC was conducting gain of function research, their employees were fanning out all over China going into bat caves.
"In the video, the researchers scale the cavern wall, their headlamps ghostly blue.
“If our skin is exposed, it can easily come in contact with bat excrement and contaminated matter, which means this is quite risky,” says Tian Junhua ["associate chief technician in the Wuhan CDC’s pest-control department, but he has a reputation as a swaggering adventurer in his work with bats and insects"], one of the bat hunters.
...
The video was released by national science authorities and Chinese state broadcaster CCTV on Dec. 10, 2019, and circulated on social media. It’s a high-quality production, designed to promote China’s world-leading viral research.
...
Tian and his team from the Wuhan Center for Disease Control and Prevention are filmed catching horseshoe and pipistrelle bats and collecting samples of guano, in search of new bat-borne diseases and the basis of new vaccines. Tian talks about the need for caution. “It is while discovering new viruses that we are most at risk of infection,” he says, though he is shown handling sample vials without wearing full protective gear.
...
In 2017, Tian told the state-run Wuhan Evening News he once forgot personal protective equipment and was splattered with bat urine, leading him to quarantine at home for two weeks. On multiple occasions, bat blood squirted onto his skin while he was trying to grasp the animals with a clamp, he told the paper."
So the questions would be - exactly when did the Wuhan CDC move to be close by the market? The move finished on December 2, 2019 - how long was the move and how was it conducted? Does the science require an intermediate animal or could it have jumped directly from bats to humans with the cave exploration?
I am now scared that the next pandemic will be started by debaters on either side doing gain of function research to shore up the evidence for auxiliary points in their debate.
There's so much stuff on this page that Chrome is having difficulty rendering it even on a high-end desktop. Consider closing comments and creating a separate page for that.
This is an issue on most of Scott's posts; Substack just isn't designed to handle lots of comments. Given that this has been an issue for years, it seems unlikely they care to fix it. SSC was much better in this respect.
Substack just isn't designed to handle, period. It's unbelievable that a text-based site could possibly run so poorly, and yet here we are, on computers so powerful they can emulate computers *emulating* computers from twenty five years ago, chugging along worse than Slashdot twenty five years ago, Livejournal twenty years ago, Blogspot fifteen years ago, or Reddit ten years ago.
Sites with lots of text and comments are a *solved issue*. I shouldn't have to literally download Firefox just to browse the comments for more than ten minutes without crashing (on Firefox it "only" crashes every twenty minutes or so).
I remember when "nested" comments first came out (different from "threaded" view, which was more like old-school UBB) it could get a little bit laggy... on a Pentium II. A computer far less powerful than the cheapest burner phone out there today. And it still wouldn't crash, per se.
At the outset of the pandemic, I was at about 15% lab leak, 85% natural origin, and thought it unimportant because I didn't think we'd ever have enough evidence to do better than guess. By the start of this year, after way more than 15 hours of studying way more evidence than I imagined we would ever have, I was at 80% lab leak, 20% natural origin, After this, maybe I should update to 70% lab leak, 30% natural origin. But I'm going to indulge in a bit of emotion and make that 25% lab leak, 100% annoyed.
Annoyed in that it seems like the rationalist and rationalist-adjacent community, parts of which at least I value and respect, has decided that on this one issue we're going to flip a $100K coin to decide what to believe, and I suspect anyone who dissents will at best be quietly dismissed as an ignoramus by much of the community. I will grant that by paying $100K, we got a coin that is *slightly* biased towards the truth,
Seriously, one-on-one personal realtime debate is a *lousy* way to find the truth. There's a reason we don't have debates as a central feature of e.g. scientific conferences. Throwing in high-value bets doesn't improve things. The process selects for the most skilled and charismatic debater, and in some cases for the favor of biased judges, with the truth mostly just serving as a tiebreaker.
In this case, there was no tie to break. Miller, by all accounts including this one, is an exceptionally capable debater. Wilf, while clearly skilled as an entrepreneur and possibly as an analyst, does not seem to be very good at in-person debate. From this very thorough account, Wilf completely failed to raise some of the very strong points in favor of the lab leak hypothesis, failed to call out Miller on some of his indefensibly weak points, and agreed to a structure where he'd open by spending the first third of the debate on the very weak assertion, "maybe it didn't come from the Wet Market" even though that is in no way necessary to the lab leak case. Or maybe he did raise the strong points and call out Miller's weak points, but in a way that neither Scott nor the judges noticed.
As for Scott shifting is odds from 50-50 to 90-10 on the basis of this debate, I think that requires p<0.20 for "Wilf isn't very good at this and left the winning arguments at home", which I don't think is defensible. You'd need a statistically significant sample of debates to be confident you have seen both sides well-argued.
I will continue to value and respect the rationalist and rationalist-adjacent community, particularly Scott's corner. But it has a clear blind spot in its enthusiasm for clever and novel truth-seeking mechanisms (bets, prediction markets, debates, etc) on the basis of their cleverness and novelty without adequate regard for their in many cases obvious flaws.
I'm not taking the results of this debate too seriously. I'm taking the fact that the debate exposed me to the arguments on both sides, that I spent a long time grilling the debaters on their arguments, that I read their followups and blog posts, and then that I read many other blog posts and a couple of papers on this topic seriously. Plus the surveys of experts and superforecasters.
The post has long cut and pasted sections of unchallenged borderline dishonest arguments from Peter. The quotes are so long that sometimes I had to double check whether it was you or Peter talking. Many of the arguments come from a paper which is flunking a careful replication attempt, not mentioned. Weissman's probabilities are not included because the items didn't match up with the rows of a spreadsheet built from Peter's analysis or something. There is not much evidence of research beyond the debate here.
I'm glad to hear that and, knowing you, not terribly surprised. But this,
"For what it’s worth, I was close to 50-50 before the debate, and now I’m 90-10 in favor of zoonosis"
does kind of read like it was just the debate. And, as I said, I don't think it would be reasonable to make that kind of shift on the basis of just a debate, because it requires unwarranted confidence in the process and participants. You say that the debate exposed you to the arguments on both sides, but in reality it only exposed you to a subset of those arguments. Ideally, that would be a comprehensive subset including all of the strong and relevant arguments on both sides, but I am skeptical that is the case here. And I'm certain that you have at least some knowledge of missing arguments from your other readings, but most people (including most rationalists) don't.
Unfortunately, as I see this debate spread out into the broader internet, it is largely being stripped of nuance and turned into "Two really smart people held a very thorough debate on COVID origins, and now all the smart people agree that it's almost certainly natural zoonotic origin, nothing more to see here, move along".
If it turns out it was natural origin — as I am more likely to now believe — it’s sad it will be that much harder to end gain of function research. Even if the lab did not leak the pandemic, the risk that the lab would leak a pandemic eventually was still too damn high. Hopefully this event will nonetheless lead people to take the risks more seriously.
I’ve spent some time in virologist subreddits since covid, and I find their attitude towards these risks to be terrifying. They remain very blasé about it. It appears to be deep in their culture, similar to how civil engineers are culturally committed to designing for cars, and architects are culturally against ornamentation.
The science has to decide whether it’s a lab created virus or not. However created in a lab is not the same as lab leak. In fact your prior for this being a leak should be close to, or at, 0% since it’s the only time a major pandemic has happened in that fashion.
When you look at the distance between the lab and the market and the fact that the origin is clearly the market if lab created pathogen is proven true then the most likely explanation is an external attack.
> So for example, suppose there are fifty things about a virus. You should expect at least one of those to have a one-in-fifty coincidence by pure chance.
Minor math error here I think: You should expect *exactly* one, not at least one. It's equally likely to be more than one or less than one.
Thank you to Saar for running this in good faith, and thank you to Scott for a great write up. One of the key things that stood out to me while reading the analysis was that there are a few relatively simple factual items that were disagreed on that might be relatively easy to chase down if we weren't worried about governments trying to obfuscate things. There's also a few simple pieces of information that could shift my thinking a bit too, like where do the lab workers live? I wouldn't expect a lab leak outbreak to be centered around the lab, but around one of its employees homes.
I noticed in myself a high risk of falling victim to a classic 'david and goliath' story too, which is potentially a very bad bias. Almost everyone reading that first section is going to be rooting for Peter, and enjoy it everytime he seems to win a point.
I also found myself with the David vs Goliath bias, despite having no opinions on COVID origins and not particularly caring about it.
Despite this, Saar’s arguments (as described here) often ended up fairly weak due to him leaning on what seems like a lot of “bad claims” about things like the pre market cases that Peter easily dismissed just by following the logical consequences that should have occured if they were true. I think this ends up with him showing weaker than his argument really was.
Also agree, huge props to Saar for doing the thing in good faith, and hopefully this taught him a lesson that when someone bets a large % of their NW versus your proportionally small bet, they will prepare 10-100x as hard as you will (because they have so much more to lose), and the fact that they took the bet in the first place implies they are likely already exceptionally good at whatever you are betting on.
I must admit that I don't care nearly as much about whether COVID was zoonotic or not, as I do about the government paying social media sites to enforce a particular orthodoxy about the question.
Some of this sounds like an “Encyclopedia Brown” level of argument. E.g. the idea that because the University of North Carolina said they’d do furin site insertion in a grant application from years before the pandemic, therefore the WIV cannot have done any insertions of their own.
Research institution capabilities, and researcher ambitions, and the real world generally, are a lot more complicated than that reasoning implies (I say this despite putting little weight personally on COVID being engineered).
Likewise with the idea that the WIV would have fully released the details of all viruses they have sequenced.
I haven’t watched the debate though, so maybe it was better than Scott makes it sound.
It strikes me that whether you should debate theories you think are crazy depends on how popular they are. Consider:
1/10000 believe in theory A. You debate someone arguing A, and you win 99% of the audience (considering what humans are like, this is super impressive). But now, 1/100 of the audience believes in the crazy. And *no-one* can convince 99.99%.
66% believe in theory B. Now when you debate someone arguing theory B, even if you only convince *half* the audience, things have improved!
This, of course, assumes that the people believing what you think is crazy can be convinced in the first place. This might well be true about lab leaks or *possibly* even UFOs; it will *not* be true about creationism (evidence isn't even a factor in why someone is a creationist).
The JFK assassination is another of those cosmic jokes. Sensible people should not believe in conspiracy theories about it. But dear Lord, was Lee Harvey Oswald a weirdo, the rare person who had been involved with *all* the kinds of people who _might_ be suspected of conspiracies; and murdered before he had a chance to testify; and with an investigation that couldn't have seemed more suspicious and cover-up:y if it tried.
If a lone gunner killed a reformist Russian president, if the gunner was himself killed by a man with mafia links and if that man died of cancer a few months later we would all be fairly conspiratorial.
Really fantastic read. I already leaned towards zoonosis but this removed most of the remaining doubt. It so valuable to see one side put up their strongest arguments and then see if the other side can knock them down, which in this case they did pretty dramatically.
“… And so on and so forth, until we end up with the final calculation: 86% chance Putin doesn’t have cancer, too bad.”
What Rootclaim is doing is interesting. But should not such hard-core Bayesianism also include sensitivity analyses? That is, investigate how robust the conclusion is (here: “86 percent chance”), at least for minor-to-moderate changes in priors, age-related cancer statistics, and the like?
Using sensitivity tests in cost-benefit analyses is usually recommended. It appears that Bayesianism could benefit from such tests, for the same reason.
It would also quell the skepticism people like me automatically feel when encountering predictions spelled out in fine-grained percentages. (Plus, why stop at percentages? Why not add a few decimal places as well: 86,xy percent chance Putin doesn’t have cancer– that would make the prediction even more “accurate”….)
Do not get me wrong – I respect what Rootclaim is doing. But I would be (genuinely) interested in why Rootclaim does not add/construct some kind of confidence interval around their percentage-estimates. (And why Bayesians in general do not do this.)
This post caused me to strongly update against Bayesian methods. Maybe it's the impossibility of seeing accurate priors, or the tendency to engage in cherry-picked absurd probability estimates. Or maybe it's the qualitative difference between statistical probabilities prospectively estimated vs retrospectively measured. I no longer have confidence this method produces more signal than noise.
For example, what prior do you set for an outbreak in WSM? You might think that's low, but given multiplicity you're not really looking at a prior probability anymore, are you? Both a spillover event and a lab leak might have come from any number of places. We'd tell some plausible-sounding story about why that place was unique, then set absurd priors as a fence around that story. But at heart it's still a post-hoc story, and all the prior-setting in the world will only serve to reinforce your story with more plausible-sounding stories. This seems to be what Saar is fighting against in making adjustments for out of model variance, but that seems like the wrong approach.
The better approach is probably abandoning Bayes - at least explicitly/strictly. I'm not yet convinced about the whole Bayesian enterprise, but it feels like it's on much more shaky ground for me today than it ever has since I first learned about it.
"This post caused me to strongly update against Bayesian methods."
This is also my major takeaway from the post & from the comment section.
I sometimes teach students who enter occupations where they have real power over other humans. Such as child protection officers, who in my country have a lot of influence when deciding in a child should be taken from its parents and moved to foster care.
I have toyed with the idea to include some "real" Bayesianism in their curriculum, with regard to calculate the risk of various forms of child abuse in various types of families. (As of now, I only make references to what I half-jokingly label "everyday bayesianism", which is basically what Scott talks about early in the post - also captured by the expression "inference to the best explanation".)
This post, and this extensive comments section, has made me update (as they say here) strongly against this idea. Few things are more dangerous than professionals with real power believing that "science" can give them analytical tools that makes if quite certain when sufficiently serious child abuse goes on, that drastic measures are called for (and can be defended in the courts).
Putting numbers on inherently uncertain things creates professional peace of mind. It is a temptation in all professions given power from the state to dramatically interfere in the lives and health of other human beings. So no wonder that "Bayesianism" is an increasingly popular catch-word in the helping & caring professions, as elsewhere. But judging from this blog post & comments, "real" bayesianism is a danger, not a help, in reducing type I errors (not interfering when one should) and type II errors (interfering when one should not).
After thinking it over more thoroughly, I'm not sure the problem is that Bayesian thinking is wrong, per se. The problem is that some of us have been convinced to apply it more broadly than it should be. (Here I'm thinking of Scott's "P(A|B) = [P(A)*P(B|A)]/P(B), all the rest is commentary.")
Simple heuristic: before you try to apply Bayesian reasoning, ask yourself, can you accurately fill out the equation? "Not quite, but I think I can estimate-" No. The answer is no. Move along, or you'll be getting yourself into trouble, trying to apply (even loosely) principles that will only lead you to error.
I personally agree it was not a lab leak but a pretty important was lost in the debate (or at least poorly factored in). Namely China was making hiding evidence. While this may impact priors…the bigger impact is that, if it was a lab leak we only know what information was released (which almost certainly would be anything that boosted their preferred narrative) and do nit have all the evidence that was presumably withheld (which would be all the evidence they could suppress that went against the preferred narrative).
It’s kind of like sample bias in that the evidence you use to make the assessment is systematically biased.
I unsure how you assess these probabilities because they are fundamentally unknowable.
I did a poor job articulating this and want to highlight that I do not presume to know what the CCPs preferred narrative even is and it could have changed. They could and have started by withholding all evidence at the local level (this does appear to have happened) but then at higher levels decided on a new narrative. We just don’t know and cannot reliably intuit this with any degree of precision
The point is that we can assume the evidence we are using is biased in a systematic way but have no reasonable way to assess this.
"maybe you should lower your disbelief in each hypothesis to something more reasonable, like 99%. But now the chance that the sun rises tomorrow is 0.99^100, aka 36%. Seems bad."
All these hypotheses are very correlated; extreme probabilities of base events are not required, just extreme probabilities of conditionals.
'If all 99 other hypotheses are false, and I'm not making some silly maths mistake or whatever, then hypothesis 100 is 99.999999% false' is not necessarily overconfident even for a human.
I have done my own extensive research and believe the evidence is compelling that there was a lab leak. IMO, Peter has a real axe to grind and his strong personal conviction against the lab leak hypothesis has made him blind to common sense.
The odds that the virus evolved free of human-engineering are probably much less than in 1/10,000. Peter has no idea how to figure those odds and frankly no one else does either. The actions of the Chinese government in stonewalling are far more convincing to me that there was a lab leak. If the Chinese government didn't have something to hide why be so opaque???
For an adversarial setup (such as debates or US court cases) to converge on something close to the truth, the case for each side needs to be represented at a much more similar level than was done here. Even if an innocent man is accused of a crime (so he *knows* that the truth will set him free), it is still generally recommended that they hire a competent lawyer to represent their side, as opposed to defending themselves. This is probably not as important in countries with a less adversarial court system.
Calling Peter a better debater doesn't quite capture what happened here. I don't think there was a great difference in rhetorical skill, but there was an extreme difference in both the level of preparedness and the commitment to the cause. This case involves 50+ pieces of circumstantial evidence of various degrees of strength, some favoring lab leak and some favoring zoonosis. All the ones favoring lab leak were heavily contested, but at least to me it did not seem like the same was true for many of the ones favoring zoonosis. This was probably a mixture of a few things. Saar didn't seem to not know enough of the precise details to effectively push back against some of the bolder zoonosis arguments, such as how confident we should be that we actually have a good picture of the first cluster. In other words, he conceded far too much, allowing the zoonosis side to look much stronger than I suspect many people still believe it is. Secondly, I don't think Saar fully understood what his "role" in this debate was. Instead of being a soldier for his side, he mostly stuck with his 2-3 year old argument for what he thought the "truth" was. Peter, on the other hand, seemed eager to actually win the debate, so he made the maximally strong argument for his cause, not granting anything, as opposed to trying to converge on the "truth". So for every piece of evidence, he was able to put on the "maximally pro zoonosis" spin, perhaps resulting in his absurdly low final lab leak probability.
It is perhaps interesting to think about how this compares to adversarial collaborations, which are popular in our community. Using the language of Julia Galef, one can perhaps think of an adversarial collaboration as a "dance" between two scouts. Similarly, a debate can perhaps be thought of as a battle between two soldiers. The incentive is to win, not to find the truth. The rootclaim debate looked like a battle between a soldier and a scout, with the predictable result.
(Disclosure that I support in BiosafetyNow, but these are my views and don't necesarily reflect those of BN)
The Twitter origins debate, which Scott understandably shies away from (ref. his most recent tweets, and ofc I certainly don't blame him) represented the richest, most boisterous most tumultuous, most rancorous, most acrimonious form of the origins debate I've ever encountered online. The current ACX substack post & its comments don't reflect the Twitter debate's diversity of opinions, evidence, and individual credibilities IMO.
The key items I feel are insufficiently highlighted (even if they are mentoned at least once) from my scan of the comments section:
• the value of 'adverse inference' facing:
° PRC's gov
° the most outspoken virologists (+ 1 evolutionary biologist, + 1 disease ecologist, and yes not *all* virologists) especially when relevant materials turned up from their side only following subpoena
• No progenitor or 'sister' virus with an MRCA dated to shortly before outbreak has been found in animal populations despite extensive sampling.
• I don't feel any post-outbreak (and some near-pre-outbreak) data out of the PRC should be accepted on faith; rather it & all conclusions drawn therefrom should be placed in its own little suspect epistemic box. By contrast, any data in non-PRC hands pre-outbreak is 'gold'. The remaining quadrant, data & analysis produced in the ordinary course ex-PRC but post-outbreak, is tainted by the need to avoid 'rocking the geopolitical boat'. Often we can't reference such data that for example underlies post-outbreak intel findings, like that of the ODNI (atop DoE, FBI, CIA DIA...) so I can only give it the 'silver' label treatment.
And transcending all these quadrants, verified data that turned up ex-PRC from leaks, FOIAs, sequences reconstructed from SDAs, etc. is 'platinum'.
1. Peer reviewed papers since the debate that show ascertainment bias in early case data and intermediate genomes that undermine the multiple spillover theory indicating lineage A arose first so the market cases are not the primary cases (Weissman 2024, Lv et al (2024), Stoyan and Chiu (2024).
2. Miller's claims about the furin cleavage site are dubious:
Miller incorrectly claimed the N501Y mutation would result from passage in hACE2 mice (he mixed them up with BALB/c mice). WIV was performing in vivo experiments in transgenic (human ACE2 expressing) mice and civets in 2018 and 2019 in SARS-like CoVs.
I think both sides may have also overlooked the MERS furin cleavage site shares several structural and functional similarities. The sequence looks quite similar. In 2019 WIV researchers were involved in MERS research. Yusen Zhou who died in mysterious circumstances in May 2020 and was involved in an early Covid-19 vaccine was also involved in this research that involved manipulating the FCS. t.co/7zcSUPR60T
Broad Institute biologist Alina Chan also observes the S1/S2 FCS PRRA insertion in SARS-CoV-2 generates a Class IIS restriction enzyme site (BsaXI). This was used by WIV and Ralph Baric at UNC previously.
Dr Andreas Martin Lisewski discusses similarities with a MERS infectious clone described in 2017 here. The argument no one would engineer an FCS like this overlooks the MERS example.
I think it's wrong to characterize this as a tragedy for Saar and Rootclaim. Sure the contest didn't validate their methods -- but I think their purpose was bigger than just winning internet arguments. Their goal is basically furthering scientific knowledge. Every model and method is flawed, but some are still useful. I see this as a heroic story, and Saar's stubbornness doesn't negate his accomplishment. By gracefully hosting and funding this exercise, he's made us all better off and advanced our knowledge of important issues. And that is the most important thing.
So the Chinese want us to think that they caused a global pandemic?
Although not infallible, humans are actually pretty good at ascertaining the intentions of enemies and potential enemies, and that is why it is a universal human inclination. Being oblivious to the likely motivations and intentions of other people is a big evolutionary disadvantage.
> Newtonian mechanics wasn’t pseudoscience when Newton discovered it, but if someone argued for it today (against relativity), that would be pseudoscientific.
I disagree strongly with the ontological assumptions here. Newtonian mechanics and relativity are both great theories. Newtonian mechanics is superior for terrestrial, human-scale phenomena because it provides a simpler representation of the physical laws. Relativity is superior at speeds approaching c because the complexity is justified by the accuracy gained. To claim that Einstein > Newton for modeling the trajectory of a baseball would be the same error as saying that Newton > Einstein for modeling a black hole: both arguments would be domain-of-applicability errors.
I self-identify as a flat earther, because for 99% of daily life, the flat earth theory (viz., conceptualizing the Earth as a flat plane) is simpler and more useful than conceptualizing it as a globe. The error made by "actual" flat Earthers is not that their theory is bad - they have an excellent and very useful theory - it's that they misunderstand their theory's domain of applicability. Flat Earth theory gets you to the grocery store; global earth theory gets you to the moon.
To state my core point a bit more explicitly: the "best" theory (for any particular purpose/problem) is that which most simply encodes all information that's relevant (to the particular thing being studied). In some cases this is Newtonian mechanics; in others, it's relativity.
I think most points here are good points to make, but I also think it's useful as a general caution against this type of exercise being used as an argument at all! So I'd obviously caution against anyone taking your response itself as a reasonable attempt at an estimate of the "correct" Bayes factors, because this is all very bad epistemic practice! Public explanations and arguments are social claims, and usually contain heavily filtered evidence (even if unconsciously). Don't do this in public.
That is, this type of informal Bayesian estimate is useful as part of a ritual for changing your own mind, when done carefully. That requires a significant degree of self-composure, a willingness to change one's mind, and a high degree of justified confidence n your own mastery of unbiased reasoning.
Here, though, it is presented as an argument, which is not how any of this should work. And in this case, it was written by someone who already had a strong view of what the outcome should be, repeated publicly frequently, which makes it doubly hard to accept the implicit necessary claim that it was performed starting from an unbiased point at face value! At the very least, we need strong evidence that it was not an exercise in motivated reasoning, that the bottom line wasn't written before the evaluation started - which statement is completely missing, though to be fair, it would be unbelievable if it had been stated.
Counterposition here- given the verified attempts of multiple authorities to shape the "narrative" around covid, plus the powerful incentives for them to do so, does this mean that the sum of evidence available to assess the situation is probably unreliable and unfit for the purpose of this debate? The added layer of information control typical of the Chinese government makes the situation even more opaque.
Maybe this isn't a debate which can be settled based on the available information. Maybe that is by design.
The real question for me is this- why is any of this important in the long run? The obvious answer is that it has some bearing on how the world might respond to future pandemics. If it is yet another zoonoses then all the efforts to date to prevent them didn't prevent this one and probably wont prevent the next one, so maybe we should attempt to apply democratic pressure to make governments do more. I'm not sure they can do much more either way.
If laboratory research is behind covid, then this is also not completely unprecedented (though the level of genetic engineering might be somewhat new). Again, the technology isnt something that can be easily regulated by governments. Nucleotide sequence synthesis is relatively centrally controlled by a few large companies with the technical capacity, but that is only one of many ways to generate genetically modified pathogens. Again, we could attempt to apply democratic pressure to democratic governments, but again that would likely have limited impact.
The only thing that really shines through with the whole covid affair was the government and health authority overreach, panic, incompetence and opportunism. Once again none of this is new. The corporate capture of the medical system has been progressing for decades. At best the whole covid affair is just another reminder that authorities do not often have the interests of the individual as their main priority, and even when they do their capacity to deliver functional outcomes is usually highly limited. All trends point to this situation worsening in the future.
Novel pathogens will likely always remain potential black swans, which means that most of the time authorities will over react to their emergence relative to the actual risk posed. Over time this will probably build up broad social resistance to government responses until a genuinely threatening pandemic emerges. The relative difference in perceived threat between pathogens that are novel and fast moving versus those that are familiar and slow (such as tuberculosis) is also worth contemplating. Our information ecosystem thrives on novelty. Nobody is going to buy a newspaper that reports the same steady rate of slow deaths from tuberculosis or malaria. No government minister is going to win the fanatical devotion of the masses by heroically fighting such unexciting killers.
>"Peter seems to have a photographic memory for every detail of every study he’s ever read. He has some kind of 3D model in his brain of Wuhan, the wet market, and how all of its ventilation ducts and drains interacted with each other."
Mostly I think it's just keeping up with the Standard Arguments™️, from papers and preprints by zoo crew (Holmes, Andersen, Garry, Rambaut, Worobey, Pekar, Rasmussen, Débarre, Crits-Christoph, Neil etc.—the "Proximal origin" authors and their collaborators). They're the group of Western virologists who ~yearly come out with a new paper finally proving market zoonosis for real this time (Andersen et al. 2020, Holmes et al. 2021, Worobey et al. 2022, Pekar et al. 2022, Crits-Christoph et al. 2023a, Crits-Christoph et al. 2023b, etc.) and get a ~yearly media circus about it. Maybe he also read some pop-sci books like Chan & Ridley's "Viral" for pro-lab (2021) and Quammen's "Breathless" for pro-zoo (2023). Probably also the Chinese HSM-sampling paper (Liu et al., 2023) and WHO-China joint study (2021), following zoo crew in how to interpret them.
E.g. Peter's heat map slide is just Fig. 1 of (Crits-Christoph et al., 2023a), which is an elaboration of Fig. 4 of (Worobey et al., 2022); and his "drains" argument is from Fig. 2C of (Crits-Christoph et al., 2023b).*
Bloom, JD. (2023). Association between SARS-CoV-2 and metagenomic content of samples from the Huanan Seafood Market. Virus Evolution 9: vead050. https://doi.org/10.1093/ve/vead050
Chan, A. and Ridley, M. (2021). Viral: The Search for the Origin of COVID-19. HarperCollins.
Crits-Christoph, A. et al. (2023a). Genetic evidence of susceptible wildlife in SARS-CoV-2 positive samples at the Huanan Wholesale Seafood Market, Wuhan: Analysis and interpretation of data released by the Chinese Center for Disease Control [preprint]. Zenodo. https://doi.org/10.5281/zenodo.7754299
Crits-Christoph, A. et al. (2023b). Genetic tracing of market wildlife and viruses at the epicenter of the COVID-19 pandemic [preprint]. bioRxiv. https://doi.org/10.1101/2023.09.13.557637
Pekar, J. et al. (2022). The molecular epidemiology of multiple zoonotic origins of SARS-CoV-2. Science 377: 960-966. https://doi.org/10.1126/science.abp8337
Quammen, D. (2023). Breathless: The Scientific Race to Defeat a Deadly Virus. Simon & Schuster.
Worobey, M. et al. (2022). The Huanan Seafood Wholesale Market in Wuhan was the early epicenter of the COVID-19 pandemic. Science 377: 951-959. https://doi.org/10.1126/science.abp8715
*The heat map is misleading: it was based on implausibly assuming all sampled stalls were sampled equally; in fact the wildlife-corner stalls were sampled far more heavily (Liu et al., 2023). For a critique of dismissing ascertainment bias, see my (Bahry, 2023). For a critique of Crits-Christoph et al. 2023a, see Bloom (2023). For a critique of their later, also-misleading heatmap (Crits-Christoph et al. 2023b Fig. 2a), see (https://www.biorxiv.org/content/10.1101/2023.09.13.557637v1#comment-6279754825).
I find the "drains" argument mildly interesting (Crits-Christoph et al., 2023b Fig. 2c): nonzero evidence, but just one real datapoint. (The two downstream of the racoon dog stall are *in general* downstream of a big chunk of the western wing and the entire western wing respectively, so those mean little.])
I guess I don't quite understand your point. Scott points out that Peter has an eidetic memory for the arguments in this debate, and you respond: "Well, he just memorized the studies and arguments, especially those that support his view."
Aren't you saying the same thing as Scott? Surely, we wouldn't expect Peter to actually author the papers himself. So what else is there to do than defending his position using other people's work?
And if Peter's understanding of the research is actually much shallower than it looks to a layperson (which is I think what you're saying), that seems fairly embarassing for Rootclaim, no?
My point is that Peter was well-prepared—but it was the high end of the normal range of preparedness for someone who'd been following the topic, not savant-level (compare to Scott's "I can’t imagine how many mutations it would take to make me even a fraction as competent as Peter was").
E.g. to make the drains argument, he didn't need a working 3D model of every brick and rivet in the market; he just needed to follow zoo crew on Twitter, see one of them share Crits-Christoph et al. 2023b, read it make the drains argument, and see its Fig. 2C.
Note that this comment isn't intended to disparage Peter, just to bring things down to Earth—and to give some pointers on where the arguments came from so they can be seen in their original publications, so readers can evaluate them further, etc.
Scott- I can roughly translate my factors into the grouping you use in your table.
Net priors using your grouping)
me: 12 you: 16
combo of all HSM/lineage factors
me: 1 you 0.002
FCS-ish
me 25 you 25
reasons WIV wouldn't do it
me: in priors you: 0.17
cover up success
you 0.5 me: did it?
other factor restriction enzyme pattern
me 70 you NA
The RE pattern is new, and I think you'll find my argument at least reasonable.
The big disagreement is about Worobey/Pekar. If I were to argue for ZW, I'd stay the hell away from Pekar, whose errors are now verging on something worse than "wrong".
Worobey does correctly point out that there was a spreading event at HSM, but so many specifics (lineage, internal DNA-RNA correlations, Wuhan share of relevant species (if any) don't fit that ZW has to rely on the non-market versions, which actually were generally considered before hand to have more of the priors than a market version- including by DEFUSERs.
Why don't I then get stronger LL odds? There's a lot of uncertainty in the priors. I do "robust Bayes" to take that into account, integrating over a broad fat-tailed distribution of priors. One tail keeps ZW in the game. Changing likelihood factors makes much less difference in the final odds than you might expect, because the enormous likelihood ratio has already been diluted down by the uncertain priors.
Lots of people in the comments seem to have additional evidence one way or another that they want to add in. Maybe we need a coronavirus version of TalkOrigins.
I just cannot get myself to the point where I take anything being stated about this as reliable. For 2-3 years, everything that was being said about COVID by "experts" was either on what Zvi calls Simulacra level 2 (the public health establishment; saying X, despite ~X, to make you believe X and act in a way they want you to) or on what he calls Simulacra level 3 (Twitter users saying something not because THEY believe it's true, or even that they expect YOU to believe it, but just to signal that they're good non-racist Science-lovers). So even while I find Peter's points as you've summarized them to be pretty reasonable, he hurts himself IMHO by making the characterization of lab leak as a "conspiracy theory", because accusing an opponent of that also happens to be great tribal signaling. And the data on which all of this is premised comes from sources like the WHO which is clearly willing to lie for a variety of reasons.
So despite finding Peter's arguments here pretty reasonable, I can't get over the hump. I also, were I as dishonest as the PH establishment, would be inclined to just say it was a lab leak to induce public fear of gain-of-function research to help eliminate it. I can live with that having been wrong for this one virus, but still wanting to ban such research because next time it may NOT be wrong, but since most people don't reason that way then zoonosis being true is quite a blow to efforts to end this dangerous stuff. So let's give public health a dose of its own medicine and keep a nice healthy percentage on lab leak regardless.
Even after this excellent post, I find Michael Weissman's v5.7 analysis from March (https://michaelweissman.substack.com/p/an-inconvenient-probability-v57) to be the most compelling information available online to date, and it contains significant evidence updates from his v5.6 post that you linked in the write-up from January. I've kept an open mind through the years about Covid origins and found the Rootclaim debate to be especially interesting and cause for shifting my probabilities, but I have not encountered anything that has so significantly updated my probabilities than Michael's latest v5.7 analysis does. I'm curious if people have read that analysis in full and what everyone's thoughts are.
You seem to initially be interested specifically in Michael's Bayesian probabilities to compare his results to your table (which I think he posted recently in the comments above). I understand that you likely have dedicated dozens of hours to this post already and reading through a massive updated analysis is likely not especially appealing, but just looking at the probabilities alone does a disservice to the immense amount of research, evidence, and analysis that is included in Michael's write-up. For what it's worth, I've read an incredible amount of high-effort research and debate on this topic throughout years of following this, and Michael's research/analysis is the first one that I find convincing. In addition to more convincing Bayesian logic, his research includes crucial compelling evidence/information that came out after the debate (or was available but dismissed or potentially misrepresented at the time of the debate). I am trying to be as fair to both sides as possible and I honestly can't see how someone can read Michael's v5.7 post in full and come away with a 90% probability of Zoonosis. In my mental model, the two most likely reasons are that either people haven't read Michael's 5.7 analysis due to low exposure of a small blog, or I am fundamentally misunderstanding something.
Either way, I want to know what people think. If Michael's post isn't as convincing as it seems to me, I'd like to know why so I can update. If it is convincing, I'd like others to know so they can update. @Scott, I historically have placed a lot of value on your opinions and I'm genuinely interested in what your thoughts are here.
Personally, I don't have a lot of emotional/ideological investment in the issue so wouldn't sit through the debate but it great to get a detailed enough write up of the main arguments and their problems. I guess I should be over it, but I remain amazed by the level and vehemence of the responses here and elsewhere. Lab leaks remain possible whether or not it happened here so there's an ongoing risk anyway. There are plenty of risks. In the long term, I'm actually more worried about the intentional release of an engineered virus than an accidental leak.
100k is about 0.05% of Saar's net worth. I assume it is a much much larger share of Peter's net worth. More to lose. More to gain.
The incentive for debate performance is massively skewed toward Peter in this case. Worth noting when analyzing the truth-seeking utility of this format.
Sorry if this is noted elsewhere, it is a big comment section.
This was such an interesting read. There are a lot of smart people here in the comments continuing the discussion, but I just wanted to chime in and say THANK YOU for writing something so compelling to read for somebody (me) who has no prior knowledge or real opinion on the subject(s). Legend ⭐
I believe it is absolutely impossible that an animal or animal specimen transported to Wuhan lab for study caused the pandemic.
I only believe that an animal or animal specimen transported to the Wuhan wet market to be EATEN, could have possibly caused the pandemic. Obviously.
I believe the Chinese govt is LYING when they say no pangolins, bats or racoon dogs were sold at the market at that time. I believe the Chinese govt is LYING that the only COVID found in the market were on fish, who can't get COVID .
I absolutely believe the Chinese government when they say with exact certitude when and where the earliest cases were, so I can see my dots on the map. The cases neatly surround the fish on the map where the COVID was found.
I believe the Chinese govt when they say they took down their virus database in Sept because of fear of hackers. Cause you know, hacking had just been invented. I also believe the Chinese govt when they say they increased safety protocols at Wuhan just because. I understand the Chinese govt blocking attempts to get to the bottom of this and silencing doctors and scientist who tired to warn the world.
I believe the former head of the CDC, the FBI and the Department of Energy are all wrong on their beliefs of the Covid orgin.
"Either a zoonotic virus crossed over to humans fifteen miles from the biggest coronavirus laboratory in the Eastern Hemisphere."
As they say, 'you couldn't make it up'. Or, 'truth is stranger than fiction'. It does *seem* very suspicious co-incidence that a new virus that would turn into a world-wide pandemic just happened to pop up out of nowhere on the doorstep (as it were) of a specific institution dedicated to doing this kind of research, but there was no connection. And yet we seem to be forced to accept that this is so.
Congratulations and gratitude for sitting through 15 hours of video to produce this post! And I did appreciate the Linear B/Lineage B joke.
Isn't it a reverse correlation issue though, the lab is situated there because it's an area where coronaviruses were found in the past.
No. Wuhan is both far from bat habitats and far from south China, where wildlife consumption is popular.
I would also add that concerns about a lab leak in Wuhan are not post hoc. There were concerns about lab safety in Wuhan in the scientific literature published from 2015 onwards. If you were to ask people in the know prior to the pandemic - where in the world would there most likely be a lab accident with a coronavirus, Wuhan would have been top of the list given that it has the world's largest collection of bat coronaviruses, and did incredibly dangerous experiments on them at BSL-2.
Concerns about the wet market are not post hoc either. An Australian virologist posted pictures and video from visiting the HSM back in 2014, noting it would be an ideal breeding ground for coronavirus pandemic.
Actually, Wuhan was used as a negative control for coronavirus antibodies in a WIV study prior to the pandemic. i.e. It was thought a v unlikely place for a spillover prior to the pandemic.
Shi Zhengli has said she never thought a spillover would occur in Hubei
Also, the Australian virologist you refer to (Eddie Holmes) was in Wuhan because he was visiting the Wuhan Institute of Virology! He collaborated with the WIV on RatG13, one of the closest known relatives to covid, in 2018 (and then apparently forgot that he had done this_
...and therefore... what? I don't understand your point here at all. I know he was visiting WIV. We already knew about RatG13. So?
Gonna combine 2 threads to reduce confusion:
> Shi Zhengli has said she never thought a spillover would occur in Hubei
Did she think a spillover would occur in the WIV? It's kind of weird to have this sentence right after your prior one... are we just cherry picking experts now? Some people identified wet markets, including HSM, ahead of time. Some people identified WIV ahead of time. Seems like kind of a wash to me.
There's extensive bat populations in the cave systems of Enshi in western Hubei aren't there? And the whole province is within the range of the rufous horseshoe bat. There wouldn't be large populations in the area around Wuhan itself since it's heavily urbanized, but hunters would bring them into the city to sell them given that its the provincial capital and richest city in the nearby area by a far margin.
Also raccoon dogs are common across rural areas of central and eastern China and are another possible source.
Obviously I don't know what the internal reasoning of the Chinese CDC was for deciding where to put their research into coronaviruses. But Wuhan, as 7th largest city in China, fairly centrally located, and in the region of the relevant animals would seem a good choice. Chengdu is the only larger city in central China, but is much further west and has less tech sector, so presumably would be harder to get good people
You are on the right side for the wrong reasons...
https://jimhaslam.substack.com/p/6-if-you-believe-the-wiv-engineered
You also couldn't make it up that the virus appeared in a wet market that was identified years ago as a plausible place for a pandemic to appear. These kinds of coincidences are the rule, rather than the exception, for real-world events.
Re-reading your comment, I realize you meant that as a joke, sorry.
I was once at Disney World (over a thousand miles from my hometown of population 3000) and met a person from my hometown who also happened to be there that day. We had 0 idea the other was going.
This is a lot less weird than if I met them at a random gas station in rural Mongolia, but it's still pretty weird!
While on a high school trip, my classmates & I ran into someone else from our class (not associated with the extracurricular organization that had arranged the trip) on the second level of the Eiffel Tower.
I once was on random free city tour in another country ~5,000 from my home with maybe 10 people in total and met another person who lived within five miles of me back home (same suburb, so not like a city of millions either).
Or as Jon Stewart put his perfectly natural skepticism:
“‘Oh, my God, there’s been an outbreak of chocolaty goodness near Hershey, Pa. What do you think happened?’ “Like, ‘Oh I don’t know, maybe a steam shovel mated with a cocoa bean?’ Or it’s the [expletive] chocolate factory! Maybe that’s it?”
How many labs do coronavirus research?
I've seen the argument that it's suspicious that COVID appeared near one of China's only two BSL-4 labs, but in this debate the *proponents* of the lab leak hypothesis said that the WIV recklessly did coronavirus research in a BSL-2 lab, not in the BSL-4 lab. (The WIV has a BSL-4 lab as well as BSL-3 and BSL-2 labs.) It's less suspicious that COVID popped up near a coronavirus lab if dozens of labs do such research than if this is one of the few such labs in the world.
Yeah, the emphasis on the pro-lab leak side puts less emphasis on the BSL-4 designation and much more weight on "there are/were two labs that were at the forefront of bat coronavirus research, WIV and one in North Carolina." Plus WIV seems to have been doing GoF research.
15 miles is a long way away for it to travel without it being found first close to the lab or close to the families of the lab.
Cities are designed to make distance and geography irrelevant. They aren't perfect about it, but the intent is that your interactions are defined by your social and economic network, and the physical infrastructure allows you to hop between those nodes without concern for what boring irrelevant-to-you terrain is between them. You probably barely notice that stuff when you're on the freeway, and you can't see it at all from the subway,
Again, that's only an ideal which is imperfectly approached. But what we need to know isn't "what's the distance from the WIV to the HWSM", it's "how many WIV lab techs have families with a taste for fresh seafood or exotic bushmeat?" If the answer is "lots", then the 15 miles won't stop them and/or their wives from schlepping their virus-laden selves over to the market.
Or maybe there's some class or cultural divide that means nobody working in the WIV's labs is going to shop at the HWSM; I don't know. That information, which would be very informative here, is hard to come by. The geographic distance, which isn't nearly as informative, can be read off any online nap; a few clicks and calculations, So that's what most people use; looking for their lost keys under the streetlight.
My prior, starting from ignorance, is that the odds of an average WIV lab tech('s wife) visiting the HWSM, is the same as the odds of an average Wuhan resident visiting that market. Which is something we can calculate.
> Cities are designed to make distance and geography irrelevant.
They really aren’t. Cities are the slowest way to travel from A to B by car - which is what we are talking about with the 15 miles. It’s shorter as the crow flies. As for public transport we would have to look
Into it I suppose but if the lab Typhoid Mary was taking the infection to the wet market on a train then you expect the spread of the disease along the track of the line.
My belief is a zoonotic answer but if the virologists come out with a conclusive peridot that it’s artificial the lab leak isn’t the only game in town.
Several recent papers challenge the core arguments relied on for Huanan Seafood Market origin in the Rootclaim debate. Also, Miller incorrectly claimed the N501Y mutation would result from passage in hACE2 mice (he mixed them up with BALB/c mice).
1. Samson et al (2024) estimate likely point of emergence between August and October 2019. Rendering the December 2019 cases and January 2020 market samples irrelevant to origins.
They refer to horizontal gene transfer of the spike gene from BANAL bat CoVs in Laos to pangolin CoVs occurred in mid 2018. "Horizontal gene transfer" can occur in the wild (rarely) or can be induced in he lab artificially "transgenic". In other words a nice code word for lab engineered virus leak.
Samson S, Lord É, Makarenkov V (2024) Assessing the emergence time of SARS-CoV-2 zoonotic spillover. PLOS ONE 19(4):
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0301195
2. Dietrich Stoyan, Sung Nok Chiu, Statistics did not prove that the Huanan Seafood Wholesale Market was the early epicentre of the COVID-19 pandemic, Journal of the Royal Statistical Society Series A: Statistics in Society, 2024;,
https://academic.oup.com/jrsssa/advance-article-abstract/doi/10.1093/jrsssa/qnad139/7557954
3. Lv et. al. (2024) found new intermediate genomes so the multiple spillover theory is unlikely (it was anyway given lineage A and B are only two mutations apart as François Balloux, Virginie Courtier-Orgogozo and Jesse Bloom have observed). Single point of emergence is more likely with lineage A coming first (as Nick Longrich points out below). The market cases were all lineage B so not the primary cases. Their findings are consistent with Caraballo-Ortiz (2022), Bloom (2021).
Jia-Xin Lv, Xiang Liu, Yuan-Yuan Pei, Zhi-Gang Song, Xiao Chen, Shu-Jian Hu, Jia-Lei She, Yi Liu, Yan-Mei Chen, Yong-Zhen Zhang, Evolutionary trajectory of diverse SARS-CoV-2 variants at the beginning of COVID-19 outbreak, Virus Evolution, Volume 10, Issue 1, 2024
t.co/50kFV9zSb6
4. Jesse Bloom (2024) published a new analysis showing that genetic material from some animal CoVs is fairly abundant in samples collected during the wildlife-stall sampling of the Huanan Market on Jan-12-2020. However, SARS-CoV-2 is not one of these CoVs.
Jesse D Bloom, Importance of quantifying the number of viral reads in metagenomic sequencing of environmental samples from the Huanan Seafood Market, Virus Evolution, Volume 10, Issue 1, 2024,
t.co/rorquFs1wm
5. Michael Weissman (2024) shows a model with ascertainment collider stratification bias fits early Covid case location data much better than the model that all cases ultimately stemmed from the market. George Gao, Chinese CDC head at the time, acknowledged this to the BBC last year - they focused too much on and around the market and may have missed cases on the other side of the city).
Michael B Weissman, Proximity ascertainment bias in early COVID case locations, Journal of the Royal Statistical Society Series A: Statistics in Society, 2024;
https://academic.oup.com/jrsssa/advance-article-abstract/doi/10.1093/jrsssa/qnae021/7632556
6. Chen et al (2024) discuss how using the Grunow-Finke assessment tool lab origin looks at least as likely.
Chen, X., Kalyar, F., Chughtai, A. A., & MacIntyre, C. R. (2024). Use of a risk assessment tool to determine the origin of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Risk Analysis, 1–11. https://doi.org/10.1111/risa.14291
7. The anonymous expert who identified coding errors in Pekar et. al. leading to an erratum last year has found another significant error. Single spillover looks more likely. t.co/GAPihZu51P
8. The argument that an engineer wouldn't make the furin cleavage site with the features of SARS-CoV-2 overlooks it resembles that of MERS in several structural and functional ways, and the sequence looks quite similar. In 2019 WIV researchers were involved in MERS research. Dr Andreas Martin Lisewski discusses similarities with a MERS infectious clone described in 2017 here. t.co/fAVUlJu0TK
9. Broad Institute biologist Alina Chan also observes the S1/S2 FCS PRRA insertion in SARS-CoV-2 generates a Class IIS restriction enzyme site (BsaXI). This was used by WIV and Ralph Baric at UNC previously. The full DEFUSE proposal available since the debate strengthens the argument of Bruttel et al. Specifically, the use of BsmBI, 6 fragments, and leaving the sites in). https://usrtk.org/covid-19-origins/scientists-proposed-making-viruses-with-unique-features-of-sars-cov-2-in-wuhan/
What you say about raccoon dogs here is mistaken. Raccoon dogs are not a plausible intermediate host for sars-cov-2 on the basis of information that has been known since 2021. There are several considerations.
1. Xiao et al (2021) - https://www.nature.com/articles/s41598-021-91470-2%E2%80%8B%E2%80%8B%E2%80%8B , which includes a co-author of Worobey et al (2022), a leading zoonosis paper states in table 1 that the raccoon dogs were wild caught in Hubei, not farmed as you assert in the piece. This alone rules out raccoon dogs as plausible hosts for two independently sufficient reasons. Firstly, there is unanimity in the literature that the bat ancestral virus to SARS-CoV-2 is in southern Yunnan or South East Asia. Everyone agrees with this, including Shi Zhengli. If a species was wild caught in Hubei, then there would be no explanation of how it acquired the ancestral bat virus, given that Hubei is 1000 miles from southern Yunnan.
Secondly, a mystery of sars-cov-2 is how it acquired the furin cleavage site that makes it so transmissible. There are 850 known sars-like coronaviruses, and only one with a furin cleavage site. According to private messages exchanged by proponents of zoonosis, the furin cleavage site could not have been acquired in the market because the density of animals was too low (only 3-4 per cage). When avian influenza acquires a furin cleavage site that occurs on farms with thousands of chickens densely packed, i.e. not in the wild and not when there are a handful of animals in cages in a market. https://usrtk.org/covid-19-origins/visual-timeline-proximal-origin/
2. Wang et al (2022) https://academic.oup.com/ve/article/8/1/veac046/6601809 also confirms that the raccoon dogs were wild caught in Hubei. What's more, Wang et al (2022) tested 15 wild raccoon dogs of suppliers of Wuhan markets, including the Huanan market, in January 2020 and found them to be negative for SARS-CoV-2. On average, 38 raccoon dogs were sold across the four markets in Wuhan from 2017 to 2019. So, the 15 raccoon dogs likely comprised nearly the whole inventory of raccoon dogs that would have been supplied to the Huanan market at the time.
In short, the raccoon dogs supplied to the market were tested and were negative. It is very strange that the raccoon dog narrative has persisted for this long given this public information. Even the strongest proponents of the raccoon dog hypothesis have walked back their bold claims that raccoon dogs are thost.
Xiao et al (2021) has a list of species sold at the Huanan market. I would encourage you to read that list and suggest which animals you think are plausible, and I will tell you why they are not actually plausible.
I don't think there's a consensus that the ancestral virus is from Yunnan or SEA. No plausible immediate ancestor has been found anywhere. For what it's worth, high seroprevalence for SARS-related viruses have previously been found in Hubei:
https://www.science.org/doi/10.1126/science.1118391
Second, the line between wild-caught and farmed animals in the wildlife trade is blurry. For instance, there was a wildlife farm in Hubei where SARS was detected in palm civets. The farm contained a mix of wild-caught and farm-breed civets:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1153763/
I am still not convinced that racoon dogs were the intermediate host, since SARS-CoV-2 evidently has a very broad host range.
While it is true that Sarbecoviruses have a wide range, direct bat ancestors to SARS-CoV-2 have a much narrower range considering bat and viral dispersion patterns. Still under investigations though, and politically sensitive topic.
My point was specifically about SARS-COV-2, which can infect mammals from several different orders. And we don't know whether the restricted bat lineages are descendent from an ancestral generalist lineage, or if SARS-CoV-2 is descendent from a more restricted lineage.
Si et al has a preprint out (zhengli shi is a co-author as well) adressing some of it by analyzing what makes RBD have broad or narrow ACE2 tropism. Seems broad tropism is ancestral
There are SARS-like coronaviruses in bats in Hubei, but they are too distant from sars-cov-2 to be the ancestor virus.
Shi Zhengli of the Wuhan Institute of Virology stated in July 2021:
“We have done bat virus surveillance in Hubei Province for many years, but have not found that bats in Wuhan or even the wider Hubei Province carry any coronaviruses that are closely related to SARS-CoV-2. I don't think the spillover from bats to humans occurred in Wuhan or in Hubei Province”. (Cohen interview)
This is confirmed by wider bat sampling efforts by the Wuhan Institute of Virology and EcoHealth Alliance prior to the pandemic. (Fan et al 2019 - WIV team; Hou et al 2010, Latinne)
The WHO Joint Study report noted:
“Tests on samples of more than 1000 bats from Hubei Province showed that none was positive for viruses related to SARS-CoV-2” (p106)
The samples tested had 74% to 90% sequence identity to SARS-CoV-2,(annex F table 4) compared to 96.1% for RatG13 (from Yunnan), and 96.8% for BANAL-52 (from Laos).(Temmam)
From January 7th to 18th 2020, Wang et al (2022) collected samples from 334 bats around Wuhan. All were negative for SARS-CoV-2 or its progenitors.
Hassanin et al (2024) reported the “the results of several field missions carried out in 2017, 2021 and 2022 across Vietnam during which 1,218 horseshoe bats were sampled”.(Hassanin) They conclude that “For SARS-CoV-2 however, phylogeographic indicators provided very high support for an origin in the zone covering northern Laos, southern Yunnan and north-western Vietnam”.
Pekar et al (2023) argue that the ancestral host for SARS-CoV-2 is ‘very unlikely’ to be from Hubei. *Pekar et al are the main proponents of the zoonosis hypothesis*
“Considering the large distance separating the closest-inferred bat virus ancestors from where SARS-CoV-1 and SARS-CoV-2 each emerged (Fig. 3i,j) and the high dispersal velocities necessary to traverse the aforementioned distance (Extended Data Fig. 18), it is very unlikely that the lineages descending from the closest-inferred bat virus ancestors of SARS-CoV-1 and SARS-CoV-2) (which reached, respectively, Guangdong Province and Hubei Province solely via dispersal of these viruses through their bat reservoirs.”
Another point - If the animals came from a farm, why were there approx 1-2 spillovers only in Wuhan city sufficient to cause an ongoing outbreak? If the farm had hundreds of raccoon dogs, why were they not numerous spillovers at other locations the farm was supplying? This is much easier to explain on the lab leak hypothesis.
What is the standard structure of exotic meat suppliers to wet markets? Are they usually big operations that supply many markets on a weekly basis, or are they usually small mom-and-pop operations that turn over their entire inventory in a single transaction to a single market? I'm sure I'm presupposing a million false things in the way I've phrased this question, but I think some of this information would be relevant to knowing how much and whether your question is a significant one to ask.
Hello, farm size varies a lot - for SARS-1 civet farms had around 100 civets per farm, but some bamboo rat farms have tens of thousands of animals. So, a lot depends on what the intermediate host was and how big the farm was.
The go to source on the suppliers to the Huanan market in particular is this washington post article by michael standaert https://www.washingtonpost.com/world/asia_pacific/china-covid-bats-caves-hubei/2021/10/10/082eb8b6-1c32-11ec-bea8-308ea134594f_story.html
which is cited by Worobey 2022. They say that the suppliers of the Huanan market were farms in Hubei.
They reported that reported that authorities closed 290 farms in western Hubei that accounted for between 450,000 and 780,000 animals. This suggests that there were 1,551 - 2,689 animals per farm. The number of spillovers posited by zoonosis proponents is around 8. None of these spillovers have actually been ascertained, they are the product of a (bad) model, and none of the official confirmed cases in the wet market worked at wild animal stalls.
Is there any reason to expect that the farms that were closed were relatively equal in size, rather than some being tens of thousands of animals and some being only a few dozen? Is there any reason to expect that the farms that were closed had a distribution of sizes that is representative of the set of all potentially relevant farms?
There is a lack of information on this front. The farmed live animals sold at Huanan were mainly from Enshi and "A clampdown on Enshi's wildlife trade at wet markets began on Dec. 23, 2019, according to state media, eight days before China publicly acknowledged the new virus. The head start in Enshi doesn't mean officials found something amiss: It could have been preventive, as rumors emerged of market vendors falling mysteriously ill in Wuhan. But it means evidence regarding Enshi's wildlife trade was erased before the world was aware of the existence of a novel coronavirus."
It would be very nice to know what the specific farms supplying the market were, but we don't have much information. But certainly one has to posit that the farms could not have been very large, otherwise there would have been numerous zoonotic spillovers at other markets, which were not observed
So, do you think the Wang et al (2022) and Xiao et al (2021) papers are fraudulent or mistaken? Have you thought about calling for a retraction. Wang et al literally sought out the suppliers of raccoon dogs to the Wuhan markets and tested the wild caught raccoon dogs. All negative.
They only tested for viruses, not antibodies. If you tested 100 random humans today, your chance of finding any specific endemic virus (e.g SARS-COV-2) would be quite low.
They also only state that some raccoon dogs were wild-caught, not that all were.
As noted above, they tested the entire inventory that would have been supplied to Wuhan and said they were all wild caught - they were not farmed. Xiao et al (2021) also says the raccoon dogs for sale at Huanan were wild caught. Xiao et al (2021) includes a co-author of Worobey et al (2022), the main zoonosis paper.
They tested their tissue and blood samples by PCR. Zoonosis proponents posit that the approx 20 raccoon dogs for sale at the Huanan market in the month of November 2019 caused approx 8 zoonotic spillovers, of which 2 led to an ongoing outbreak (per Pekar et al (2022). Despite testing of the suppliers to that market in January, they were all negative.
There are, to recap, three independently sufficient reasons to rule out raccoon dogs (1) wild caught, so no explanation of FCS acquisition; (2) wild caught in Hubei, so ~impossible that they were exposed to the ancestral bat virus; (3) the animals at the market suppliers were tested and tested negative.
You seem very focused on a very narrow scenario for zoonosis, that is very similar to the consensus position for the SARS-CoV-1 outbreak. If you accept any scenario for one hypothesis, but any scenario for another, of course you will end up favoring the second hypothesis. If you could prove definitively that it couldn't be raccoon dogs, that's evidence against a zoonotic origin. But it's far from devastating. There were plenty of other susceptible animals at the market. Additionally, it was a dense area with an unusual amount of traffic from people with close contact to wildlife. I think you really need evidence of engineering in order to believe in a lab leak, and I with a molecular biology background, find the evidence on that profoundly unconvincing.
I would also argue that since it's CCP policy that the outbreak didn't happen in Wuhan, we should be skeptical of statements from Chinese researchers working in China.
Xiao does not seem to declare that all raccoon dogs sold at the market were wild-caught.
>Approximately 30% of individuals from 6 mammal species inspected (labelled W in Table 1) had suffered wounds from gunshots or traps, implying illegal wild harvesting (Table 1).
>Raccoon dog fur farming is legal in China; however, due to a drop in fur prices, raccoon dogs are now frequently sold off in live animal markets, augmented by wild-caught individuals."
If raccoon dogs in any specific population were subject to any particular viral epidemic that burned itself out due to herd immunity, it's possible that no virus would be sampled two months later.
Xiao et al Table 1 marks them as wild caught on the basis of injuries and vendor interviews.
As noted above, on average, 38 raccoon dogs per month were sold across the four markets in Wuhan from 2017 to 2019. So maybe like 20 were sold at the Huanan market in November 2019, which is less than one per day. Wang et al 2022 states
"We immediately started a surveillance investigation on the
origin of SARS-CoV-2 during 7–18 January 2020. Consequently,
the lung, liver, and intestinal tissue samples were collected from
mammals, which were captured in the rural area (Changxuanling and Yaoji towns) of Wuhan by three local traders for
vendors at animal markets including Huanan Seafood market
during 7–18 January. "
I.e. these are locally captured wild animals. The 15 raccoon dogs likely comprised nearly the whole inventory of raccoon dogs that would have been supplied to the Huanan market at the time.
"If a species was wild caught in Hubei, then there would be no explanation of how it acquired the ancestral bat virus, given that Hubei is 1000 miles from southern Yunnan." How often do viruses travel a thousand miles? I was under the impression it's common.
"Even the strongest proponents of the raccoon dog hypothesis have walked back their bold claims that raccoon dogs are thost." Who are these strongest proponents?
It is true that viruses often travel more than 1000 miles, the point is that they do not do this via bats which do not migrate that far. e.g. it seems that SARS traveled ~1000 miles from Yunnan to Guangdong via the civet trade. But with SARS-CoV-2, for animals wild caught in Hubei, there is no explanation of how they would have been exposed to the ancestral bat virus.
eg Angie Rasmussen is in many ways the face of the zoonosis hypothesis and a co-author on many zoonosis favoring papers. She said to the Atlantic “This is a really strong indication that animals at the market were infected. There’s really no other explanation that makes any sense.”
Then she said this after the Bloom critique "As a co-author of the “raccoon dogs saga” report, it’s disappointing to see a journalist outright lie about our work. We never claimed to find infected animals."
https://twitter.com/angie_rasmussen/status/1651976430773731336
> She said to the Atlantic “This is a really strong indication that animals at the market were infected. There’s really no other explanation that makes any sense.”
> Then she said this after the Bloom critique "As a co-author of the “raccoon dogs saga” report, it’s disappointing to see a journalist outright lie about our work. We never claimed to find infected animals."
I'm not sure what this is supposed to prove. There's no contradiction.
If I go into a bathroom and smell shit, there's a really strong indication that someone took a dump in there. There's really no other explanation that makes any sense. But that doesn't mean I found a turd in the toilet.
I was responding to this: "Even the strongest proponents of the raccoon dog hypothesis have walked back their bold claims that raccoon dogs are thost." Who are these strongest proponents?"
She said "there is no explanation that makes sense apart from an infected raccoon dog". She then said that this doesn't imply that a raccoon dog was infected.
No. She's saying that all signs point to an infected raccoon dog, and that they have not found an infected raccoon dog.
A detective finds a body. All signs point to a gunshot wound. He has not found a bullet or a gun. It doesn't mean he's changed his mind, or that it's now necessarily more likely that the guy lying on the ground with a big hole in his head died of strangulation.
We might have to agree to disagree here.
Except that animals were not all wild caught in Hubei. This may be part of what tripped up Saar during the debate -- lab leak supporters have spent the past four years repeating things that aren't necessarily true and creating an epistemic bubble for themselves.
In reality there are many ways a virus from southern China could have made its way to a market selling wildlife in central China. For instance the markets in Wuhan seem to have sourced raccoon dogs and civets from farms in a mountainous region called Enshi about 630km to the west of Wuhan, nearly halfway to Yunnan province. Enshi has some of the largest cave systems in the world and there's no testing of bats or farmed animals from there that's been released so far. This is the region, western Hubei, where animals were found infected with SARS-CoV-1 back in the early 2000s. https://www.independent.co.uk/asia/china/covid-coronavirus-bats-caves-hubei-b1940443.html
We also know that the Huanan Market also imported at least some animals such as bamboo rats directly from southern Yunnan province, as well as Guangxi. And these are only the sources we know about...it's unclear exactly where the raccoon dogs sold at the time of the outbreak came from.
Meanwhile the evidence at the market does seem to point to raccoon dogs. The store where they were sold appears to be the most likely source of the outbreak within the market. Positive samples were found on a cage used for holding them, as well as a tool used for stripping fur off the animals. That doesn't prove they were the species that spread the virus to humans but there's definitely not a good basis to rule it out.
As noted above, there is consensus in the literature that the ancestral bat virus is not in Hubei. So, you are going against all of the scientific literature here, including Pekar et al (2023), who are leading proponents of zoonosis.
The bamboo rats you are referring to I believe were not live animals but frozen or refrigerated animal products. See the Gao teams' report - https://www.nature.com/articles/s41586-023-06043-2
Worobey et al cite the washington post article which says that:
“A person with knowledge of the Wuhan market supply chains, who spoke on the condition of anonymity to protect his contacts, told The Post that live animals sold at markets in Wuhan were sourced from Hubei, particularly Enshi and Xianning prefectures, as well as from Hunan and Jiangxi provinces.”
Further confirmation for this is provided by the fact that:
“A clampdown on Enshi's wildlife trade at wet markets began on Dec. 23, 2019, according to state media, eight days before China publicly acknowledged the new virus. The head start in Enshi doesn't mean officials found something amiss: It could have been preventive, as rumors emerged of market vendors falling mysteriously ill in Wuhan. But it means evidence regarding Enshi's wildlife trade was erased before the world was aware of the existence of a novel coronavirus.”
I don't understand why you think the raccoon dogs were not wild caught locally. Xiao et al Table 1 says this explicitly, as does Wang et al 2022
Yes but did you read that Washington Post article? It's actually the same as this article I posted in the comment above: https://www.independent.co.uk/asia/china/covid-coronavirus-bats-caves-hubei-b1940443.html
It seems to be saying that many animals were from farms in Enshi where civets and raccoon dogs were raised. That's a significant distance from Wuhan and yet you're pointing to testing of the 15 raccoon dogs from within the Wuhan city limits itself.
As far as the ancestral virus not being from bats in Hubei that may be true but like I mentioned SARS-CoV-1 also somehow infected animals in western Hubei. So there are certainly some unexplained details but I don't think any of this makes a spillover to humans from raccoon dogs or civets impossible at all. Remember the zoonotic origin doesn't need to be something as simple as it spilling over from bats to one species of animal that then infects humans at the market. It could have spread between multiple species just as COVID did after infecting humans.
Also the reported shutdown of these markets in western Hubei a week before the Huanan Market was closed is interesting and that alone would seem to call the lab leak explanation into question.
The 15 raccoon dogs were traced from suppliers of raccoon dogs in the Huanan market. It's not that they randomly sampled raccoon dogs in Wuhan. They did this over 11 days. Around 38 were sold on average per month. so, the raccoon dogs they tested likely accounted for the entire raccoon dog inventory in Wuhan at the time.
The wapo article is about the farms supplying the market, but Xiao et al says that not all of the animals sold at the market were farmed. Worobey et al also says this!
fwiw I think the infected animals for SARS in western Hubei is disputed though haven't looked into it.
Shutting down of the farms is not evidence against lab leak. Some parts of the government clearly thought it was a zoonosis and acted accordingly in Dec and Jan, which is eg why they disinfected the market in Jan.
1) It's not clear if that report was indicating that ALL animals were wild-caught or just some. There was an interesting incident back in 2020 where a Chinese environmental NGO wrote to the Wuhan Forestry Bureau (responsible for regulating the wildlife trade in Wuhan) asking about animal sales at a single store at the Huanan Market: https://cbcgdf.wordpress.com/2020/05/31/breaking-news-cbcgdf-policy-and-research-department-received-a-response-about-the-dazhong-livestock-game-stall-at-the-wuhan-south-china-huanan-seafood-market/
The Forestry Bureau responded with a list of sources for the store -- which included raccoon dogs sourced from a farm where mink and foxes were also bred. So in at least some cases, they were sourced from farms.
2) This study doesn't say that all raccoon dogs were wild-caught and testing 15 animals from one location is a ridiculous number to draw a conclusion from.
Overall the wildlife trade in China is difficult to track with animals being raised on farms but also poached from the wild, often shipped long distances, with some such as pangolins trafficked illegally. We never got a full accounting of animal sales at the Huanan Market -- in spite of the post above by the Chinese NGO, this information was left out of the eventual WHO report. There was no mention of the sources for raccoon dogs and civets, two of the most-suspected species, and China denied there was any evidence live mammals had been sold at the market at all anytime in 2019.
The article you link to refers to an advertising sheet from 2015. Xiao et al (2021) is a study from 2017-2019 involving observation of the animals and interviews with sellers.
15 is not a ridiculous number to draw a conclusion from. If the monthly sales were 38 per month across four markets in Wuhan, then the January testing by Wang et al (2022) of the supplier is half of the inventory that would be supplied to that specific market in a month.
Worobey et al said of the raccoon dogs photographed in Dec 2019 "“appear to be local, wild-caught common raccoon dogs rather than farmed raccoon dogs and that their plush coats are consistent with those observed in the winter”. edit - reference is the worobey preprint - https://zenodo.org/records/6299600 fig 3
No the article I linked involves an environmental NGO asking the Wuhan Forestry Bureau about that advertising sheet, which lists nearly a hundred species for sale, far exceeding the total number of species you're allowed to sell in China.
The Wuhan Forestry Bureau responds by giving a list of sources for that store at the market, and it includes raccoon dogs from a specific farm. This is the ministry that was responsible for regulating the animal trade in Wuhan.
Meanwhile the Xiao study you mentioned simply indicates that some of the animals observed appear to be wild-caught. So this is a good example of why it's important not to draw such a strong conclusion when there are all these different facts swirling around. Let's at least wait until China gives a source for the raccoon dogs that were sold on the western side of the market -- they've never provided this information amazingly.
Ah I see, that's interesting I hadn't seen that.
Still, the referenced farm is in the east of China, which is even further from the relevant bats than Hubei. In general, fur farms in China are overwhelmingly in the east of China.
Yeah I agree but this particular store, located on the east side of the market, was already less likely to be the origin. The question is why this information, sources for the most suspected species, was left out of the WHO report. There are no sources listed for raccoon dogs or civets at all, in spite of them clearly having been sold there and in spite of China already having both sampling data and seemingly also sales records showing they were.
WHO investigators later said CCP officials pushed them to exclude any mention of these animals from the 2021 WHO report and denied there was any evidence of any live mammal sales at the market during 2019. So China seems to have censored almost all of the information (beyond the 15 locally tested animals you mentioned) about the most likely intermediary species.
Roko (the one infamous for the basilisk) has argued that what might have appeared as a shift in expert consensus is in fact "manufactured consensus." This is worth considering. Link: https://www.lesswrong.com/posts/bMxhrrkJdEormCcLt/brute-force-manufactured-consensus-is-hiding-the-crime-of | I also believe an algorithm somehow sorting, weighting evidence cleverly could solve the problem of "too much evidence." The CIA, in fact, at least has used one such algorithm, invented by Richard Heuer.
I think the fact that the expert class went out of their way to suppress the lab leak hypothesis, and that lab leaks of this nature have occurred in the past, is probably damning enough. My impression is that Roko hasn't really introduced any new arguments that Peter didn't cover in the debate, and Peter isn't being funded by the CDC, to my knowledge.
Everybody plays "Cover Your Ass" whether they're innocent or guilty. So that's not a viable argument. Nobody likes to be accused of having made a mistake, whether they have or not.
Is this supposed to excuse any possible degree of pre-emptive collusion and fraud to cover up for institutional incompetence?
No, it's supposed to say that suppressing a hypothesis is not damning evidence of the hypothesis being correct. The hypothesis might be correct or incorrect, suppressing it is bad anyway.
I agree. That's why I'm saying it's a 'damning' indictment of our reigning technocrats.
I regret to inform you that technocracy is unusually *good* on this point. If you can reliably find a strain of human that doesn't reflexively hide their misdeeds, we sure could use them!
Everyone? Some naive people talk to cops believing they will get their side of the story out there
Destroying evidence, like oh idk dna sequences the wuhan insitute were working on at the time; suggests a estimation that they expected they couldve been at fault, I would expect innocent people to hide the evidence and do a very private ananlysys and ask a lawyer if it helps them.
>So [playing "Cover Your Ass"] is not a viable argument.
It *is* a viable argument in the sense that it dampens and modulates the strength of other arguments.
Arguments about what the WIV can or cannot do, arguments about the number of cases centered around the wet market, and arguments about which DNA sequences were found and when they were found, all need to have their Bayesian impacts divided by a selection-bias-from-a-Maoist-Totalitarian-Government factor.
OK, but you've got to include EVERYONE, not just the people you don't like. It wasn't *just* the various segments of the government of China, it's also the investigators, people who have a public positions ahead of time (and "public" can be a argument with another expert).
Just about everyone who made a public statement on any factor related to COVID during the early days had at least one mistake that they wanted to "de-emphasize". There were "public experts" in the US claiming that we could keep COVID out after it had already been circulating for months. And the border closings were a farce. The quarantines were not really enforced, and lots of people were allowed to essentially skip them. (Not that it really mattered since COVID was already circulating.)
This would be the perfect placement for the Peter Parker-Harry Osborn meme.
I already distrust the FDA, the CDC, the WHO, and the CCP. You don't need to sell that distrust to me.
The expert class went out of their way to suppress the lab leak hypothesis in part because Trumpist Republicans were promoting the lab leak hypothesis and the expert class is reflexively opposed to anything associated with Donald Trump. And in part because discussion of the lab leak hypothesis, regardless of its factual basis, would imperil the funding and prestige of themselves and their colleagues.
Neither of these depend on the lab leak hypothesis being true or probable, so the fact
of the suppression carries little information regarding the truth of that hypothesis.
More generally, innocent people reflexively engage in coverups often enough that "Coverup, thus Guilty!" is a weak argument at best. The bit where innocent people welcome thorough investigation of every accusation made against them because they are confident it will prove their innocence, is sadly disconnected from reality.
Agree on the second part but scientists also dismissed the lab leak hypothesis because they genuinely found it to be less likely than zoonotic origin. The private Slack messages everyone likes to cite where scientists initially discussed the possibility it came from a lab end with them saying that. https://jabberwocking.com/i-read-the-entire-slack-archive-about-the-origin-of-sars-cov-2-there-is-no-evidence-of-improper-behavior/
Also important to realize that the idea of lab origin was pushed by a number of different groups, not just Trump Republicans. The first version of it was actually from the CCP themselves, who tried to claim SARS2 was a bioweapon the USA attacked them with at the Wuhan Military Games. From there it was picked up by anti-CCP groups in the Chinese diaspora, which had a surprising degree of influence on certain Trump officials, and the storyline changed to focus on the Wuhan Institute of Virology. Other governments such as Iran and Russia came up with their own versions blaming the USA.
So there's really an ocean of BS on this topic and you can't blame scientists for saying it's less likely when it actually does appear to be less likely. None of this resulted in the idea itself being suppressed. It's probably gotten something like a hundred times or more coverage in the media as the animal trade origin. Governments and other groups are incentivized to promote some version of this storyline for geopolitical reasons -- in China's case to evade responsibility for the pandemic, in the USA's case to more directly blame China. Meanwhile the real origin gets lost in the confusing back and forth of competing claims and accusations.
"It's probably gotten something like a hundred times or more coverage in the media as the animal trade origin"
Really? You want to dig into the archives of NYT and show me a hundred-to-one ratio of articles propounding the lab leak theory vs. zoonosis?
I also think 'Biorealism' made a fair point at the end of the comments section-
"There is at least one argument they used knowing it was misleading. They used Ron Fouchier's argument (without acknowledgement) that WIV would have used a well known reverse genetics system despite Andersen saying that wasn't the case on at least two occasions. He noted they had been created "on a whim" and on 20 Feb if anyone thought these were hard to create from scratch a group of researchers had just created one in a week. They should have acknowledged this.
The Nature reviewer clearly took too hard a line as well. Their initial manuscript was far more balanced. Moving from lab origin was not "necessary" to "not plausible" was not justified. As David Relman emailed Francis Collins there need to be an impartial assessment which acknowledged the lack of evidence either way. A Dept of Defense Working Paper dated 26 2020 noted their conclusions were based on "not on scientific analysis but, on unwarranted assumptions"."
Just doing a quick search on Google, even NYT, which has been criticized by lab leak supporters for their supposed biased coverage, has about twice as many stories that mention "Wuhan Institute of Virology" vs "Huanan Market." And overall in the media the number of articles on the animal trade origin -- i.e. looking into the animal trade in China in relation to the outbreak -- is almost nothing vs. those covering the lab leak controversy.
This is strange because even if you think it leaked from a lab there are specific people in China, vendors and workers at the market, who could tell us much more about what happened, and yet the focus has shifted almost entirely to American scientists and politicians. You then have to think about the fact that many experts believe the lab leak never happened at all!
As far as the Proximal Origin paper I think it's fine to say the conclusion should have been less strong, maybe they should have said "lab origin is less likely" rather than "not plausible." But what's funny is you can also find a quote from Richard Ebright from the same time period (but before Proximal Origin was published) where he says "Based on the virus genome and properties there is no indication whatsoever that it was an engineered virus."
So clearly this was not an unreasonable conclusion to come to, unless you are saying Ebright has no idea what he's talking about, and where it falls apart IMO is this idea that in writing this paper scientists were somehow suppressing the lab leak theory. Just weeks after the paper was published the president of the USA was repeating stories about how someone at the lab had started the pandemic. The State Department launched an investigation that resulted in the stories about "sick WIV scientists with COVID symptoms." Steve Bannon ran a PR campaign to promote Li-Meng Yan's bioweapon allegations. Congress issued multiple reports and held hearings about it. We're now going to see yet another investigation by Rand Paul, who doesn't exactly seem neutral. You can't say this idea was really suppressed that much...instead there's a narrative people repeat over and over about how it was suppressed.
"Just doing a quick search on Google, even NYT, which has been criticized by lab leak supporters for their supposed biased coverage, has about twice as many stories that mention "Wuhan Institute of Virology" vs "Huanan Market.""
If you use those exact terms, yes, you get 450 results vs. 170 in favour of the Wuhan lab being mentioned. If you search for "wet market" vs. "lab leak", you get ~40,000 results for the former vs. ~16,000 for the latter, so the balance of coverage flips. If you search for the same terms on google outside of the NYT website, you get about 3 times more mentions of the Wuhan lab or 'lab leak' than you do for the wet market, but on the other hand, "chinese animal trade" returns 89 million results as opposed to 274,000 for "chinese biolab". So I'm not sure the zoonotic origin story is being downplayed, precisely.
My argument is not "cover-up, therefore lab leak is true". My argument is "cover-up, therefore untrustworthy"- that the effort to suppress the lab leak hypothesis was a crime in itself, and the expert class are still 'guilty' in that sense.
My understanding was that the Proximal Origins paper was already published by the time Trump made a public statement in early April, but I don't have a ref handy.
Yeah that's correct, although it had been promoted earlier by other GOP politicians such as Tom Cotton, and outside the government by former Trump officials like Steve Bannon. Actually I don't think Trump himself was ever too invested in this idea. There's a hilarious interview he did with Sherri Markson where it quickly becomes clear he doesn't have much more knowledge than a random person. It was more a project of Mike Pompeo, the Secretary of State and former CIA director, and a group of people working under him
Proximal Origins was in review by the time of Cotton's editorial.
I think the conclusion might not have been fully exaggerated yet though.
Ultimately the Proximal Origins controversy becomes much less interesting IMO if there was no actual inside knowledge about what happened, and even Emily Kopp of US Right to Know told me she doesn't think that's the case, although she does think scientists engaged in a "coverup" of some kind. Nothing ever prevented anyone from writing a different paper saying lab origin is plausible and many did.
Hence my specifying "Trumpist Republicans" rather than "Donald Trump". As everythingism notes, Trump himself was not really the problem, but there were plenty of people in Trump's orbit advancing the Lab Leak theory.
And advancing it well beyond what the evidence would have supported at that early date. It was absolutely reasonable for the "expert class" to push back against that, and at least understandable that they would overcorrect in the other direction.
This is one of the situations where I don't feel any more confident trusting the expert consensus than looking at the papers and/or debates and trusting my own reasoning to evaluate the arguments.
A good example I personally like to use of this definitely being the case would be from COVID as well: during early 2020, the Finnish Institute and Wealth and Welfare (THL) officially claimed IFR of [alpha-variant, in immunonaive population, before best practices for treatment were discovered, as average within a population with a lot of elderly people] COVID to be <<.01%, purportedly repesenting the best available scientific evidence, while at the same time some .5% of entire population of Bergamo in Italy had died (and similar IFR could have been calculated from Diamond Princess data, etc). Here, rational evidence easily trumps claims made based on scientific evidence: a child could see if explained to that no matter how many unidentified cases there might be, there's no way for IFR to be lower than total excess mortality within the population. And additionally, I have a clear model where the people working in the institute with their virology doctorate doctorate degrees went wrong: they rely (indeed, might be contractually obligated to, I haven't checked) on scientific studies, but as none had been published at such an early stage, they pattern-match to the most closely analogous study, which might be about flu or something. Most people, even experts, just aren't epistemologically nuanced enough to use rational evidence, as Scott has often written about in "The Phrase "No Evidence" Is A Red Flag For Bad Science Communication".
When expert scientists are discussing statements within established science, I would trust their claims to not misrepresent the science, and since the scientific method approximates Bayes' theorem pretty well, I have a high degree of trust for the claim. This isn't such a case! Science works, but it advances one funeral at a time more or less, and there's no way this debate has gotten to a point of established science, indeed the polls lean to one direction but also specifically show the question ISN'T settled.
Sure there can be. Not all deaths are caused by COVID, and both Diamond Princess and the population of Bergamo were heavily age-skewed. Not a representative population, that's why serosurveys were useful to determine the true number. You say the Finnish government were also making claims about the elderly but did you check that the numbers line up there?
There are other issues with early IFR estimates, for example in Italy there was mass abandonment of care home residents by the mostly foreign carers who were scared and often wanted to return home in anticipation of border closures. Deaths reported as COVID were actually in some cases caused by abandonment e.g. dehydration. Early on doctors were also way too ready to put people on ventilators, partly because they thought that would keep the virus out of the hospitals and themselves (it didn't). But ventilators are a very destructive last resort option. Once doctors stopped doing that so much the death rates went down.
The point made in the link is classic the-social-aspect-of-opinion-formation. The world of domain experts can be analysed as a community (more-or-less closely integrated), complete with opinion leaders, followers, rebels, go-betweens, outcasts and every other (social) role related to the goings-on in any community.
The negative spin on these (social) processes in the linked post is unneccessary, though. Partly because such social processes are inevitable/unavoidable, partly because they may also be conductive in getting the joint reseach process from A to B. (The importance of all social factors on opinion formation is Janus-faced.)
(...and ACX/self-declared rationalists is also to a certain extent a community, of course.)
I have no horse in this particular race, but I do have a lot of expertise in some of the areas rootclaims "investigates" (especially the stuff related to Syria and chemical weapons) - where their analysis is so shoddy and laughable it's indistinguishable from Youtube conspiracies - and the biggest surprise to me here is that anybody really bothers with rootclaim in the first place? The more you learn...
Even if you do not agree with the values of the (prior) probabilities, you can choose your own, and replicate the process rootclaim uses. It is far from "shoddy" and research on Bayesian methods modelling these kinds of events is highly valuable.
I have nothing against Bayesian reasoning. Use it all the time. But the priors, the "evidence", the weights attached were so out-there, I just assumed arguing with them was like arguing with Alex Jones who has taken a stats course. Unfortunately statistical methods don't cure brain worms (and in fact might even provide them with a rationalist veneer).
Can we then Aumann agree that: the Rootclaim method with better priors and handling of evidence would not necessarily be shoddy, the current analysis is or might be, and the method itself remains valuable enough that it would be good if they continued to work on it (in Scott's framework, (2)).
With good-enough priors, yes it would be good enough (by definition). But how good would those have to be? Is it possible to have priors good enough that the method isn't just laundering assumptions through some fancy statistics, as meta-analyses or financial projections or suchwhat often are? I'd answer these questions as "at least 10x better than we have" and "not really, no."
Strange place to comment, I was aiming for agreement with Tobias.
How so?
You happened to laugh at an analysis where we turned out to be spot on. There is no now clear verified video evidence of the opposition carrying out the Ghouta sarin attack.
https://blog.rootclaim.com/new-evidence-2013-sarin-attack-in-ghouta-syria/
If you are going to claim your analysis is spot on, please link to a credible independent source. Otherwise this comes across as we believe this unlikely thing and used our analysis you find shoddy to conclude we were right so you should not consider our analysis shoddy.
Follow the link. It describes external forensic work which you can verify yourself.
I did follow the link. My point is if you were vindicated I would expect to be able to read about it in an independent source. Otherwise I need to look at all of your sources and see both which are in support of your claim and which you are claiming support you and look at the evidence in detail. Lab leak as a plausible theory has been printed in the NYTimes since at least 2021, has any mainstream newspaper published anything about this?
How often do mainstream newspapers publish anything about Ukrainian misdeeds in the war? Probably something like 1/1000 ratio compared to the Russian ones, whereas a reasonable prior of their prevalence would be like 1/10 at best. Similar sort of bias is probably at play with Syria.
I'm not discussing the ratio of coverage but here is an article from le monde that is easily searchable that backs up the claim that Ukranian's committed war crimes. https://www.lemonde.fr/en/international/article/2022/04/09/ukraine-s-military-accused-of-war-crimes-against-russian-troop_5980121_4.html. If there was anything that resembled reasonable evidence, it would get some coverage the trouble is Saar has motivated reasoning that is not convincing without external collaboration. Otherwise the best you can claim is that there is some evidence that the syrian weapons were a false flag and again it's not your (Saar's) method that proves the analysis is spot on, its how strong you trust the conclusions from the grayzone/wikileaks that the appearance of impopriety means the conclusions are false.
Sorry, I thought it was obvious. These outlets have no incentive to publish such findings. Sadly these are the kinds of things you need to verify yourself.
Agreed.
How do you account for “one of the most powerful political actors in the world which has a history of controlling information flows runs the area in question and can be manipulating what evidence is available?”
I get that you have to work with the evidence you have. When the evidence you have is passing through a filter that has pretty strong incentive to hide its own wrongdoing (if that were to exist) how does that fit into your analysis?
Did they assess a probability that, ie, certain research records where the researchers tried that ~exact~ furin site were destroyed by the CCP? Wouldn’t that just be a “naked prior”?
The CCP were trying to cover up zoonotic origin as well at first, with arresting Dr Li and cleaning up the site. Because it was seen as politically embarrassing, as China said it had fixed wet market issues after SARS. So that applies on both sides.
(The Chinese official position now is to heavily imply it came from abroad)
Also a question of whether, in the lab leak hypothesis, they'd have been able to cover up large amounts of evidence effectively, given how inept the initial covid response was. And officials at different levels acting at cross purposes to cover their asses
This is not true. George Gao said in January 2020 that it was from the wet market. After leading sampling there and the case search, he changed his mind in summer 2020
If you look at what usually happens with research-related accidents, they are usually shrouded by obscurity for years. This is likely even more true in China. Compare e.g. the 1977 Russian flu, sverdlovsk anthrax leak, various SARS leaks. In some cases finding out that an accident occurred can take decades.
I think you're expecting a government to act in a unified way. I'm sure if you searched you could find statements supporting every side. And I *have* encountered (reports of) official Chinese statements claiming it came from abroad. So that is one of the positions that they are (or were) taking. I've also encountered denials that COVID exists (e.g. by the mayor of Wuhan), though that was early in the epidemic.
I was responding to the claim that the government was trying to cover up the wet market origin in early 2020. This is inconsistent with the head of the China CDC saying it came from the wet market in January 2020.
Obviously, all of their incentives are to cover up a potential lab accident.
The problem is that everybody has incentives to cover up, whether or not they are at fault. If someone things something MIGHT be read as placing the blame on them, they'll try to suppress it. This makes things quite murky.
FWIW, I believe that if the disease originated (as a human disease) in, say, New York, but didn't get the last mutation that it needed to become pandemic until someone in Wuhan caught it, we'd see largely the same response. You can't judge things by "who's trying to cover things up".
But you have to ask why George Gao would change his view to denying the wet market origin if it came from the lab. That's a strange thing to do if China's intention was to cover up a lab leak. What these statements from China and the various lab narratives/hypotheses have ultimately done is leave the origin unresolved within the domain of global public opinion, an outcome that may not be the worst from China's POV. They're able to avoid officially taking the blame for the pandemic and can control the narrative domestically to say the virus came from outside China.
That would only work if the research was completed but not published before COVID happened. If someone was studying furin cleavage sites in 2018, they would want to publish a paper saying "hey, we tried cleavage sites X, Y, and Z and only X worked" and China would have no reason to think that should be kept secret because COVID hasn't happened yet.
Even a 3 minute skim of anyones work here suggests they probably believe "math is so great you get the best answer anyway", and something something wave at criminal forensics existing
This is discussed in the debates and in Scott's summary. It is a common theme with conspiracy theories (term used literally) to allege that the conspiracy is both capable of and motivated to make the evidence come out to whatever happens to be the case. However, just because you can imagine a hypothetical in which this might happen, doesn't mean it's remotely likely, especially when characteristics like who is involved and how competent they are change from argument to argument. China is not omnipotent and frankly probably doesn't care much about lab leak vs zoonosis. Let's consider the epidemological data pointing to the market, which is often alleged to be tainted.
1. The Chinese government first tried to cover up the existence of the pandemic at all, and failed. Also, as Peter points out, if it started much earlier, you would have many times more deaths, which is even harder to cover up.
2. They also tried to cover up (or at least, were incompetent) the wet market, by claiming animal testing showed no results, but then it turned out they hadn't actually tested live animals (roughly, I might be getting details wrong here).
3. Now they claim the virus started in the US; if they were trying to fake evidence of this, they would have had cases cluster near an airport or hotel or something, not a market that is almost exclusively visited by residents.
4. There's multiple different lines of evidence that point to the market, which makes it increasingly unlikely that China could manipulate all of them in a way that's undetectable.
5. If you don't believe the market data, why do you believe the pandemic started in Wuhan at all? Couldn't China be trying to hide, like a secret bioweapon plant in a different city? This is possible, but of course there's no evidence for it.
Maybe you can help me understand how this would fit into a Bayesian framework, but here’s how I’d think of it:
Suppose someone did indeed to a research and create that exact strain of virus that was ostensibly “not what you’d do” if you were doing research. This would be pretty direct evidence overwhelmingly in favor of the lab, right?
Yet if that evidence existed it would be far easier to track down and destroy, precisely because someone would know “holy crap I did this” and that specific person would have a very strong incentive for that evidence not to go public.
I don’t believe that “China make the evidence go however it wants”, I just don’t see how to account for “people who knew they caused this could destroy that evidence far more easily than we could find it, and would have a very strong incentive to do so.”
> Suppose someone did indeed to a research and create that exact strain of virus that was ostensibly “not what you’d do” if you were doing research. This would be pretty direct evidence overwhelmingly in favor of the lab, right?
I don't follow what you're saying here. It looks like "Suppose it was created in a lab, that would be evidence in favor of the lab." Can you rephrase? Is there a statement missing here?
> Yet if that evidence existed it would be far easier to track down and destroy, precisely because someone would know “holy crap I did this” and that specific person would have a very strong incentive for that evidence not to go public.
I agree that *if* some lab workers knew they had created covid, and knew how they had done it, they probably would and probably could destroy at least some evidence. It's not at all clear to me how well this would actually work (i.e. would any evidence remain? would there be evidence of this sort of cover up?) Do you have more specifics?
(Even in the case of lab leak, there might not have been any such cover up--I have seen the theory that the leak was totally accidental and came directly from samples, so they might not have even realized at first even if they were to blame, and WIV workers were behaving pretty normally through December at least. Conversely, even in the case of zoonosis, panicked workers could have engaged in some sort of cover up if they had some vague idea they might be considered responsible. So FWIW I don't think that either "there's some evidence of a cover up at the lab" or "there's no evidence of such a cover up" would be definitive evidence either way.)
> I just don’t see how to account for “people who knew they caused this could destroy that evidence far more easily than we could find it, and would have a very strong incentive to do so.”
Criminals try to hide evidence all the time, but still get convicted (or convinced to take plea deals). It's actually really hard to successfully hide, destroy, or fake all of the evidence for something this big, and also do the same for the cover-up, and for that cover-up, etc. There were whistle blowers for previous outbreaks, including lab leaks of Sars 1, as well as for Sars 2, from the hospitals, in spite of Chinese crackdown.
You can certainly put a cap on how confident you are in a theory--it's not like the judges, or Miller, or Scott, or me, or the superforecasters, are >99.9% confident in zoonosis. Seems like mostly in the 80-95% range. But it's not correct to just throw all the evidence out.
I agree with most (all?) of what you’re saying but still don’t understand something. I couldn’t glean this from the article.
How did “people have an incentive and ability to destroy evidence” influence the computed odds here?
Yes i understand that destroying evidence would likely happen either way (due to panic, not wanting to be blamed, not knowing the true cause) and that it can be hard to totally cover something up and that people often are convincted despite trying to cover up evidence, etc.
I’m not questioning the outcome of the debate.
I’m just trying to understand how “people had an incentive to destroy evidence” influenced the computed odds. It seems like the conclusion is “that doesn’t matter because there are multiple trails of evidence that point to the zoonotic case”, but I can’t tell if this is “it shifts odds by 5%” or “we just didn’t compute that because it doesn’t seem necessary” or what.
> How did “people have an incentive and ability to destroy evidence” influence the computed odds here?
Miller spends a fair amount of time in the debates on why the early case data is reliable enough, in ways that can't easily be faked, as well as reasons to believe that China is neither capable of nor willing to fake the data in a way that explains the evidence. So I think the conclusion is "that's theoretically true but it seems unlikely to explain anything."
I think the highest level answer is "everyone rounded the BF used for each piece of evidence down, based on how likely they thought the evidence to be wrong." If you want the specific numbers the judges used, you would have to dig into their reports, or perhaps ask them directly. As far as I can recall, I don't think Rootclaim really dove into this point, but rather just used it at a very high level ("those data could be wrong, so the evidence provided is weak") so the judges may have just taken it has most likely that the data are accurate.
Keep in mind that the data merely being limited (which is known to be the case) doesn't necessarily tell you much, because you have to explain why the cases cluster at the *market* specifically. If want to claim the pandemic started at the lab, it's not enough to say that maybe there are cases missing, because it would still be a hell of a coincidence that the cases we do know about are all at the one place in the city most likely to be the origin of a zoonotic pandemic.
Thank you! This seems to answer my question. I appreciate your patience here, it’s a real service.
A few minor niggles...
> 1. The Chinese government first tried to cover up the existence of the pandemic...
It's a bit more nuanced than that. When people started coming down with mysterious cases of a pneumonia-like disease during the first weeks of December, the local hospital MDs and administrators alerted the health authorities of Wuhan, and they in turn notified the Mayor of Wuhan and the Governor of Hubei. The mayor and the governor tried to hush it up because a big party convention was scheduled to be held in Wuhan (early in Jan IIRC). The mayor and governor of Hubei Province decided to go forward with the conference. They felt a small outbreak of some unknown pathogen would reflect poorly on their administration, and they didn't alert the national authorities to the outbreak—until people were crowding the emergency wards, and they couldn't cover up all the hospitalized patients. (This would have been around the third week of December when they couldn't cover it up anymore—although it didn't reach the Western media until the first week of January.)
Local dithering and silence wasted two critical weeks, but some of the hospital MDs notified the folks across the river at the WIV, and they were able to alert the National Health Commission of the PRC that they suspected they had a SARS-like outbreak on their hands. Not long after the MDs were beaten and arrested by local Gong An Bu, probably at the behest of the local authorities who were embarrassed about the outbreak happening on their watch (again, it's unclear who directed them to do this, but the Mayor and the Governor had the motive).
The national health authorities took the warning seriously, and they informed Chairman Xi that the situation was potentially quite serious. He immediately fired the Mayor (and then the Governor) for incompetence and he sent in a hand-picked troubleshooter to handle the situation. (I wasn't able to find out what happened to the mayor and governor. I don't know if they were allowed to fade into private life or were maximally disciplined.) Once the National-level health authorities assessed the situation (and it took them about a week), they alerted the WHO at the end of December, and the Chinese started sharing data with the WHO, the ECDC and the CDC. The local coverup wasted two critical weeks, but it was at the local level, and not at the national level.
> 2. They also tried to cover up (or at least, were incompetent) the wet market...
Due to the case cluster, the Chinese national health authorities immediately focused on the wet markets in that district. SARS1 and some Avian Flu outbreaks had started at wet markets, so the wet markets within the initial case cluster were immediately suspect. And once it was clear that they had something dangerous on their hands, public health authorities immediately shut down the wet markets in Wuhan (I believe they shut them all down all over the city). The animals were disposed of, and the market stalls were disinfected. From a public health perspective, this was a wise decision—but from a data-gathering perspective, they lost the chance to test the animals when they incinerated the carcasses. Again, I may be misremembering some of the details but they invited the WHO to visit the site in mid-January. Either the Chinese or the Chinese with the WHO collected swabs of DNA and RNA from the stalls. So the disinfection happened after they collected the DNA and RNA. The Chinese sequenced the virus and they sent the genome to the WHO, to ECDC, and the CDC on January 15th. It wasn't until March when the CIA director and Trump started accusing the Chinese of covering up a lab leak that the Chinese stopped cooperating with us. Then naturally, they blamed us for creating the virus.
It's important to note that wild animal farming is a 7 billion dollar business in Hubei, and Wuhan is the second only to Guangzhou as a transshipment center for the animals and their byproducts. It's a big business, and the authorities don't want to see it shut down because it lines their own pockets. So their impulse is to cover up instead of disclose because of all of the negative attention they've garnered worldwide.
What's important to realize here is that the CCP doesn't want evidence to exist for *either* of these theories. In both scenarios, the CCP would be partially responsible, which is obviously bad for their optics. Ultimately, COVID didn't hurt China relative to other countries, and so they don't really have any incentive to prevent this from happening again. Any evidence that would narrow down the truth, any truth that would implicate China, would have been destroyed already.
I enjoyed this piece - I tried to follow this story in 2020 and have always leaned more zoonosis than lab leak (I believe I said 90-10 publicly in 2021, probably closer to 95-5 now).
Despite this, I've often been frustrated and annoyed by mainstream virologists who dismiss LL with poor reasoning, and almost always find myself defending (more reasonable) LL proponents.
I guess that this annoyance has led lots of people in rationalist-adjacent circles to lean further LL than they should have. I wonder if this is a common rationalist failure mode we need to update on.
Without caring about which one is actually true, we can look at the costs of the beliefs of they were true or false. If LL is true, and you fight against some sort of conspiracy, that is some positive U. If LL is false and you fight against an imaginary conspiracy, that will worst case cause skepticism and more investigation. Believing that it's zoonosis means you will be enforcing a conspiracy if you are wrong, which is negative U. Otherwise slightly positive U to most people, but less than what you'd get if you were right and fighting a conspiracy. Eliezer says one should believe that which is true, but actually this sort of analysis of costs and gains of being wrong / right is very useful, because it can help to navigate the unvertainty of what you should believe in.
If we want to go Hansonian about it, saying you think lab leak is at least plausible signals sophistication and open mindedness
I did not take signaling that precisely into account but this is absolutely true and might have some wider implications (I did not have solid concrete examples at time of writing).
It certainly is an important point. Every statement you make about the probability of something is at the same time a signal you send about what kind of person you want to be perceived as, and/or want to perceive yourself as.
This is unavoidable.
You're leaving out the utility of finding another reason to blame some group you dislike. I feel this is a major driver of many human beliefs.
Unless you go all Pascal's wager (which is obviously bullshit), your beliefs don't have immediate consequences in the world.
In the brains-as-Bayes-engines model, beliefs are probabilistic in nature.
What you do about your beliefs is then another question subject to cost/benefit analysis regarding your utility function. You might want to state your true beliefs, exaggerate them, stay quiet, pay lip service to the opposite side or even do your very best to argue against your own beliefs.
There are some corner cases where you might chose to stay strategically ignorant due to direct effects of your beliefs (e.g. infohazards, unconditionally knowing about your top 100 health abnormalities might encourage hypochondria in all the cases where you don't have serious health problems, or psychiatric disorders being at least somewhat memetically infective in populations), but for the most part the Litany of Tarski holds: if X is the case, then you want to believe X.
I think Pascal's Wager is true. Probabilistically speaking, life is meaningful and consequential enough to treat it as such
If I had to give a number on a test I might say that our life is 60% likely to be materialistically determined. To strengthen the point, we'll go with 99%
Now multiply the likelihood by the consequences. Again, to strengthen the point, let's say it's probably not very consequential to both exist in a 'meaningful' and non- random universe and correctly believe that it is. So 99% random or probabilistic existence x 0 consequences of being correct or incorrect about that. And 1% non-random x .1 consequences of it being true. Point one units of consequence weighed against 0 seems very noticeable in an absolute sense
With my own weights of likelihood and suspicion that human beliefs and actions can be wave like, propagating and amplifying despite small discrete effects, then my take is that I am compelled to act as if my life and beliefs are potentially consequential and subject to free will
>If LL is false and you fight against an imaginary conspiracy, that will worst case cause skepticism and more investigation.
The worst case of convincing people that someone is out to get them is that they decide to preemptively attack those people.
It's a little weird to cite Eliezer as the person who says one should believe that which is true! This really seems like something that should be much more fundamental than one guy who's famous for an unrelated issue (AI safety).
He wrote less wrong and also a screed on what "the simple truth" is, as well as "Harry Potter and the Methods of Rationality", not "Harry Potter and the sealing of horcruxes" or similar.
I don't think you can say he has *no* claim on being the truth guy, or he isn't famous for wanting to believe true things.
Thank goodness he came along in the early 21st century and upended thousands of years of philosophic practice by inventing the idea that people should believe true things to be true!
I mean, that might not be a novel idea, but it's not really universal either. You'll find plenty of people insisting that sometimes believing or teaching false things -- even tricking yourself into believing things you don't know to be true -- is a good idea. (Plato and Pascal being two famous examples.)
To be the Truth Guy, he doesn't have to say it, he has to be the first or the best in some way. A lot of the problem is that EY's fans tend to be naive about the history of thought, so it all seems new to them. It's not even the first iteration of out-of-the-mainsteam philosophy that claims to answer everything.
The thing called "rationalism" has little to do with rationalism as the term is understood in mainstream philosophy. It is better understood as part of a pattern where a succession of small, insular groups engage in a very self confident style of amateur philosophy. Unlike other cult-like groups, they claim to have a strong orientation to science, logic and reason(and to be able to implement them much better everyone else).
Examples include Yudkowsky's "rationalism", Critical Rationalism (Karl Popper's philosophy, articularly David Deutsch's version), Ayn Rand's Objectivixm and Korzybski's General Semantics. There are often several such movements overlapping each other, some are reactions to others, and some borrow ideas from others.
In more detail, This kind of Thing is characterised by:-
• Being science-orientated , but having much more specific claims than "science good"
* Being strongly to stridently irreligious and anti religious. Also opposed to New Age thinking and continental pilosophy.
• Being largely outside of mainstream academia etc
• Being an insular group that mostly talk to each other
• Having difficulty in communicating with outsiders , in any case, because their own theories are expressed in a novel jargon.
• Centering on a charismatic leader, with a set of mandatory writings
• Having an immodest epistemology..which claims to be able to solve just about any problem..
• ..which is based on a small number of Weird Tricks. Nothing requires the same level of deep study as mainstream academia, and the key ideas are often brief slogans.
> To be the Truth Guy, he doesn't have to say it, he has to be the first or the best in some way.
I used lower case truth guy, clearly in reference to him becoming well known in some respect for this. I use it in the same way that I would call a somewhat competent local mechanic the "car guy" or I can call Bill Nye "the science guy" without claiming he's the best person in the world at science, or that he invented science.
Or that I was replying to the statement
> It's a little weird to cite Eliezer as the person who says one should believe that which is true!
Which, like, do you believe that you should only cite the things from the best formulation or inventor? Maybe start auditing Kenny about what other dastardly crimes against citation he has committed would be better.
I feel it's not a coincidence that this type of misreading is common to, as one would say, "haters",whom all seem to share the following traits, excuse me they are Characterized by:
> Perceive any group seeking improvement to be automatically suspect
>Incapable of reading even slightly positive sentiments about their target and stay silent
>seem to predictably misread incredibly common use of language, seeing "good vibes" and assuming it means GOAT or inventor being a common example
>Incapable of putting forth or discussing any object level example, preferring instead to talk in innuendos about how there's a large population of informed disagreement, yet fails to link or write something. If an example is provided, it is full of guilt by association and sneering arrogance and little else, and they converge on it like little flies, congratulating each other about how many good points have been made.
> does not seem to be aware of their own constant isolated demand of rigors, or that many of their objections can be answered with less than five minutes of thought
> become obnoxiously self congratulatory about how they used middling to low wit to dunk on their targets
You see, listing a bunch of negative traits is a great way to be truth seeking. If someone tries to engage, it's because they are insecure and you, yes you, personally are correct, no matter how obviously derailing or irrelevant that list is. If you can play to an audience, that's all that matters. :-)
Well, you know, one can choose the quality of their friends, but not their enemies. There are ways to choose the quality of your enemies, but the world is not ready to hear them.
I am puzzled that you include Karl Popper here, his view on the demarcation criterion between science and non-science (i.e. that a hypothesis should at least in principle be falsifiable) is very mainstream among scientists. (I do not know Deutsch' interpretation, though.)
If you believe in malicious plots that aren't real you reduce social trust which is one of the main predictors of sociatal succes. Messing with that has huge potential negative U.
I think this is definitely a common failure mode in general, and I observe it way more often in rationalist circles.
Personally, I experienced the exact same thing with the mask and airborne debates. I still get an emotional reaction every time I see someone wearing a cloth mask.
FWIW, cloth masks are a major help in many environments. They slow down the speed of the air ejected by someone who is infected, so it tends to remain near them. They also absorb moisture, so any wet globules ejected are likely to be absorbed.
Note: Almost all masks are more effective at preventing an infected person from spreading the disease than they are at preventing someone from catching it.
I fully agree with you, but the standard response is that (1) wearing masks is "dehumanizing", (2) we will all catch the diseases anyway, sooner or later, and (3) many people wear the masks incorrectly anyway. Therefore, the proper way to signal sophisticated thinking is to refuse wearing the masks. Especially when the cost of doing so will mostly be paid by the others.
Yeah, I think. There's a common failure mode of overcorrection from "don't blindly follow the mainstream" to believing the most popular contrarian option by default.
You see a similar thing in the left with people over correcting from "America is always right" to supporting genocidal dictators as long as they oppose the USA.
As Scott described in the article, Lab Leak is the "mainstream" view among the general public, even if it's contrarian in academia and prestige / elite / media / whatever you want to call it circles.
JFK conspiracies are also the "mainstream" view. I think it's possible for "mainstream" to be a slight "term of art" even if the raw survey data says otherwise.
Well, there is a big Oscar-winning Hollywood movie heavily implying that the official JFK narrative isn't the whole story, so calling that view mainstream seems reasonable in any sense.
> You see a similar thing in the left with people over correcting from "America is always right" to supporting genocidal dictators as long as they oppose the USA.
Ironically, nowadays the right tends to do the same thing, due to associating the US with Evil Wokeists. It really is a trip, seeing politics flip to basically the exact opposite of 20 years ago.
Well there's the question of how far this was an overcorrection. It may very well be that it was correct to put (as a random number) 80% on it given your state of information & willingness to investigate & priors, while 90% was too much.
(I think it is easy for social-posts, like notable blogposts, pointing out that some belief of the community was likely wrong to end up having the same problem of overcorrecting to "don't disagree" rather than "what degree did we do worse than we should have")
I think that there's almost no difference between 80% and 90%. Even if one believes in objectively correct priors, someone who was off by this much is basically as right as possible.
I was using them as random example numbers, the same logic holds for 5% vs 30%.
True.
And there is a danger if offering more precise probability estimates, in particular when offering predictions about the future (and even if you offer them only to yourself): The danger that you may convince yourself that you have more precise knowledge that you actually have, or can possibly ever have.
> I wonder if this is a common rationalist failure mode we need to update on.
Absolutely. I sometimes feel like a very large fraction of comments on any topic on this blog are reaction formation by rationalists who are annoyed at something that MSNBC watchers are saying and have talked themselves into the diametric opposite to rationalize their annoyance.
It's often worse than that--it's rationalists who are annoyed at a caricature of what MSNBC watchers are saying, and have talked themselves into the diametric opposite of that caricature.
True!
Nate Silver justifies his belief in lab leak pretty much entirely with frustration with the mainstream media's attempts at narrative control in 2020 (nevermind that the mainstream media has been breathlessly *promoting* lab leak ever since). It really is sad to see this from someone who should know better.
I don't think this grapples with the problems the Mr Chen case poses for the zoonosis argument. Both sides in the debate agree that Mr Chen contracted symptoms on 16th Dec. This suggests he was exposed around 10th Dec. He thinks he was exposed on the train or in hospital on 8th Dec.
This raises two problems:
1. Mr Chen lived 30km away from the Huanan market on the south of the river, and never went anywhere near it in December, according to his interviews. His case was only recorded because he happened to be transferred to the top tier Wuhan Central Hospital, which is a sentinel hospital for respiratory outbreaks in Wuhan. It is in the north of the river from his local hospital because his relative happened to work there. This is evidence of a geographical bias in he case search. How many other cases were missed on the south of the river because they happened not to visit top tier hospitals, which were largely near the market.
Geographic bias in the case search is also confirmed by:
- Literal statements to the effect that 'we focused too much on the market' by the man in charge of the case search https://www.bbc.co.uk/sounds/play/m001ng7c
- A mountain of evidence in official Chinese documents of a market-bias in the case search. https://journals.asm.org/doi/full/10.1128/mbio.00313-23
- This short statistical argument -https://academic.oup.com/jrsssa/advance-article-abstract/doi/10.1093/jrsssa/qnae021/7632556?redirectedFrom=fulltext&login=false Cases with no links to the market were closer to the market than linked cases. This is evidence FOR ascertainment bias in the case search.
2. Given that there was a confirmed case 30km from the market around 10th of December, this suggests widespread community transmission across Wuhan in early December and late November. i.e. not that there was a zoonosis in early Dec. The first confirmed case was on 10th/11th Dec (even for that case, it is unclear whether the first case was market linked). There was no published retrospective case search. The first case was initially thought to be 1st Dec. Usually if you do contact tracing and a case search, the first official case goes backward in time. In this case, it went *forward* in time to 10th Dec.
As I wrote in an earlier post, I think it was circulating in rural China long before it was officially detected. Catching it on the train fits in with this quite well. My main hypothesis is that a "merchant" carrying produce (of some sort) to the wet market was the source of that epicenter. (That could actually be widened a bit, as all that is required is that someone infected in a rural village for some reason visited the wet market area for awhile.)
Remember that prior to special tests being developed, COVID was being diagnosed as "atypical pneumonia", at least in a Washington (I presume state rather than DC) hospital.
>Remember that prior to special tests being developed, COVID was being diagnosed as "atypical pneumonia", at least in a Washington (I presume state rather than DC) hospital.
Good point!
Personally, I'm somewhat skeptical about the "geographic focus" evidence because there so very many asymptomatic cases. Even if there had been consistent diagnosis from patient 0 onwards, we would have only been seeing the tip of the iceberg.
( I have to admit, I haven't been following LL vs. Z terribly closely. )
It's common for COVID to have an incubation period of only a day or two, so the fact that he developed symptoms on December 16th doesn't exclude that he was exposed on December 14 or 15th.
Rai et al (2022), a meta-analysis of the covid incubation period for patients infected by early variants from January to March 2020, found that the average time from exposure to symptom onset is around 5.7 days, with a 95% confidence interval of 5.2 to 6.3 days. This suggests he was likely infected around the 10th December.
https://link.springer.com/article/10.1007/s10389-021-01478-1
I think you're misreading the paper. The 95% confidence interval means that there's a 95% chance that the mean incubation period is between 5.2 and 6.3 days. It doesn't say anything about how likely a 1-2 day incubation period is in an individual.
I haven't read the paper (I only read the abstract), so perhaps there's information there about how common a 1-2 day incubation period is.
Ah yes, I think you're right. In either case, in expectation, he was exposed around 10th Dec, and thinks he was exposed on the 8th
Does when he thinks he was exposed matter at all? Why should we assign any weight to his speculation? He had symptoms in the hospital on Dec 16th which seems very consistent with contracting COVID in the hospital a few days before that.
I agree it is consistent with that, I am just saying that the average time to symptom onset is 6 days. he could also have been exposed ten days earlier, per this logic, which is a further update from zoonosis and these two effects cancel out, no?
Not sure why this is mysterious.
The first officially confirmed case, the shrimp vendor, got sick December 11. So she was probably infected December 5. She probably wasn't the first human case, because she didn't interact with the raccoon-dogs; more likely a raccoon-dog vendor got (quietly) sick first and spread it to some other vendors. It's not even clear she was in the second batch of cases - only 5-10% of COVID gets detected, probably less when it was first starting, so she could be anywhere from one to four transmissions away from human Patient Zero. So human Patient Zero probably got sick in late November or early December.
Let's say Patient Zero got COVID November 27 (with high error bars). He infects other people at his workplace (the market), and some of those people infect still other people at the market, or other people elsewhere (for example when they're on the train home from work). COVID case numbers double every 3.5 days, so by 12/11 when the shrimp vendor starts feeling sick, there are somewhere around 16 cases (with high error bars - and maybe more if there was a second spillover for Lineage A, but ignore that for now). Any of those people could have passed it to Mr. Chen on the train on the 8th/10th/11th/whenever.
See above comment re this illustrating bias in the case search.
My final comment. It is important to consider how different SARS-CoV-2 is to other zoonoses. I have a challenge for zoonosis proponents to find me a zoonosis with all of the following features
- Spillover occurred after 2000 when sequencing became much cheaper
- There were more than a hundred human cases
- There are zero infected animals.
This characterises SARS-CoV-2, but no other zoonosis meets these criteria. Why?
What evidence do you have that there are zero infected animals? I only see evidence that none of the animals that have been tested were infected with an essentially-genetically-identical pathogen, but there are many animals that were not tested. (I assume the claim you're making is about essentially-genetically-identical pathogens, since there are plenty of samples from animals of coronaviruses closely related to SARS-CoV-2.)
I think there are also some viral hemorrhagic fevers that haven't had their animal reservoir identified yet, though I'm not very familiar with this, and I don't know if any of them have had over 100 human cases.
Yes I am saying that they haven't found an infected animal in their testing efforts, which were very extensive, and included testing of species sold at the market in Hubei and across China. Usually, for a zoonosis with more than 100 human cases, they find numerous infected animals and the first people infected mostly interacted with those animals. This did not happen for SARS-CoV-2.
This is true for: SARS-1, MERS, H7N9 avian flu, 2018-2021 Langya Henipavirus outbreak in China, NIPAH, and all other zoonotic coronavirus spillovers with more than 100 human cases.
I guess you mean at the initial detection? Since obviously many (many) infected animals have been found subsequently.
yep i mean infected with a progenitor
Some important context here is that according to the WHO, China also claims to have found zero infected animals total, during the course of the entire pandemic, even after COVID swept across the country infecting (most likely) hundreds of millions of people.
It seems wildly implausible that there are really zero animal infections (including those passed on by humans) in a region as large as the entire continent of Europe, that also has millions of susceptible species being raised on farms. More likely China is either failing to identify infected animals or failing to report this information to the world.
2013-16 West African Ebola outbreak; almost 30k cases, no animal intermediate. This is probably the most notable zoonotic episode of the last 20 years apart from SARS-2, I'm surprised you missed it. There are also 7 other Ebola outbreaks that match your criteria. We've been studying Ebola for over 40 years and have yet to determine the animal reservoir. It took 20 years to identify the reservoir for HIV-1's progenitor. Sometimes finding the reservoir is easy, sometimes it's hard. Typically it is easy when you have lots of cases and the virus is not very efficient at human-to-human transmission, because that necessitates lots of separate zoonotic events, which necessitates lots of infected animals. For something that spreads fast (i.e., the kind of virus likely to start a pandemic), you don't need a big reservoir, so you have a smaller target. For example, we *did* find the reservoir for the 2009 flu pandemic, but it took 7 years: https://elifesciences.org/articles/16777
I hadn't actually looked into ebola yet, so I take it back
I thought the ebola host was known to be bats?
also hadn't looked at the swine flu one, I was overconfident there.
I agree re HIV, but that was before 2000 and sequencing was much more expensive.
There is strongly suggestive evidence for bats -Marburg virus has been isolated from bats, we've been able to identify antibodies in various bat species that react to Ebola, and in one case we were able to isolate a short bit of RNA that was similar in sequence to Ebola, but nobody's been able to isolate the virus, nor is it clear how it might get to humans from bats.
HIV reservoir was confirmed to be chimps in the mid-2000s, well after sequencing technology was cheap enough to do it. Main limiting factor was just that many chimp populations don't have it, plus having to go into the jungle and track chimp troops and isolate fragile RNA from their poop...not at all trivial to do.
Doesn't SARS1 meet these criteria?
Wikipedia says they found "SARS-like" viruses in civets, which made them think civets were the natural reservoir, but I don't think they ever caught SARS itself in the wild.
There's different types of evidence for infected civets, some of which comes from higher seroprevalence among civet traders.
In May 2003, Guan et al (2003) identified SARS-CoV-like virus in animals in a live-animal market in Shenzhen, Guangdong Province, China. Guan et al (2003) also tested for antibodies among workers in the market. They note that “8 out of 20 (40%) of the wild-animal traders and 3 of 15 (20%) of those who slaughter these animals had evidence of antibody, only 1 (5%) of 20 vegetable traders was seropositive.” This suggests that the majority of the infections of the 11 people with close contact with animals were zoonotic.
Among 508 animal traders, 66 (13%) tested positive for IgG antibody to SARS associated coronavirus by ELISA, while the control groups including hospital workers, Guangdong CDC workers, and healthy adults at clinic had an antibody prevalence of 1–3%. Among animal traders, the highest prevalence of antibody was found among those who traded primarily masked palm civets (72.7%), wild boars (57.1%), muntjac deer (56.3%), hares (46.2%), and pheasant (33.3%). Those for cat, other fowl, and snake were 18.6%, 12%, and 9.2%, respectively.
We have no such evidence for covid. As I have noted, none of the cases in the market worked at the animal stalls; there is no evidence that they had close contact with animals.
That's an interesting result and something similar was later found for wildlife workers in Myanmar and Laos -- antibodies to SARS viruses but the workers themselves having no memory of being sick. However one reason we have no such evidence for COVID is because this kind of testing for workers at the Huanan Market was never released by China. There also is an article stating that one of the first to get sick at the market interacted with animals, although it's unclear to me whether this person was included in the WHO report or not. The first known case at the market, the shrimp vendor, has said she thinks she may have been infected by sharing a toilet with the wild game traders.
Thanks for your deep and well-reasoned dive into a truly important topic. For me, your most important take-home messages were a) Bayesian reasoning has serious practical limits; b) we need to take the potential dangers of viral lab leaks seriously; and c) we need to take the potential dangers of natural zoonotic spread very seriously. One point I haven't seen mentioned is that the Earth harbors roughly 10**27 virus particles.
I'd guess that most of them are bacteriophages...
Probably, and most are marine. But it's a big number even if you cut it x 1,000. It's a big number even if you cut it x 1,000,000. It's a big number even if you...
Well, you _could_ think of it as "just" about one kilomole or so... :-)
It's also interesting how b and c support somewhat opposite reactions to gain-of-function research. If zoonosis is the bigger threat, we'd want to learn more about how this happens.
good point. I'd say that on the one hand there probably *should* be more gain-of-function research, but it needs to be performed far more cautiously. ALL such work on virus should be performed in BSL-3 conditions, using only virions that have been genetically crippled to be intrinsically non-virulent. I guess the question is whether the scientific "community" can ensure that some jerk somewhere doesn't do something really stupid. Unfortunately, we tend to underrate the raw power of Stupid.
a) is particularly important. What this blog post has shown beyond reasonable doubt is that "real" Bayesianism is not for amateurs.
A little knowledge is a dangerous thing. Better to only teach students who will be tasked to make difficult decisions in their professional lives (such as medical doctors, psychologists, investigators, judges and child protection officers) awareness of cognitive biases & where heuristics may go wrong, and NOT to also teach them formal Bayesian analysis - as they are unlikely to have the time & information & sufficiently advanced methodological skills to do it properly (plus be aware of all the pitfalls).
My rule of thumb for something like this is 'whose evidence stands up to scrutiny'. I am not smart enough to understand any of this stuff, but I vaguely knew about the evidence beforehand from podcasts and social media. And then Peter just explained it all away effectively without significant pushback. This never happens in UFO debates. Or political conspiracies. Someone spends a billion hours researching sounds smart makes 30000 claims and then you pick 5 at random and 4 are nonsense and 1 is true or misleading or misinterpreted. I notice that my plausible theories of lab leak after hearing Peters argument about the implausability are *not* those advanced by rootclaim and so I assume they are obviously wrong I just don't know why but I'm confident Peter could convince me instantly. Rootclaim's inputs fell apart under scrutiny', which makes me doubt his statistical analysis. Garbage in garbage out. Note my real confidence about who is right is super low because I am dumb. But the debate definitely moved me away from lab leak 70-30
You are conflating the spoken debate part, where Peter had far superior memorized knowledge, to the written parts, where there is basically no zoonosis evidence that survived scrutiny. We'll elaborate in a post soon.
I look forward to reading it, and I have in anticipation 'rolled back' my change in beliefs until I do!
“Okay we got our asses kicked in the mutually agreed format *that we ourselves proposed*, but now after the fact we want to re-litigate in a format that plays more to our advantages” *really* feels like shifting the goalposts in an unfair way.
If you take that route I think it would only be fair to give Peter access to the same resources you have in doubling down on the research and time to refute your refutations.
i was referring to the written parts of the debate. anything that didn't require real time responses.
I agree it should reflect poorly on Saar that he didn't realize that sophisticated debaters are in fact a huge problem (see: https://web.archive.org/web/20190129225836/http://commonsenseatheism.com/?p=1437 and to be clear I don't begrudge him for not knowing about a link rotted post from 2018 by a niche atheism blogger).
But I feel you should be engaging in the object level too, and realize that a spoken debate in fact is not the best at eliciting truth, for usual status reasons as well as the reasons Saar cited:
1.Debate importance is rated via time spent on a topic, not its actual importance, see Scott's point about the UFO detail debater.
2. It mostly explores how many layers of point-refutation each person has at the moment of debate, rather than what the reflectively endorsed endpoint would be. I would rather people take the time to decide how much each point matters to their main point, be explicit about this ahead of time and then decide if hashing it out is a good use of time. This is difficult to do in real time format.
3. If it is indeed true that Saar has found the ~one epistemology~, it seems much better to agree upon that frame (but not its conclusions) beforehand. This is not just to advantage it, but if it turns out to be as dismal as debate afterwards, well, that should give us a pretty good idea how well it generalizes.
That said, you are correct that retooling the die is a social faux pas, and that we should downrate the second round result.
I think if you are as confident on lab leak as Peter is, you should be rubbing your hands together, salivating that Saar is going to eat crow in HD this time.. Maybe not by Peter in particular, because Peter sick of this and won.
I think Saar should state up front that he commits to doing only this one round of welching / sore losering, and have it framed for everyone to see after, if he wants to dodge the "after wit who keeps coming up with reasons why he is right 3 weeks after losing the argument" accusations.
Seems more like he was fine with the format when he thought he was the more sophisticated interlocutor, and changed his mind afterwards when he lost.
Yes, this was included in my initial assessment of "penalty for not understanding that debates suck". I don't know why people keep thinking it's good to debate! Judges predictably flop the other way they started at the beginning, we know about the existence of truth independent debaters! This is an opinion I've had before I even heard it was a debate, and to my eye this mostly looks like every other case where an amateur debater loses, but had a better case (not that I think rootclaim had a better case, but rootclaim at least pattern matches to non sour grape, honest dismay at debates sucking).
I don't think sophistication is a good way of tracking truth, because sophistication is often displaying that you understand deep nuances and details, regardless of how it connects to the bigger picture. So anything that tracks sophistication tracks more of your ability to fight on specific sub topics, rather than subtopic weighted by centrality to claim.
I think I can be convinced away from this position if Saar goes on to repeat this similar behavior, or if he reflects and decides he was wrong after all. Or if Saar wasn't willing to additionally show his reasoning, or welches on the actual bet.
This is not a niche atheism blogger and the post is not from 2018.
This post has actually left me with a deep curiosity on Rootclaim itself. To be honest, rather than the Nth post on the topic of lableak I'd very much rather read a primer on how to use the method.
I read your reply and found it quite persuasive, and am tentatively back in the lab leak camp
A lesson Scott draws is that Bayes is effectively impossible to apply in complex real-world situations, and that ultimately we are forced to rely on our intuitive reasoning. We’ll just throw a little math in at the end to make sure we aren’t making some clear and obvious blunder. I suggest this means the true epigraph for ACX should not be “P(A|B) = [P(A)*P(B|A)]/P(B), all the rest is commentary,” but rather “commentary, and all the rest is P(A|B) = [P(A)*P(B|A)]/P(B).”
Well, I would argue that Hillel's standing-on-one-foot summary of the Torah was similarly too simple to apply in reality and largely just ends up being a rationalization for our intuitive aesthetic judgements, so I'd say that just makes the epigraph all the more fitting.
My main take-home point also.
Like cost-benefit analysis that is not done by a very skilled and aware-of-own-biases practitioner, hard/real Bayesianism seems to create a danger that uncertainties are hidden in the priors & underlying assumptions & the empirical material used for the calculations, rather than appearing openly for all to see.
> When psychoanalysts claim their therapies work, they don’t mean that someone who just read a two page “What Is Psychoanalysis?” pamphlet can do good therapy. They mean that someone who spent ten years training under someone who spent ten years training and so on in a lineage back to Freud can do good therapy.
And Robyn Dawes in House of Cards showed the evidence doesn't support the idea that training makes them any better.
I had forgotten that over a decade ago I'd written about Dawes vs the concept of "metis" among professionals here:
https://entitledtoanopinion.wordpress.com/2010/11/28/robyn-dawes-robin-hanson-as-antidotes-to-james-scott/
And a follow-up here:
https://entitledtoanopinion.wordpress.com/2011/01/26/bogus-expertise-as-weapon-not-of-the-weak/
Jeremy Howick (The philosophy of evidence-based medicine) cites the derogatory acronym GOBSAT related to the problems with "Metis": Good Old Boys Sat Around A Table (and agreed on diagnosis & treatment).
“One read out of 200,000,000 is completely statistically insignificant,” said Pond. “It really had no SARS-CoV-2. There is no evidence based on genetic analysis there was SARS-CoV-2 in that sample. One read out of 200,000,000 — it could have been a low level of trace contamination.”
What’s more, as Bloom’s preprint reports, Q61 was the only swab above a certain threshold for raccoon dog genetic material that contained any SARS-CoV-2 RNA at all: “13 of the 14 samples with at least 20% of their chordate mitochondrial material from raccoon dogs contain no SARS-CoV-2 reads, and the other sample [swab Q61] contains just 1 of ~200,000,000 million reads mapping to SARS-CoV-2.” When Bloom plotted the quantity of animal genetic material found in the swabs with their SARS-CoV-2 RNA content, he determined that there was in fact a negative correlation between the abundance of SARS-CoV-2 and genetic material from raccoon dogs in the swabs.
Wow this was a great read.
But also.... What happened to discussion of "NIH-funded gain-of-function research" by Peter Daczak and EcoHealth alliance? Is this the "DEFUSE" grant I'm question? It makes not mention of University of North Carolina - that's new to me, though of course I didn't delve deep into the grant.
https://www.science.org/content/article/nih-restarts-bat-virus-grant-suspended-3-years-ago-trump.
I also wish it would be possible to go and collect all the Twitter posts on this topic from when the situation was unfolding, before and after the pandemic left China, including all sorts of claims: spread through plumbing, claims of corona virus circulating months (!) before the official first case, etc. -- do an analysis of how much is real, how much is fake. For example I remember following very closely everything I could read and don't remember racoon-dogs ever being mentioned, but pangolins were.
Pangolins were mentioned in March 2020 as a hypothesis. Raccoon dogs became the more prominent hypothesis by early 2021, when the WHO investigation got samples from the market. Obviously, there are severe limitations on what you can conclude with any confidence from year-old samples, but they're not as much as the limitations on what you can conclude without any samples.
Can someone explain to me Peter’s argument "doubling rates at the wet market points at zoonosis rather than a superspreader event" ?
Intuitively, the two scenario (zoonosis / superpreader event) looks the same to me : an infected (lab worker/racoon dog) (goes at/is brought to) the market and infects market participants. I expect the same doubling rates in both cases.
A least insofar as I remember, "superspreader" was typically used in the public Covid debates to describe cases where the disease vector, for whatever reason, would spread the disease to particularly many new targets. Ie. if a typical Covid patient ended up spreading it to two others (I can't remember the actual numbers), a superspreader would spread it to twenty.
Yeah, that argument was weak and I would have liked to see it either elaborated or conclusively refuted - it's a high-load-bearing part of the zoonosis argument.
Early research showed that the original COVID variant spread bimodally: most of the spread (~80%?) was caused by a small number of vectors. How is the evidence incompatible with the idea that a WIV employee seeded infections that developed over the course of the following week?
Furthermore, even if that model isn't correct, what plausibility advantage does "infected raccoon-dog" have over "WIV employee infected a HSM vendor who then became an vector (asymptomatic or not), infecting multiple people over the course of a week or two"?
The plausibility advantage of “infected raccoon dog” is that there is a good reason for the racoon dog to infect someone (or someones) at the wet market, but nowhere else - raccoon dogs don’t take crowded public transport to and from their office job and their meat shopping.
For the WIV employee, there is an extra layer of coincidence that needs to occur for them to only infect someone else (or at least, only infect someone else in a way that “catches on” and spreads) at a wet market that isn’t particularly close to the WIV itself.
Basically, conditional on the probability of the wet market being the epicenter of the subsequent epidemic, the probability is higher that this is because it originated there, vs originating at the WIV and being carried to the market, not creating other outbreaks elsewhere in Wuhan.
That's fair: I agree that the "infected a vendor" model deserves a penalty relative to "infected raccoon-dog" (all else being equal). A small penalty for the extra layer of coincidence, and a somewhat larger one raised by the question of "why HSM and not anywhere else?" (I don't think zoonotic-origin proponents have convincingly refuted the possibility that ascertainment bias undermines key parts of their case, though I'll grant that my priors have been shifting in that direction.)
Peter's point is that it can't be spreading too much more widely in early December because otherwise when testing becomes widespread you would expect a much higher rate of infection and death therefore we should not consider the HSM initial location to be biased (e.g. there is strong evidence that the first and most subsequent early Human-to-Human infections occurred at HSM). I think Peter (and the Judge's explanations) was extremely convincing that the probability of zoonotic origin is much more likely than an infected lab worker traveling to HSM and starting the pandemic. In addition Peter did a good job reducing my opinion of the likelihood that Sars-Cov-2 is caused by Gain of Function (or any other human manipulation).
Peter did say (something along the lines of) that if you want to assume that a WIV employee left work and went to the wet market then went home and quarantined for a week you are welcome to. It isn't as if he ignored the scenario. I think the point is, if you know you are infected, why go to the wet market at all and if you don't know you are infected then why are you quarantining...
Isn't the most transmissible part of the disease right before symptoms appear? Wouldn't that exactly match this behavior, and there's no mystery?
I would be substantially wrong here if minimal virus shedding happens before symptom onset.
Yup, if you want to come up with contrived scenarios, you can indeed make the data fit any hypothesis.
https://www.goodrx.com/health-topic/infections/viral-shedding-and-how-to-prevent-the-spread-of-infection
Seems to think that this is true, and cites https://www.bmj.com/content/376/bmj.o89#:~:text=Previous%20studies%20suggest%20that%20the,on%20or%20before%20symptom%20onset
Saying that virus transmission starts two days before symptom onset.
I had the data first, before generating the hypothesis. I wasn't sure that I remembered correctly.
Do you think my scenario is still contrived? Or that I said this as a post hoc justification?
> Early research showed that the original COVID variant spread bimodally
I think that’s a very interesting point ; can you find the paper ? I failed to find it.
The links I sent to you elsewhere should cover this.
To me, this argument makes no sense. The market was either: (a) a zoonosis plus a superspreader, or (b) only a superspreader. All of the confirmed cases at the market were from human-to-human transmission. None of the confirmed cases were wildlife sellers (unlike SARS). The confirmed cases were spread across the large area of the whole market.
Your b) is actually “a lab leak origin outside of the market, that only becomes a superspreader at the market”. If the origin is not in the market itself, you need to account for the probability that an infected superspreader from the lab (or infected by someone from the lab once or twice removed) infected a substantial number of people at the market, but not apparently anywhere else.
yes agreed
well actually unless cases were missed because of ascertainment bias in the case search
Genuinely asking: But do the subsequent case rates actually support that? Sure, initially all the focus was on the wet market, but if there were other simultaneous epicenters surely that would be ascertainable after the fact?
And lab leak proponents would be highly motivated to find such data. Instead they (at least Saar does) seem to mostly concede that the wet market was the source of the first significant outbreak, plus or minus some low level earlier circulation in weakly sourced or anecdotal data.
To clarify, my initial comment was just making the logical point that the market was a superspreader (and maybe also a site of zoonosis), not that it was necessarily the *only* superspreader.
I don't see why you think other epicentres would be ascertainable after the fact? There was a retrospective case search that ended in mid Feb 2020. There was a market bias in play for around half that time and the case count actually fell for part of the case search. In my other comment, I highlight evidence for ascertainment bias. for example, if you read this report about Zhongnan hospital, a hospital next to the WIV BSL-2 and BSL-3:
"The South China Seafood Market [link requirement] alone ruled out most of our suspected cases." Yang Jiong said that Zhongnan Hospital is 10 or 20 kilometers away from the South China Seafood Market. Most of the suspected cases received in the early stage did not have a history of exposure to the South China Seafood Market, but they had obvious clinical manifestations of unexplained viral pneumonia. However, the district experts refused to identify them as unexplained pneumonia after consultation. They could not be referred to Jinyintan Hospital, and samples could no longer be sent for genetic sequencing"
https://docs.google.com/document/d/1mK2CFD8IeSYrT26BLVRuiJSnSaK-Q8-5K71byJUsmbA/edit
Because many cases in this area were not market-linked, they could not be confirmed at Jinyintan (the premier infectious disease hospital), so there was limited opportunity to identify clusters in this area.
Similarly, Mr Chen was only registered as a case because he happened to be transferred to a sentinel hospital north of the river. This suggests that other cases where Mr Chen lived (35km from the market) would just never have been found.
You have to assume not just early cases missed because of ascertainment bias, but also some explanation for why there aren't five times as many cases a couple weeks later if it had been spreading a week longer than hypothesized.
Before kvetching, I want to concede that this is a good argument, and it's been a big part of why I've updated away from assuming ascertainment bias.
However, I have never seen anyone substantiate its premises. I think that's important because there are two related areas of uncertainty underlying the point: first, that we got sufficiently accurate epidemiological data out of China during that early period and second, that we know what the R0 was at the time well enough to make the judgement.
Beyond that, the epidemiological dynamics of COVID were weird: more "punctuated equilibrium" than steady spread. Admittedly, my understanding of that is mainly rooted in subsequent history, when NPIs were in effect and people were on guard, but this is not an uncommon model for how diseases spread.
Maybe Peter addresses these issues in the debate, but I have yet to see anyone do so to my satisfaction, so my position remains something like "there's only a 50% chance we have information enough accurate to make these kinds of judgements."
(Edit - reading down thread, I see now that Peter linked to his analysis of the spread, which I'll read as soon as I have time.)
So, am I right in thinking you don't think there was ascertainment bias in the case search? I'm not sure I understand. In that quote above, you have >10 hospitalized cases at Zhongnan that don't get registered because they didn't have a market link. Assuming there are ~10x total cases, that's already 100 cases you don't see for market biased reasons, (in an area around the WIV)
Whether Patient Zero is a careless WIV lab tech who is also a wet market customer, or an unlucky wet market vendor, they both go home at the end of the day, probably a family, and very probably a normal life where they go out and do things at least until they get sick. So in either case, the pandemic starts with A: a superspreader event at the wet market and B: an infected guy going around spreading the disease the ordinary non-superspreader way. I'm not seeing how you can use observed doubling rates or anything like them to distinguish between the two.
If there were confirmed cases predating the market outbreak, those would give us great insight into what happened, but AFIK there aren't. The fact that there aren't, just tells us that Patient Zero infected very few people outside of the market and they all just thought it was a cold.
Peter seemed to have adequately debunked, "superspreader at market," based on growth rates and start dates. Why do you ignore this here?
See above comment re talking past one another.
I feel like he's leaning very heavily on the average spread, but that's essentially his argument. That if early COVID spread like we know it did later, then the size of the spread would have been much larger if it was earlier, which Saar's argument requires.
Given small numbers at outset, I don't think he can actually rule out low and intermittent spread for a few days maybe weeks before catching in a more randomized way. If someone was infected early and then only spread to one or two other people and they only spread to one other person, it may have delayed the virus from spreading as we saw it later. Once it hits a certain threshold that seems very unlikely, but below that threshold we really can't say that.
Indeed. There is a paper (can dig it up if you want) showing that (iirc) the model number of people to infect (if you're infected) is 0 and the next most common is 1. Many many chains either die out completely or take several cycles to explode.
Yes, I’m very interested if you can find that paper. It could help shed light on the the question "on average, how much time do we expect to see between the first case and the first superspreader event ?"
One key point is that Peter insist on "very little time because exponentially doubling rate", which I believe is wrong.
Yes, I agree - I think the uncertainty about the start date is far larger than he's implying from "we know the doubling rate, so therefore it's tightly bounded". If you know the exact doubling rate you can talk about aggregate dynamics using this math when the number of infected is large, but when the outbreak is small in size the variance can be very large.
Now onto the specifics - COVID-19 transmission is characterised by a large degree of overdispersion. This means most people infect very few others, and the vast majority of transmission is driven by a small fraction of the infected. Multiple studies from 2020 found that over half of infected people never infected anyone themselves. Here are some links:
a) https://www.nature.com/articles/s41591-020-1092-0 - See Figure 3
b) https://pubmed.ncbi.nlm.nih.gov/32685698/ - See Figure 2
c) https://www.nature.com/articles/s41467-020-20235-8 - "The remainder (57%) did not lead to a transmission event"
d) https://www.nature.com/articles/s41598-020-79352-5 - See Figure 2
Sure, but why would the virus only “break out” at the market? I don’t think the claim is that these early cases were some less virulent strain (the A and B types are both pretty infectious). So if it was spreading at a low rate throughout Wuhan, there are a *lot* of potential breakout sites where it didn’t break out, or create multiple epicenters at roughly the same time as the wet market.
Which isn’t totally implausible! But it’s probably less likely than “no really, the first cases were spread in the wet market and not somewhere else”.
That's the other part of Peter's argument that I didn't like, the almost slavish following of "less likely = wrong" when at any point something unusual could very well have happened. That seems to be Scott's point at the end, where we know for sure that at least one of two very unusual things actually did happen.
I don't know if I have enough knowledge to say that it is "probably less likely" or not, but that's not the final determiner of truth.
This strikes me as a weird argument to defend someone whose whole project is rigidly following Bayesian probability. “Probably less likely” IS the final determiner of (as-close-as-you-can-get-to) truth.
I get his point, but I simply disagree with it. If that means I disagree with Bayesian reasoning then I'm okay with that too.
I think you miss my point: slavish devotion to Bayesianism isn’t Peter’s position, but Rootclaim’s.
If you reject Peter’s argument on non-Bayesian grounds, then Peter has lost the battle, but Rootclaim loses the war.
Something as high as like 60-70% of people infected with Covid appear to not infect anyone (in a seronaive environment). It might even be more skewed than that once you account for under detection of asymptomatic cases to begin with (I'm guessing the never symptotic don't spread it much)!
This makes chains that die out, or chains that barely survive for a few generations and then explode completely plausible.
I understand “superspreader” to mean something like “an infected individual who sheds way more virus than usual in a confined environment creates a localized, discrete spike in infection rates”. Or in any case “a discrete event with a much higher rate of infection than the base rate expected by the observed growth rate of the epidemic”.
Peter is saying the wet market doesn’t look like that, it looks like a “normal” rate of infection and growth in cases.
Why does that matter? The two sides of the argument both need to explain why the epidemic was first detected and seemed to spread from the Wuhan market, rather than somewhere else.
Saar’s claim is that basically Covid leaked from the lab sometime prior to the wet market outbreak, and was circulating around at a low rate, but a “superspreader” event in the wet market made it the epicenter of subsequent infections. This theory takes a major hit if the spread in the wet market was not atypical, because in that case the market should just be one of many similar outbreak sites started by the various infected folks circulating around, rather than a clear epicenter.
Peter claims that the wet market was the actual origin of the zoonotic transmission. It looks like it all spread from there because it *did* all spread from there. The fact that there was nothing “super” about the spread at the market supports this theory - wet market infections don’t need to swamp other potential epicenters because there are no other epicenters.
No, superspreader is, "went to my own wedding with COVID and kissed all 200 guests.
Either way, it comes down to one person infecting way more than the average 0-2 people per case, right?
It does. But it is far more likely to be situational then something anomalous about viral shedding.
> I expect the same doubling rates in both cases.
Yes, this does make sense, which is the point. Such a scenario is a priori unlikely for lab leak, since it requires a lab worker to go to the market and infect someone there, but infect no one anywhere else, either in the lab or in between, since there are no other early clusters. To make this hypothesis seem more plausible, Rootclaim asserted that the market provided very good conditions for much faster growth of Covid, possibly even a superspreader event (i.e. a case where a single person infects a dozen or several dozen people all at once, greatly in excess of the average number of secondary infections, which for early covid is less than 2. This would explain why the market dominates the early cases, but as Peter points out, these claims have no evidence and are contradicted by available data. If the spreading rate is the same inside the market as everywhere else, then the strong clustering of early cases to the market is very good evidence in favor of the pandemic starting there (and thus, in favor of zoonosis).
Bu the upshot is that this argument is specifically a response to Rootclaim's argument, so if you don't believe the latter, then this part would be irrelevant for you.
This is a better summary than I was able to put together.
I think Rootclaim needs to defend their assertion of a superspreader event at the market, because otherwise Peter seems to have convincingly shown that it wasn’t, and if it wasn’t it’s pretty good evidence in favor of the wet market theory.
I don't think they're going to do that, and from Saar's point of view they don't really need to. This is his whole shtick with "both sides shouldn't have high BF" and Scott's analogy with the Muslim UFO spotter. Maybe it wasn't a superspreader event at the market; that was just one possible explanation. But this example (allegedly) shows that such explanations are not that rare or hard to find, and so this evidence must have a low Bayes Factor.
In something like this, you might end up with weirdly extreme probabilities simply because global pandemics are rare, and will necessarily be the result of a chain of individually unlikely events occurring. I know in theory all these probabilities are supposed to be “conditional on Covid actually happening” but that’s easier said than done.
I get the Muslim UFO Spotter analogy, but I still feel like Saar is a bit dodgy here.
Your model should account for all possibilities even if there is just a “what if this did happen, but not in one of the specific ways I already quantified” uncertainty term.
If your conclusion is correct, but for a reason your model didn’t even anticipate, you just got lucky.
As it is, it looks like Rootclaim simply didn’t consider the possibility that the wet market wasn’t a superspreader, which raises the obvious question “what else did they miss”.
> In something like this, you might end up with weirdly extreme probabilities simply because global pandemics are rare, and will necessarily be the result of a chain of individually unlikely events occurring.
I think this is likely to be true.
> I still feel like Saar is a bit dodgy here.
Oh, I agree that Saar's argument is extremely dodgy. His argument relies on the Bayes Factors that he produces being both high and extremely robust. If that is not the case, it totally falls apart. I was just explaining why I don't think Saar is likely to further support the superspreader claim, or care about it.
Which is a shame since Rootclaim kept repeating the claim in various contexts (based purely on Scott's summary. I haven't actually watched all the debate itself). Given that Peter indicated this as a week spot in their arguments, repeatedly referring to the superspreader event without justifying that claim is ... strange.
I'm confused about the "both sides shouldn't have high Bayes factor" claim. With plenty of ordinary claims, I can very easily come up with extremely high Bayes factor pieces of evidence on both sides.
Like consider the question of whether I was born in India or in North America. My last name is a very high Bayes factor piece of evidence in favor of the India hypothesis, and my first name is a very high Bayes factor piece of evidence in favor of the North America hypothesis. I haven't thought of any evidence I can provide that has a high Bayes factor in favor of the hypothesis that I was born in South America over the North America hypothesis (the fact that I currently live in Southern California is probably single digit Bayes factor evidence in favor of South America over India).
In this case, there's very good explanation available for the anomalous evidence (my father was born in India), and not every false hypothesis has an equal amount of high Bayes factor evidence.
But if you do a very serious search for anomalies related to something, you're likely to find a few anomalies in favor of some particular false hypothesis, even if most false hypotheses don't have any in their favor. There's a reason we're debating whether it was from the wet market or from the virology lab, and not considering whether it spread from pet dogs, or was planted by the CIA, or came from bats living in someone's attic.
I believe what Rootclaim argues is that if you have a high BF on conflicting hypotheses, then one of them is either wrong or p-hacked. Your name example is a good one, though, since each piece of evidence is so similar it's hard to allege p-hacking for either side. I wonder how strong each of those pieces of evidence actually is? Like "Kenny" being 1/1000 in favor of NA and "Easwaran" being 1/5000 in favor of India? I have no idea what these numbers actually are.
And of course, unlikely events can still happen. For most people, their first and last names probably don't provide such a high level of disagreement. I think this has to be the case, right? If lots of American-born people had Indian last names, then the Bayes Factor for "Indian last name" would by definition not be very large? I'm just thinking about this quickly and not doing math.
> But if you do a very serious search for anomalies related to something, you're likely to find a few anomalies in favor of some particular false hypothesis, even if most false hypotheses don't have any in their favor.
Yes, I think this is related to Rootclaim's assertion of p-hacking. But I think this is also an easy failure mode, which several people have pointed out is prevalent in the online lab leak community, of basically going on fishing expeditions with a high number of degrees of freedom and not accounting for the number of opportunities for an anomaly.
> I have no idea what these numbers actually are.
I also have no idea - I could imagine them being anywhere from a factor of 10 to a factor of 1,000,000!
In any case, I think that it's a bit awkward to talk about "p-hacking" when we're trying to do a Bayesian analysis. We know that most of the evidence people come up with is going to be gathered in a highly non-random way!
I'm a huge proponent of Bayesianism, but I don't see much prospect for explicit calculations of this sort to be that useful in this kind of format, at least in part for this reason.
For the exponential math argument and its relevance, start watching here:
https://www.youtube.com/watch?v=d1dbfoK8nSE&t=870s
It shows that the market outbreak can't just be an artifact of ascertainment bias because there weren't enough cases at that time in the outbreak, the size of the market outbreak in early-mid December is roughly what you'd expect to find in all of Wuhan.
So it can still be a lab leak, it just needs to go straight from the lab to the market and nowhere else (with both lineages). That's very unlikely and needs to be accounted for in any sensible bayesian analysis.
If you think that the exponential model is too simplistic to make that point, I've also done more detailed epidemic simulations to demonstrate a similar thing:
https://twitter.com/tgof137/status/1772417277670871113
Universe 1: LL is true. Somebody at WIV contracts coronavirus and goes about their daily life, eventually effecting numerous people in Wuhan. If all the starting cases suddenly show up at some superspreader event (a football event? an orgy? idk), this makes sense. If they all suddenly show up at some specific, non-superspreader event, this doesn't make sense (they'd infect somebody at HSM, but also at the grocery store, at the brothel, at the school their kid goes to, etc).
Universe 2: Z is true. Somebody contracts coronavirus from the raccoon dog, and goes about their daily life. Since they are probably the raccoon dog stall vendor, they spend much of their potentially-infecting-people day at the HSM, so they cause most of their "doubling every 3.5 days" infections to happen there. They're also more likely to infect people at the HSM. This same logic could apply to a superspreader event, but it's less likely - imagine somebody at the football game (e.g. an athlete) being the cause of the superspreader event at the football game.
While it's possible either way that somebody got the infection really started at the HSM, it's much more likely for it to happen in the LL universe if HSM is a superspreader, and much less likely for it to happen in the LL universe if HSM isn't a superspreader.
Wow, this was incredible.
Is it really such a coincidence that the outbreak occurred naturally nearby the lab? It's reasonable (I guess, I don't know) that the lab was strategically placed there in the first place due to the area's wildlife, which host many animals relevant to their virology research. If so, the alleged coincidence significantly fades.
Placed in the middle of a city of 10 million people to be near wildlife? No, the virologists in question were traveling over a thousand miles to random caves in the middle of nowhere to collect viruses which they were then bringing back to Wuhan because living in cities is more fun than living in primitive villages near bat caves.
My theory is that the local government of Wuhan knew they were at risk, because so many rural farmers in the area were bringing wildlife into the wet market. So they funded the WIV. As you say, a big research institute kind of has to be near a city. It's not going to be Bruce Wayne living in a bat cave.
I wish I knew more about the internal politics of China though.
It's somewhere between a pandemic started in San Diego and you discovered San Diego had a relevant lab (8th largest city in the country) and Los Angeles county has a pandemic start in downtown and there's a relevant lab in Pasadena (based on size of Wuhan vs American cities).
I don't think either one is really a fair comparison. Wuhan has *11 million people*, so it's roughly the size of NYC and LA combined. But it's also in an area where there are still a lot of desparately poor people living in the country and hunting whatever animals they can find as game meat, on a scale which just doesn't exist in the US. "Anything that walks, flies, swims, or craws..." (https://www.quora.com/Do-the-Chinese-eat-everything-that-walks-swims-flies-and-crawls)
I was using LA county (although thinking of metro area although the two are not super diffrerent in the case of LA) not LA City (LA metro area has 12 million people). My assumption is a pandemic starting in LA would most likely be caused by travelers so it wasn't meant to be exactly the same but rather it's not surprising for there to be a city/metro area with both a lab and a pandemic. When I was picking US cities (metro areas) my though it that many people had not heard of Wuhan until 2020 but it's not some podunk village.
This answer misses the point. Does it make sense to build a lab tasked with collecting samples for coronaviruses research in the London, in the same way it that it does in Wuhan?
Not sure, but isn't agreed by all in the argument that the other top lab in the world researching Coronaviruses was in North Carolina?
From the post: "Peter: As mentioned earlier, the DEFUSE grant was rejected. Further, the grant said that the Wuhan Institute of Virology was responsible for finding the viruses, and the University of North Carolina would do all the gain-of-function research.".
I don't know what exactly "responsible for finding the viruses" operatively means, but it might mean that their location is indeed non-random, and that this was their "competitive edge" so to speak.
It's "non-random" in the same sense that Miami being a major banking hub for Latin America is non-random. It's not especially close to the places that it's serving, but it is a natural location to be an intermediary between the places that it's serving and the broader world of finance or biological research.
I don't guess, I've looked into it, and no. The lab wasn't placed in Wuhan because of the bats or pangolins or racoon dogs or whatever live near Wuhan. First off, the bats we care about live far away from Wuhan, and any pangolins or raccoon dogs we care about have to have lived near the bats. Second, it's the "Wuhan Institute of Virology", not the "Wuhan Institute of Coronavirology" or "Wuhan institute of Zoonotic Virology"; it was not initially meant to specialize in any particular sort of virus that would call for any particular location.
Probably the WIV is in Wuhan because when it came time to found a big Institute of Virology, the next guy on the CCP's "we owe this guy a favor, let's build a big infrastructure project in his district" list was from Wuhan.
Thank you dashing young blogger!
I think the "lab leak" side of the debate focused too much on the probability that SARS-CoV-2 was engineered, and didn't properly account for the scenario that SARS-CoV-2 was a natural virus collected by virologists that accidentally leaked. In my opinion the latter scenario is more likely than the former.
Tell me more. My impression was that one of the strongest pieces of evidence for lab leak is the weird furin cleavage site. If you say it wasn't engineered, you lose that evidence - what do you gain in exchange?
Why? Couldnt they be stupid and reckless in multiple ways; a little bit of virus breeding, some humanized rats, some random crisper edits all in an attempt to make a bioweapon only to be deluded that they are making any progress. You mix it all together and you get a handful of engineered gene transfer events and a very unclear "zoonotic" pathway
The weird furin cleavage site isn't evidence for that. It's no more likely under that hypothesis than under the pure zoonotic hypothesis.
its evidence of crisper no?
No. CRISPR creates cuts and introductions at planned sites and with planned new sequences. Having something that no one would have planned is not particularly likely.
It is 2024, there is 0% doubt that the Furin cleavage site is natural and was not introduced by an engineer. Long list of reasons include (not comprehensive):
-All previous FCS created by researchers (for example here in 2006, or here in 2009, all published years before the "secret" DEFUSE proposal suggesting similar work got played up as a smoking gun) have avoided using an "insertion", and rather chose to "substitute" - rewrite - existing nucleotides in this position to change the encoded amino acids to create an FCS. This is necessary and basic biology because the addition of new amino acids to any spike protein will likely disrupt its structure and break the whole thing.
- Curiously, an additional, unnecessary proline residue is part of the insertion. Proline is not part of any FCS motif, however, the amino acid is considered a helix-breaker; something that will mess up your protein structure with a high probability. Why risk introducing that?
- Even more curious, the whole sequence insertion is "out-of-frame", meaning the nucleotides were not introduced neatly but rather staggered for no reason, again something that scientists would not do ordinarily (or at all)
- However, the FCS insertion is right next to an RNA sequence motif prone for copy-choice recombination, arguing for this natural mechanism bringing it about (and copy-choice recombination is reading-frame independent and favors out-of-frame insertions 2:1)
- The GC content of the insertion is 83%, very unusual for artificial introductions
- The coded amino acids themselves make an odd, non-canonical, and weak recognition motif (RRXR) for furin-like proteases; a motif that might ordinarily not work compared to strong motifs such as RRSRR (which would usually be used by virologists).
- The FCS cleavage at S1/S2 destabilizes the original (Wuhan-Hu-1) spike protein, which only got compensated partially when the D614G mutation emerged quickly in the early pandemic, arguing for a recent acquisition, not design or cell culture
- Furthermore, potentially because the FCS in the original SARS-COV-2 is not that stable, it is lost very quickly in Vero cells and other typical cell culture systems. Selection pressures seem to run against it in these research-associated amplification systems; so how exactly did an FCS-containing virus that first has to be amplified in cell culture before it can infect a researcher escape from a lab and do so with an FCS?
But these alone is not proof. Who knows what weird and arcane experiments those Chinese researchers would cook up in a lab, right? (Some lab leak truthers with no molecular biology experience would say. For actual genetic engineers, the situation is already clear as day) Fortunately, we do not need to rely on experience or speculations about any motivations or experimental procedures Chinese scientists might have followed.
This is because nature just has more secrets hidden in the FCS that humans could not have come up with in the first place:
After intensive mechanistic studies of the FCS in SARS-CoV-2, researchers discovered a hitherto unknown synergistic interplay between the odd, suboptimal FCS and the genetic backbone of the virus (which we by now know is natural because of recombination).
- We have learned that pre-cleavage of SARS-CoV-2 S1/S2 by furin is what boost TMPRSS2-mediated cell fusion and thus impacts cellular tropism and transmission dynamics (ergo infectivity)
- But it turns out, that was not the full mechanistic story. In nature, no genetic element acts alone
- The viral QTQTN amino acid motif is an uncommon natural sequence in several sarbecovirus spike proteins directly upstream of the FCS sequence
- In cell culture experiments, QTQTN is not stable and repeatedly lost, similar to the FCS. Loss results in impaired viral replication.
- QTQTN determines how tightly the loop harboring the FCS is bound to the spike protein, thus it regulates spike protein stability and shapes how well the Golgi-bound furin-like proteases have access to recognize the FCS motif and pre-cleave at S1/S2
- The QTQTN motif is also glycosylated and loss of that glycosylation impairs viral replication as well.
- Researchers showed beautifully that there is an intricate functional interplay between the FCS, the loop length, and glycosylation: These separate but co-dependent elements work together synergistically and all elements are ultimately required for efficient FCS pre-cleavage (and with it, impact viral replication and pathogenesis). In nature, no genetic element acts alone.
Co-dependency and synergy between genetic elements are hallmarks of evolutionary selection; these type of complex interactions are extremely hard to design for even with perfect knowledge of all structural components, which nobody had in 2019. In the case of SARS-CoV-2's FCS, there is no conceivable way how engineers could have designed the observed - but hitherto undiscovered - mechanistic synergy purposefully.
Take this together with all the other unexplainable design choices listed above; and you have your answer. The polybasic cleavage motif recognized by furin was not artificially inserted
Isn't the out of frame point relative to the backbone used? On the lab leak hypothesis, the backbone used is unknown, so we don't know the insert is out of frame
Nope. Out of frame with respect to the used reading frame for the whole spike, and that is based on the coding amino acids that barely differ between the various spike proteins because they are functionally conserved
I see, thanks. Is it possible that there might be differences between the spikes of some virus that they had that we don't know about that would make the insert in-frame?
Short answer no, longer answer: still no, but never say never; but we would need some really strong evidence to suggest something like that being possible
Yes, it's relative to whatever nucleotides are there already. Though I tried aligning it to every known similar virus and it was out of frame for all of them.
In theory, the leading Serine could have been spelled with different codons, in whatever virus the lab started from. There are 6 codon choices for serine, only one would work, so I called this a 1 in 6 chance at the debate. Probably a crude way to do it, but Rootclaim wanted to quantify everything.
So then isn't it wrong to say that the insert was definitely out of frame? It is correct to say that if there were an unknown virus that the WIV following the workflow outlined in DEFUSE, the insert could have been out of frame but could not have been. From a Bayesian point of view, one has to weigh this against the possibility that the furin cleavage site emerged naturally in a farm in Hubei. The 850 sars-like coronavirus don't have a furin cleavage site, suggesting it is selected against in nature.
One also has to think about the timing of the acquisition when weighing the probabilities. The furin cleavage site emerged the year after the WIV was on a proposal to insert the site. The mutations in the virus also suggest that the furin cleavage site emerged recently around Oct/Nov. Choe and Farzan (2021) argue that “The notable emergence of D614G suggests that the acquisition of a destabilizing furin site was a recent event”. So, zoonosis has to posit that the furin cleavage site emerged only once naturally, was never found in nature despite them looking, and caused approx 1-2 spillovers in Wuhan city.
Yes. I dealt with it in a probabilistic manner by assuming the original virus could have different codons.
I suppose what you really want is the ratio, i.e. 1/3rd of natural inserts would be in frame, but let's assume closer to 100% of lab inserts would be in frame, so you could change my numbers that way. But I was already being generous by assuming that the natural viruses have even priors for all codons at that spot, when it's actually out of frame for every one we know about.
The presence of the furin cleavage site itself is dealt with separately in the probabilities. To some degree that may just be included in your priors -- i.e. the judges both assumed a serious coronavirus pandemic on the order of 1 in 100 years, it wouldn't be serious without the furin cleavage site, so it's double counting to include that.
Saar used a factor of 2, that the lab pandemic is twice as likely as the natural pandemic to have such a site. That's probably double counting, but maybe not, since his pandemic prior was 1 in 20 years. I could see going up to a factor of 5 here since we don't know exactly how often the serious coronavirus pandemics are.
Then the last question is do you get separate odds for it being an "inserted FCS" vs just an FCS. This is the real thing Saar thinks is important, he gave it a factor of 50 at the debate but later said he privately thinks it's 1 in a billion, or something like that. He thinks that the lab would totally put a 12 NT insert into a coronavirus but nature could never do that.
Strangely, he actually has that backwards -- no lab in history has ever inserted 12 NT's to make a cleavage site. They typically mutate existing amino acids, rather than inserting. On the other hand, nature has shown that it repeatedly does 12 NT insertions at random spots, including sometimes at or near the FCS. So the odds here probably actually lean towards nature, if you understand how virologists normally work (Saar does not).
It's not the least bit clear that the DEFUSE grant suggests to "insert the site" -- that's a highly motivated reasoning that few virologists agree with.
Finally, the fact that this is a sarbecovirus with a furin cleavage site is kind of weird -- it's never been observed in such a bat virus. But those are enteric viruses, not respiratory viruses, so the FCS is selected against. In an intermediate host, it's quite likely that this feature would be selected for. We just don't know of many sarbecoviruses which have spilled over into something else... I think that's just SARS1, SARS2, and a couple pangolin viruses? Maybe there are others I don't know about.
So it's kind of a 0 divided by 0 situation, going off prior odds. 0 out of 5 labs have done anything like this. 0 out of 3 natural viruses were just like this.
Very hard to look at that situation and confidently assign some strong bayes factor for the lab.
Why is all the above incompatible with forced rapid evolution via e.g. serial passage in humanized mice?
Many technical reasons. Maybe the simplest: Viruses adapt to their host system. The original SARS-CoV-2 could not infect mice. Even with mice carrying a human copy of the ACE2 receptor, the virus shows none of the characteristic adaptations that it acquired once it learned to infect mice (alpha, delta varients) that one would have expected to see earlier if it has ever been passaged in such a system. The real problem is actually transmission, but that is more complex to quickly write out here.
Thank you for this thorough explanation of that particular point! I hadn't followed it properly before.
The existence of the furin site means relatively little because no FCS implies no pandemic and we wouldn't be asking the question. The strongest evidence for engineering is how closely the virus matches the DEFUSE proposal.
Weissman does devote some space to what he calls the ZL hypothesis and finds that even if there was a purely zoonotic spillover, it probably made it to Wuhan via research.
Good point -- it's easy to double count in a bayesian analysis, so if you start with the odds for a serious natural pandemic and a serious lab leak pandemic, you probably can't go on to include things like "the virus has a furin cleavage site" or "the virus binds to human ACE2". (Rootclaim did both in their analysis)
Weissman gets this right (and may have been the first to explain it). That seems more relevant to me than your victory over Rootclaim.
I lean towards lab-leak and am skeptical early Covid would have developed in nature.
Beyond that, what I don't think is addressed, I think it's significantly more likely that a new pandemic coronavirus that pops up in the one city in East Asia with the biggest coronavirus labs, rather than any of the other hundreds or thousands of cities with wet markets, would be an entirely natural virus that made its way there due to improper safety precautions by the lab. The lab collected coronaviruses, was full of both resident and visiting researchers who work with coronaviruses, and was known for having lax safety standards.
Even if Covid wasn't engineered in the lab (I think it was), I think it's more likely that it would be a natural virus that appeared in the wet market by way of proximity to the lab than by way of an infected wild animal brought directly to the wet market from the wild.
This would be essentially a zoonotic lab leak but I would award it to team lab leak.
I don't think there are a billion people that live in cities as prone to wet lab spillovers as Wuhan is. Given that 11 million of those people live in Wuhan, this is at most a factor of 100.
Weissman is his blog post (linked by others - https://michaelweissman.substack.com/p/an-inconvenient-probability-v57) makes a pretty substantial seeming point that Wuhan actually had a very low share of the wild animal consumption trade. So Pr(Wuhan|Zoonotic, wet market) <<< 0.01.
Scott, I don’t understand understand your question. The only reason the lab leak is considered a possibility at all is the improbability of a coronavirus pandemic starting right next to one of very few global coronavirus research centers (indeed a facility which at that time had sampled and sequenced the closest-then-known relative to COVID-19).
Additional claims about furin cleavage sites may support certain scenarios within the lab leak umbrella, but their strength or weakness does not detract from that starting improbability.
What you gain in exchange is that you no longer need to assume that the Wuhan virologists were actually doing any engineering of their viruses.
And the furin cleavage site doesn't look engineered to me, someone who wanted to engineer a furin cleavage site would use different amino acids (as the debater mentioned).
If it originated in the wild at anytime at all that’s where it would have spread from. No doubt it has been subsequently studied though.
I'm guessing that Wilf didn't argue this theory for two reasons:
1) He views it as much less likely than a leaked gain-of-function experiment. Rootclaim's website shows a probability of 89% of gain of function escape, but only 3.2% for a zoonosis collection accident.
2) According to their public discussion, it was explicitly out of scope of the debate:
Peter Miller:
>Would a zoonotic collection accident be considered a lab leak or zoonosis, for the judging of this debate?
Saar Wilf:
>The judges will have to decide which of the two hypotheses is more likely: GoF vs unassisted zoonosis. So it's possible they are both wrong, and there will still be a winner.
https://www.astralcodexten.com/p/contra-kavanaugh-on-fideism/comment/12792183
That's really interesting! I had convinced myself that zoonotic collection accident was the main plausible lab leak hypothesis, though gain-of-function leak is also relevant.
It is unfortunate that there's not much discussion of the three hypotheses explicitly - everyone seems to present it as a binary choice (though they sometimes differ between which two options are presented).
I don't know of any fourth hypothesis that is meaningfully supported by anyone.
To be honest, I find the zoonotic collection lab leak theory to be the least interesting variant of the lab leak theory. If all that a Wuhan researcher did was move an existing virus into a populated area, then they did nothing for the virus that could not have been done by a miner who stepped in bat shit.
Right. But the lab is continually going out and looking for new caves full of unsampled viruses, while miners look for a site with minerals and keep working there.
If the pandemic was caused by 'zoonotic collection', there is every probability the pandemic might otherwise have never happened. Scientists take their samples from very remote locations, and deliberately get close to bats in a way that most miners would be averse to.
This is intuitive, because no pandemics have been caused by miners, despite there being many more miners underground at any moment than virologists.
In other words, it would be a massive deal. And going forward it would have huge implications for what precautions scientists should take when sampling (if they're permitted to sample at all).
I don't think it's likely for bat virus sampling to produce a pandemic. In the whole history of the WIV, collecting 20,000+ samples, they only ever managed to culture 3 sarbecoviruses. Usually it's just enough RNA to get a sequence but not so much that you can get yourself sick.
There are also signs that the starting virus was already adapted to a respiratory system, whereas the bat viruses are enteric viruses. The furin cleavage site is the most notable such feature, but there are other smaller ones. So it was in some species other than a bat, before it jumped into humans.
I suppose if researchers went out sampling raccoon dogs or some other animals and a researcher got sick, that could do it. But I don't see why the hypothetical scenario of researchers going out and doing that should be any more likely than zoonosis at the market, where we know for sure they were importing lots of exotic animals.
The main thought I have is that while there’s a lot of people gathering wild animals for wet markets, the relatively few lab workers going out to gather wild animals are specifically looking for viruses, and will thus be picking up disproportionately many (whether it’s overall greater or equal or less is not obvious).
Another way to look at it is that the only known traffic of sarbecoviruses from the location of all of the closest known relatives to covid was via science, not via the animal trade.
Exactly. The reason to at least consider 'zoonotic collection' is that we know researchers were going and continue to go collect samples from remote locations where the closest known relatives to coronavirus exist. Only one cross-species jump is required.
In contrast, all animal trade scenarios currently involve multiple cross-species leaps between unknown animals with essentially no idea as to how, when, or where that occurred.
I don't understand why you think it's impossible for an enteric virus in bats to evolve into a respiratory virus in bats?
The issues with extreme odds, focusing on the big picture, and eating one's losses are all related to the same phenomenon: The high complexity of the underlying processes.
Reality is not generated by a simple "zoonosis" factor competing against a "lab factor", as one would think if one naively took the simplest Bayesian models seriously. Rather there was a wet market and a lab and a bunch of people all doing a bunch of different things, etc.. Way too complex to explain. (Though because of funky math where we condition on the fact that there was a pandemic, we also can't ignore that there is a competition between these two factors; they just aren't the generators of reality, and therefore not of the evidence either.)
It is extremely common to have pieces of evidence with extremely low priors, because these pieces of evidence contain a lot of information about the specifics of these underlying processes, and probability drops exponentially with information. Notably, contra focusing on the big picture, accurately integrating this evidence requires building a narrow, specific picture of what happened, because otherwise you don't really know how to evaluate the interactions between the probabilities when conditioning on multiple things.
As you build these complex, low-probability models, that means you end up eating a bunch of losses, in rough proportion to the evidence you add to your model to account for the complexity. If your resulting model of what happened doesn't have very low prior probability, you're not thinking in enough detail.
Important link: https://www.lesswrong.com/posts/JD7fwtRQ27yc8NoqS/strong-evidence-is-common
Thanks for the comment and the link. Useful when warning bright students not to be too much born-again Bayesians. (Beware of the young convert.)
Yeah, I was very unclear what's going on with the claim that it should be unlikely to find high Bayes factor evidence on multiple sides of a debate.
It's certainly unlikely that any particular high Bayes factor evidence exists for a false hypothesis - but many high Bayes factor pieces of evidence for the *true* hypothesis are *also* very unlikely!
But if a lot of people are digging very hard, they're going to find a lot of unlikely things. And if there are some false hypotheses that are in some sense "close to true", you'll probably find a few things that strongly support them.
I'm somewhat unsure about the math for Bayes factors per se. Assuming you're doing your Bayesian updating correctly on a good model, you should be well-calibrated, so you shouldn't be updating outside the 0.1% to 99.9% range until you actually have a ~final conclusion. This sure seems like it means that strong Bayes factors should only line up for the true hypothesis.
I think the way one gets extreme Bayes factors is model misspecification. Like both hypotheses often need to eat the odds for the information contained in a piece of evidence, as they both need to be updated to give high probability to that information, so the low odds for the information doesn't directly give strong Bayes factors. But then the models should be ready to predict similar information in the future, and if they confidently disagree on what "similar information" means, that gives you the crux for what settles the debate.
But if one of the models get updated wrongly, for instance by not adding a factor that explains the information, then when similar information comes around, the model that was updated wrongly gets forced to eat this information too, which might turn into an extreme update against it.
Did either side get into the US intelligence assessments at all?
It seems to me that by far the strongest argument for lab leak is the US government conclusions to that effect. These agencies have an awful lot of raw data that no one else does, and the resources to put a lot of analytical effort into the problem. Obviously, they're fallible but they have access to the same scientific data as everyone else plus whatever kinds of useful classified raw data the US government has scooped up over the years and they employ experts on stuff. That's enough for a pretty decent presumption in their favor, especially when so much of the debate centers on data from Wuhan that could easily have been manipulated.
I'd find the zoonotic origins argument a lot stronger if someone from that side could explain what the FBI and DOE may have seen that led them to get it "wrong" based on some premise other than incompetence.
I also found it odd that official government positions were not discussed. Even if one side could dismiss them, you'd think they'd still talk about it?
Of course, independent researches without access can't talk about classified data they don't have, and that makes the argument an appeal to authority instead of something that can be modeled and debated.
You can absolutely talk about classified data you don't have -- historians and journalists do it all the time and learning to read through redactions is a fundamental skill in those fields. There are still valid inferences to be drawn even if you can't see the underlying material; you read around the redactions and see what could be under there and what it could mean.
In this case, for example, the fact that different intelligence agencies disagree with each other indicates there is not some kind of classified smoking gun. Like, if there were tapes of the scientists saying "oh shit we did this" intercepted by the NSA, then all the agencies would line up behind that.
We also know that the FBI and DOE both think a lab leak is the explanation, but according to the declassified report, they reached that conclusion for *different* reasons than one another. We also know that the DOE changed its position between 2021 and 2023 based on some kind of new intelligence received in that period, whereas the FBI thought it was a lab leak all along. Those are some very interesting data points.
You can also consider what the intelligence community is geared towards understanding and what kind of information they could plausibly have. These are not public health agencies. They're geared, more than anything, towards gathering and analyzing information about the Chinese government and military. So, they real comparative advantages in understanding the implications of things like Chinese cover-ups or disinformation but no comparative advantage in a debate about furin cleavage sites or whatnot. That's also important and gives us some sense of what could fill in the blanks.
I haven't done the work myself to work through this (and was hoping someone else would because I'd be very interested) but there are also plenty of bread crumbs in what has been publicly released that you can start following.
The most important exercise, though, is just to try to put yourself in their shoes and think "What could they know that would lead to conclusion X" and especially "What could they know that would *incorrectly* lead to conclusion X." If you can't give a compelling answer to that second one, then you have to rethink your own confidence level.
Those are fair points, but the problem is made more complicated because in addition to not having access to the evidence, we also don’t know in any detail what the conclusions were. The DNI says that the FBI and DOE both “assess that a laboratory-associated incident was the most likely cause of the first human infection with SARS-CoV-2.” That could mean they think that a researcher collected the SARS-CoV-2 in the field and become infected by it. It could also mean that they think SARS-CoV-2 was genetically engineered in a lab and then transmitted to humans. Those are rather different scenarios.
If you are the US president, you can perhaps trust what the intelligence services will tell you.
If you are the public, you can not trust them at all.
FFS, we are talking about groups who have worked hard to stage coups and went out of their way to torture prisoners in gitmo. Psyops are one of their tools. Do you really believe they would draw the line at lying to the public?
Before the NATO warning about Putin became true, my general stance was 'the tell for lying in public statements from the intelligence community is that their lips are moving'. I still refuse to update on them, even if they tell the truth, they don't want you to believe that because that because it is true (simulacrum level 1) but because it is advantageous for them if you believe it (level 2).
Yes, I am old enough to remember Powell trying to convince the UN security council with his 'intelligence' about Saddam's weapons of mass destruction.
You don't think Powell believed what he was saying? Is Powell even an intelligence guy?
Wouldn't one simple reason be that US is in a cold war with China, and in such a situation intelligence agencies might be predisposed to prod people towards explanations that place China in a worse light than other poential explanations?
Weren't there several intelligence agencies that reported their overall conclusion was moderate support one way, and several that reported their overall conclusion was moderate support the other way, and none that reported their overall conclusion was strong support for intelligent design?
The FBI assessed at medium confidence that COVID was a lab leak. No one else got above low confidence.
The Department of Energy (really meaning the US national labs) assessed at low confidence that COVID was a lab leak.
The National Intelligence Council and 4 unnamed agencies assessed at low confidence that COVID was zoonotic.
The CIA and one other unnamed agency didn't reach a conclusion one way or the other.
While it seems reasonable to assume government agencies might have better evidence than the public, what's actually happened on this topic has been people in the government pushing stories that end up being false or exaggerated.
These have included claims such as scientists at the lab being the real first cluster, that SARS2 was part of a bioweapons program, that there's evidence a safety incident happened at the lab in late 2019, that a vaccine scientist was executed by being thrown off the roof of the lab, that the virus was already spreading at the Wuhan Military Games. All of these are from US government sources and all of them appear to be false, exaggerated or lacking evidence. Exactly what leads high-level national security officials to push things like this without good evidence isn't totally clear -- maybe it's the intersection of US-China geopolitical rivalry with a complex, unknown topic -- but all of this suggests the FBI and DOE's conclusions might need to be taken with a grain of salt as well.
We also know that neither of their assessments changed the views of any other agencies, which seems surprising if they really had strong evidence. IMO it's more likely their assessments simply represent the view of particular analysts at these agencies, who came up with their own theories about what happened based on limited evidence like everyone else.
Am I the only one who Peter lost at the end when he starts asserting that any evidence that disagrees with him comes from liars? And his "evidence" that people are liars is stuff like "it was reported in the Daily Mail" and "he was a drug user" and "his cat got it too" (by which Reed obviously meant tested positive, which can absolutely happen to animals, inanimate objects, and all kinds of other random things). Followed by, "and everyone else was also a liar".
I have a friend who is convinced he caught COVID in Jan 2020 before anyone knew what it was or what to watch out for. He had symptoms that were an exact match. Officially, COVID did not exist here at that time. Is he also a liar? I assert that he is not - he has no reason to lie about this, and is an honest and upstanding gentleman. I therefore assign it a 1:10,000 chance that he is being deceptive, and thus COVID clearly spread far earlier than officially stated.
I also have a heuristic about people who dismiss evidence on the basis that it appeared in a tabloid, or wasn't in a peer reviewed papers: naive and gullible. Here's some Bayes for you: what's the probability that a journal publishes an article full of photoshopped images and garbled machine-generated text? My prior for this happening in an academic peer-reviewed journal is high, because there keep being incidents where whole editions of journals get retracted due to every article being gibberish. My prior for it happening in the Daily Mail or any tabloid is zero because it's never happened. Clearly, the Mail is far more reliable than peer reviewed journals and Peter should accept that its evidence overrules anything he's read in journals.
Rephrased: most of Peter's evidence implicitly assumes the testimony of China, virologists and epidemiologists is reliable (e.g. the ridiculously precise stated doubling period), when we know it's not. Many times he makes statements to which a good response might be, "and how do you know they aren't simply wrong/lying", but Saar was too generous to do that. Peter did not return that generosity!
Ultimately, although it was interesting and Peter's performance was impressive, I don't know if this kind of debate helps. I remain in camp Lab Leak.
> most of Peter's evidence implicitly assumes the testimony of China, virologists and epidemiologists is reliable (e.g. the ridiculously precise stated doubling period), when we know it's not. Many times he makes statements to which a good response might be, "and how do you know they aren't simply wrong/lying”
They were lying for about a month. That’s why Li Wenliang was arrested for spreading malicious rumours. However he was the one who first mentioned the market.
Yes, I don't understand this (paraphrased) claim by Peter: (EDIT: this seems to be Scott making a mistake in paraphrasing Peter)
> He also told the Mail that his cat got the coronavirus too, which is impossible.
'Impossible', thus implying the man was lying? I was under the impression that, quite aside from cats having tons of coronaviruses in general (FCoV being a particularly serious threat to young cats, which also seems to be a remarkable case study of the harms of the FDA), that it was not just not 'impossible' for domestic pet cats to get the coronavirus too, it was routine for them to get COVID-19, and even other cat species in *zoos* have tested positive and this was true very early in the COVID-19 pandemic and quite well publicized and well known (eg April 2020 https://www.nationalgeographic.com/animals/article/tiger-coronavirus-covid19-positive-test-bronx-zoo ). This was a topic of interest to me at the time because I like cats and have a cat and was wondering what the implications of me being inevitably infected might be for my cat, and so I remember this quite well despite my general attempt to remain ignorant of as many COVID-19 matters as possible... And double-checking now to see if all of these reports were somehow false positives or faked, I continue to see everyone like the CDC stating that it is still totally possible and routine for cats in close contact with infected humans (you know, like a *pet* cat) to be infected with COVID-19: https://www.cdc.gov/healthypets/covid-19/pets.html
Given that Peter has supposedly spent years autistically researching every last detail and this detail in particular in order to discredit that British dude, I'm experiencing sudden Gell-Man Amnesia here about the rest of his claims, as well as the supposed experts evaluating Peter's claims if they didn't flag that (I have not checked).
indeed, cats are very susceptible to covid
Generally speaking I think that:
* Peter makes a lot of very solid points; more than Saar does.
* Peter frequently manages to completely dismiss some of Saar's claims; Saar hardly ever does this to Peter.
* Peter is just generally a better debator.
But...
* Saar is arguing honestly and presenting his actual views.
* Peter is dishonestly making the strongest case he think he can for his side, which includes deliberately saying incorrect and misleading things.
I trust Saar much more, despite probably broadly thinking that the lab leak is <50%.
Yeah, I found this comment from Peter on the Manifold Market a bit concerning:
> So what's my incentive to come up with the most reasonable values possible, if my opponent just says any data I present is fake/meaningless/actually supports his theory via some twisted argument?
> By the end of the debate, it was more like, "fuck it, if he's just going to make up numbers, I'm going to stick with largest numbers for my case".
https://manifold.markets/chrisjbillington/will-bsp9000-win-the-rootclaim-chal#2C7VearXa03px5PI8p6A
The context for that is Saar asserting that the market data carries 0 weight as evidence for Zoonosis. Peter is saying that he hewed to an argument which was *clearly* unreasonable instead of contesting numbers which Peter thought were reasonably subject to debate.
>Saar is arguing honestly and presenting his actual views.
>Peter is dishonestly making the strongest case he think he can for his side, which includes deliberately saying incorrect and misleading things.
This makes sense when you remember that one of the debaters is just some guy and the other one is very wealthy. If 100k is a very meaningful part of your net worth, you can't afford to not play to win.
"making the strongest case he can for his side"- isn't that how a debate is supposed to work? It's certainly how the legal system works. It's a competitive process, you're not supposed to help out the other side. The *judges* are the ones who are supposed to be neutral and get at the truth of the matter.
Yes, and this is the result:
https://www.youtube.com/watch?v=__reD_phfbg
During the debate, Peter also cited a paper showing that the first Beijing superspreader event (another wet market, what are the odds?) had originated in frozen fish.
Oh, it's in the transcript above. Peter says:
"We know why there was an outbreak in the Xinfadi Market in Beijing: it was because the seafood stall got frozen fish from some non-Zero-COVID country, the fish had COVID particles on it, and the vendor got infected and spread it to everyone else. Something like this is true for the other Chinese wet market based outbreaks we know about it."
I think the dispute may arise because COVID is not a logically defined disease. Early on it was, back in 2019 before the PCR test for it was developed and diagnosis was handled by doctors, but once public health officials got involved the definition changed to become circular: COVID is defined as having a positive test, a positive test is defined as having COVID.
This leads to all kinds of stupid stuff like the way governments insisted that COVID testing had no false positives whatsoever, even when that clearly wasn't true. Except it kinda was true given their violent abuses of logic - if the test is positive then by definition you did have COVID, even if you never got sick.
Due to this it's possible for the cat to both "have COVID" and also "not have COVID" simultaneously. We might call it COVIdinger's Cat. The cat cannot actually be infected by the virus or get sick under any conventional understanding of disease, yet it can test positive anyway simply due to having some human-generated virus particles somewhere in its body. The PCR tests don't care and will detect tiny quantities of fully destroyed virus, which is why they could trigger in people who'd been sick and recovered weeks earlier. And because COVID is defined by the presence of a test, that's enough: the cat has COVID!
This doesn't justify Peter's claim. Both of my links confirm that cats do exhibit physical symptoms and disease traits like 'dry coughs' and 'loss of appetite', among many other symptoms like 'vomiting' or 'diarrhea', and in the case of the zoo cats, those symptoms were why they were tested at all for COVID-19 and the blood samples could then test positive for COVID-19 in addition to them being sick. Thus, it is clearly possible, and not 'impossible' for a cat to have COVID-19 beyond a circular test result, and it is established fact according to the CDC, Bronx Zoo and every other source I have checked that cats really can 'have COVID' under any 'conventional understanding of disease' and do not simply merely carry it while testing positive (which a lot of *other* animals do - just not domestic cats and a bunch of their felid relatives).
I think it's an error with Scott's paraphrasing -- Connor said that has cat died, not that it got sick. Here's his quote:
"Pictured: Connor and the cat that hangs around his apartment in Wuhan, China. Suddenly, I’m feeling better, physically at least. The flu has lifted. But the poor kitten has died. I don’t know whether it had what I’ve got, or whether cats can even get human flu. I feel miserable"
From a dailymail article:
https://www.dailymail.co.uk/news/article-8075633/First-British-victim-25-describes-coronavirus.html
I am well aware that cat's can catch the coronavirus (though my cat never got visibly ill any of the 4 times covid's been in my household). But the prior on an pet cat dying from it should be low.
Another commenter above suggested that "maybe the cat just tested positive". This is late November or early December 2019, before covid was known to exist and before covid tests were invented. The only way that Connor himself would even know if he had it is if he went to a hospital and they saved a sample and then the hospital tested it later. Connor claims that the hospital called him weeks/months later and told him he tested positive, but there's no other record of that. Either way, they most definitely did not save a swab from his cat and PCR test that.
The prior on a 25 year old going to the hospital from covid should also be pretty low. This also can't be a sampling bias thing, given the predicted number of covid patients at that time, there aren't going to be many and even less are going to speak English -- even in the case growth model that Rootclaim tried to present, I calculated there would be at most 1 at that time)
The last interesting thing is that Connor says that he shopped at the market!
Specifically, he wrote:
"My local paper back in Llandudno, North Wales, has been in touch with me. Maybe I caught the coronavirus at the fish market.
It’s a great place to get food on a budget, a part of the real Wuhan that ordinary Chinese people use every day, and I regularly do my shopping there.
Since the outbreak became international news, I’ve seen hysterical reports (especially in the U.S. media) that exotic meats such as bat and even koala are on sale at the fish market. I’ve never seen that."
So, that's not 100% explicit as to referring to Huanan market, but Huanan market was the one in the news, and there was a specific rumor about Koalas being sold at Huanan market, so it seems pretty clear that's the market he's referring to.
So, either:
A. He's lying about the whole story and just fabricating it from stuff he's seen in the news.
B. He's a super early patient who got infected at Huanan market, and he's great evidence for the zoonosis theory.
C. His story is true except that he actually shopped at a different market other than Huanan, and that makes him the super secret early case that proves lab leak.
I interpret this as most likely A. Rootclaim sees C as most likely.
In the third debate, I even tried to lay out a quick bayesian analysis for how to decide between these options. I think that's sort of the problem with Rootclaim's method -- they decided from a high level view that they think that lab leak is true, then they went looking for evidence to confirm that view, with no good filter on what evidence should be trusted, i.e. they decided that the cases in the WHO report are 100% fraudulent but Connor Reed's case is 100% likely to be covid. It all just seems like motivated reasoning, to me.
> I think it's an error with Scott's paraphrasing -- Connor said that has cat died, not that it got sick. Here's his quote:
This isn't about Connor, this is about Peter. Did Peter claim what Scott paraphrases him as claiming, and apparently no one in the debate called him on it? (I'd just check myself, but I've now searched the YT auto-transcript of like 6 or 7 of these YT videos and haven't been able to find a quote matching it; and I'm definitely not watching the whole thing, so hopefully someone who *does* know where in this rats-nest of links Peter would have said it, can go rewatch it and tell us what he said.)
I'm Peter. I claimed that it was an untrustworthy account because the cat died, not because the cat got sick (as well as all the other issues I listed above).
IIRC Saar later made some counter claims that Connor shopped at a different market and that the cat died of some feline coronavirus. None of which is supported by the original articles, just a purely motivated reading, I think.
Off topic, but I find myself less interested in the epistemology behind COVID's origins (which is still quite fascinating) and more interested in what you're planning to do with your research and persuasion superpower. There are lot of people who capable of being intellectually rigorous but not persuasive verbally. And also vice versa. To have both is quite rare. Most traditional careers after a physics degree seem like a waste.
I certainly intend to research other things in the future, and I hope to be done with this topic soon.
I have never thought of myself as persuasive at all, have never debated before or given many presentations or speeches. It just seemed like it couldn't be that hard -- Rootclaim's case is so thoroughly wrong -- so I went for it. But mostly it was stressful and not very fun, so I'll probably stick to writing, coding, researching, and things like that.
> I claimed that it was an untrustworthy account because the cat died, not because the cat got sick (as well as all the other issues I listed above)...None of which is supported by the original articles, just a purely motivated reading, I think.
We are going off on a tangent here from my original question about the paraphrase, but now that I check... how is that not supported, and how is your description supported by the original articles, exactly? The source here seems to be the Daily Mail account https://www.dailymail.co.uk/news/article-8075633/First-British-victim-25-describes-coronavirus.html which includes photos of the cat, which is clearly a kitten (an important detail to omit), and all of the references to the cat are:
>> Day 9: Even the kitten hanging around my apartment seems to be feeling under the weather. It isn’t its usual lively self, and when I put down food it doesn’t want to eat. I don’t blame it – I’ve lost my appetite too...Day 11: Suddenly, I’m feeling better, physically at least. The flu has lifted. But the poor kitten has died. I don’t know whether it had what I’ve got, or whether cats can even get human flu. I feel miserable.
I do not understand how you go from his claim "I don’t know whether it had what I’ve got, or whether cats can even get human flu. " to your slide https://www.youtube.com/watch?v=Y1vaooTKHCM#t=1h10m30s saying "Claims his cat died from covid" (which you also repeat aloud to emphasize it). You then repeated it again in the slides & aloud (without any additional source so presumably still the Daily Mail source) at https://www.youtube.com/watch?v=CTK1hWOzwns&t=1510s as "He says his cat died from covid...this is what I read in the article". Are we looking at the same 'article'? Because I don't see how you could have read that in the Daily Mail article given the only references to the cat in it which I quote above.
Connor explicitly doesn't claim the kitten got what he got (much less got COVID-19), so a claim that he didn't make cannot make him 'untrustworthy'.
Nor is the cat dying anything to make him 'untrustworthy': it is entirely ordinary for kittens to die of respiratory illnesses like various coronaviruses, to which they are especially vulnerable. (Lots of cat diseases to go around, unfortunately.) This is a major source of kitten mortality in litters, pounds (so notorious as to have their own names like 'kennel flu'), and cat colonies, or cats in general, and there is nothing all that odd about it happening or a kitten dying of it; and it dying is that much more likely if not treated (which it apparently wasn't inasmuch as Connor was busy being sick and being an expat probably couldn't call on anyone for a local vet). So it should be no surprise that kittens die of COVID-19: https://bvajournals.onlinelibrary.wiley.com/doi/full/10.1002/vetr.247 Anyway, young kittens die like he describes of FCoV *all the time* (you may have heard about the Cyprus outbreak last year), so there's nothing odd about the death itself.
See some of the articles Scott dug up, for more inconsistencies:
https://www.astralcodexten.com/p/practically-a-book-review-rootclaim/comment/52739539
Both of those articles say Connor was hospitalized for two weeks, vs previous ones saying he was sick for 2 weeks and then went to a hospital briefly.
The first article also goes on to quote him saying, "I honestly don't think the virus is as serious as all that. The people who have died must have had underlying health problems", which is a weird description for hospitalized for 2 weeks.
I suppose you can't entirely determine whether Connor's story is inconsistent or this is bad reporting.
Same with "the feline coronavirus". How would he know what the cat had? He says he was housebound, wasn't going to the vet. I'm not sure the usual timeline for FIPV. In the Dailymail, he's asking if the cat got the "flu" he had. Then in the Wales online article, he's saying he's not sure if he caught it from the cat. But if he's certain it was "the feline coronavirus", that doesn't even infect people, right?
The path to this all being important requires both that "Connor is trustworthy" and also "Connor is lying about shopping at Huanan" (or else I misinterpreted which market he was talking about) and "We've settled on a specific growth rate for the market outbreak" for which this evidence would contradict and it's somewhat dependent on "Connor got covid on Nov 25th, not during his second wave of illness in early Dec".
I think I'm very likely to win either "Connor is untrustworthy", with "Connor shopped at Huanan" as a backup.
With a certain motivated reasoning, though, you can conclude that Connor is great evidence to contradict zoonosis at the market, and therefore "SMOKING GUN FOR LAB LEAK!"
I also think I have a strong enough model, between case growth rates, genetic diversity, Wuhan serology, etc, to simply put low prior odds on dubious covid infections in November as being likely to be true.
Likewise, if Connor really got sick in November, where are the interviews with 100+ Chinese people that also did, that you would statistically expect? There are interviews available with some of the earliest people sick at the market. Did China somehow retroactively censor every other similar account from before?
It's the same thing with e.g. positive covid tests in Italy in September. I did a laborious deep dive through all those papers just in case Saar brought them up, but you can also determine they're unlikely to be valid if you have a good enough understanding of the other evidence.
In a larger sense, I think there's a certain personality type that gets drawn to the patterns and another that gets fixated on outliers. This dynamic even created a new theory, at the debate:
https://drive.google.com/file/d/1u_Ne53YLJuSPm75jjCRkeQJHGvpRogor/view?usp=sharing
I wrote some code to model and graph mutations that were reversions. There were 2 dots out of place in my graph. I sort of noticed when making it, but it didn't really register in my brain as that important.
Saar noticed the dots later and concluded they were "SMOKING GUN FOR LAB LEAK!" and they coincided with SMOKING GUN DELETED EARLY CASES and SMOKING GUN ANTARCTICA SOIL SAMPLES (which both also turn out to be misleading and unimportant claims under further investigation).
So I dug into the code and the data to figure out why my diagram didn't match Pekar's, and found it was a mundane explanation for the 2 dots and corrected the diagram.
Once you've been through this process dozens of times with the lab leak crew, you kind of develop an understanding that they're just throwing out one bad claim after another, and those rarely ever hold up.
I think this is actually a great heuristic, that's better than bayesian reasoning, for spotting conspiracy theories.
One of the first places I might have developed this approach was in 2020, when I was trying to figure out if the election was stolen, i.e. for curiosity and to know who to bet on in the political betting markets.
The 2020 election fraud people made a lot of claims that the election was stolen. People who doubted it asked "where is the evidence?"
So someone made a site called "here is the evidence"
https://hereistheevidence.com/
It looks like it currently has 1,687 claims as to why the 2020 election was stolen.
My process, back in 2020, was to go through maybe 20-30 election fraud claims and evaluate them. Every single one I evaluated was either clearly false or just ambiguous and misleading. So I concluded that I did not have to read a list of 1,000 and I bet that Biden would still win (I forget the rules of the market, perhaps I was still taking a risk on a successful coup).
It's kind of similar with lab leak -- I've evaluated hundreds of claims now and found the vast majority either blatantly or subtly wrong. Had any of them been convincing, I would have dropped out prior to the debate and moderated my stance on covid origins, maybe even flipped it. Saar credulously relied on many of them in the debate, many more transparently false than Connor Reed's covid case.
From Saar's probabilistic perspective, it's more like: the details for Connor don't even matter. Even if Connor didn't have covid, someone else did, and if he dug hard enough, he's sure he could find that person. And because he thinks that, you can't apply high odds to the market outbreak. In fact, he thinks you can't apply any odds at all to it, he thinks all existing case data is bayes factor 1 between market and lab.
I think Saar has concluded that lab leak is almost certainly true based on the location in Wuhan and the furin cleavage site. Therefore, lots of non-market early cases must exist somewhere, and he doesn't even have to look.
If someone made "here is the lab leak evidence dot com", with 1,700 different examples of early cases, I would go through 100 and notice that 95 are clearly wrong and 5 are ambiguous and whoever curated the site is probably just not very careful.
Saar would look at the list and say, "wow, that's 1,700 different claims. If even one of them is correct, then it's surely a lab leak! What are the odds that 1,700 claims are all wrong?"
I could be wrong about all this, of course. If hospital records for Connor Reed ever turn up, I will apologize, change my mind, and evaluate whether that changes the origins picture based on those details.
Alright, here's where I first made the claims in week 1:
https://www.youtube.com/watch?v=Y1vaooTKHCM#t=1h10m30s
Here's where Saar brought up the discussion again in week 3:
https://www.youtube.com/watch?v=XWMmj8ZprnI#t=55m30s
At 57 minutes, Saar infers that it was not Huanan market.
It is ambiguous, Connor just says "the fish market". I'm inferring that it's Huanan market because he says it's in the news and the thing about koalas. Saar is inferring that it's some other market, but I don't see why.
On the next slides Saar says it was "the feline coronavirus". That appears to be completely made up, if you compare it to the actual quote I provided above.
Saar then claims that there's evidence Connor went to Zhongnan hospital. There's no evidence of that, as far as I know, other than Connor's claim. And I'm not sure if Connor ever even named which hospital. There's no official record of the hospital contacting him and telling him he got sick.
And then here's where we discussed it at the 3rd debate:
https://www.youtube.com/watch?v=CTK1hWOzwns&t=25m10s
Thanks for the links. I did see those in the transcripts but I wasn't sure if they were the key ones or if the auto-transcripts had errors. I don't know what the 'fish market' or 'Zhonang hospital' have to do with anything I've asked, but moving on, I can't see where anyone in those 3 videos claims that "cats can't get COVID-19", and you do seem to acknowledge that they can in the third video when discussing how he could've tailored his story post hoc in 2020 (which would be a strange claim for you to bring up had you ever said anywhere in the debates something like "cats can't get COVID-19" like Scott's paraphrase).
So I think Scott Alexander needs to correct his paraphrase in OP: you did not say "cats can't get COVID-19", you seem to instead be saying that "cats never (ie. rarely) die of COVID-19" (which may sound similar but of course are quite different).
I do, however, still have a question about what you did say about cats; see sibling comment https://www.astralcodexten.com/p/practically-a-book-review-rootclaim/comment/52734427
My mistake; I've changed the wording in the post.
During our discussions, Saar sent me this article, in which Connor does claim it was "the feline coronavirus": https://www.walesonline.co.uk/news/wales-news/welshman-thought-first-brit-catch-17684127
This doesn't make sense to me - how did Connor know what disease his cat had? He said he was housebound that whole period, so he couldn't have gone to the vet. Also, it seems to contradict his previous story about thinking it was COVID-19.
This article ( https://www.foxnews.com/health/man-coronavirus-china-whiskey-honey ) says he was hospitalized for two weeks at Zhongnan, which doesn't match any of his other stories.
I think this is more evidence he's making some of this up.
If nothing else this quote should throw Connor's credibility into question slightly: "I think to be fair the Chinese government have been quite transparent with what's going on."
You're correct that cats are susceptible to SARS-CoV-2 -- there's even a serological study of cats in Wuhan at the time of the outbreak that found they were infected. https://pubmed.ncbi.nlm.nih.gov/32867625/
However IMO there are a number of other reasons to question the significance of Connor Reed's story. For one thing, there's no actual verification that he was sick with COVID during that timeframe in the first place, which is why he was not included in the official cases in spite of all these media stories. Other people have claimed to be unconfirmed early cases such as athletes attending the Wuhan Military Games in October, but these are generally believed to be cases of misidentification.
In his story Connor Reed says he might have been infected at "the fish market" that was written about in the Western media, where bats and koalas were allegedly sold. This only describes one market in Wuhan, the Huanan Seafood Market, where news sites incorrectly reported the sale of koalas (it ended up being a mistranslation of another species).
https://www.nzherald.co.nz/world/food-market-at-centre-of-deadly-coronavirus-outbreak-admits-selling-live-koalas-snakes-rats-and-wolves/WMMILWGWLTKZVLSK5VK6W5H7G4/
https://mothership.sg/2020/01/wild-animal-wuhan-market/
So if we take him at his word, Connor Reed may actually be backing up the market origin, although it's strange that he seems to not know the name of the market itself. He apparently lives some distance from the Huanan Market. All the details in his story could be picked out of media reports.
There was a wave of influenza going through Wuhan at the time so it seems possible that he just had a bad case of flu and confused it with COVID-19 (although he claims to have been tested). It would be interesting to follow up with Reed's friends and contacts in Wuhan at the time and ask if they know exactly what happened.
Doesn’t a lot of Saar’s argument also rest on evidence that would be impacted by Chinese meddling?
The correct response to “Chinese meddling in the data” is to reduce the likelihood that we ever “know” the correct answer, rather than to just say “therefore Peter wrong”.
Recall that a wet market origin is also very embarrassing for China. The official CCP propaganda is that it came from some foreign source.
I think the primary evidence Saar has for the lab leak is independent of China, namely the DEFUSE grant, the furin cleavage site and the improbability of a coronavirus lab being just down the road from a global coronavirus outbreak by coincidence.
Which of Peter’s arguments on the Turin cleavage site rest on “China being truthful”? The participants largely seem to agree on what the site *is* but disagree on whether or not the site is more likely to be of natural evolution or a man-made insertion.
For DEFUSE, yes to some extent Peter is taking WIV at their word. But he also points out several reasons why it’s actually likely that they weren’t doing the right type of experiments on the right strain to cause the epidemic. And why if they were doing those experiments, it’s likely they would have said so (since before the epidemic they had no reason to cover up).
Meanwhile Saar’s theory *depends on* WIV / CCP lying about what they did after not getting the DEFUSE grant. And he doesn’t actually have any proof that they did, he’s just inferring that they did from his conclusion that the furin cleavage sites were probably man-made. His argument also relies on either WIV/CCP pre-emptively covering up their activities that eventually caused COVID (while being open about their other research and desire to be involved in gain-of-function studies) OR this research occurring at exactly the right time that it had gotten done but not published before the virus leaked and they went into cover-up mode.
Conditional on this being an unintentional lab leak, yes it’s very likely it would occur in Wuhan vs. somewhere else. Conditional on it being zoonotic, there’s not really any reason it would be in Wuhan vs some other Chinese city (but no particular reason it wouldn’t be in Wuhan). How much you weight you put on that “coincidence” will largely be swamped by how likely you consider zoonotic transmission and lab leaks in the first place (which is really hard to sus out, since both are rare).
We can be reasonably sure the WIV/CCP are lying about something, because they have stonewalled or misdirected all attempts at investigation, taken virus databases offline in the middle of the night, disappeared people involved and done various other highly suspicious things, some of which are known only via intel leaks from western intelligence agencies.
The debate here didn't touch on most of the evidence, but for example the US Government has claimed that they intercepted messages in November talking about a very severe situation at the WIV, and there is evidence they was a sharp uptick in activity there (hacked mobile phone records), and that they were rapidly acquiring "air incinerators" and other things. Air incinerators would imply they suspected or found a leak in the lab's air ducts or air handling equipment. It's been a while since I last looked at this but the weight of evidence for something having gone wrong at the lab is very high, unless you assume the leaks are all invalid (which would lead to a question of motive, as the establishment is very keen on zoonosis as a theory).
I do not buy the idea that it's unlikely they were doing such experiments:
1. It's very likely that the PLA were involved or at least monitoring the research going on there. All this stuff is highly tied up with military research, hence why DEFUSE was submitted to DARPA in the first place. Thus we can't assume everything would have been made public.
2. Peter argues the WIV would have been totally open and transparent before COVID, even though they became immediately 100% opaque afterwards and that has never changed since. If they're so innocent, why the sudden switch? More plausible: they have something to hide and thus were plausibly also previously hiding something.
3. I don't think WIV/CCP have ever commented on the DEFUSE grant or whether the WIV did or did not do the suggested research, did they?
4. The research occurring at the time of a leak and before it got published, assuming it was destined to be published at all, seems very plausible, no?
5. The "FCS is natural" argument makes some bold assumptions about the nature of whatever was being done, why and what could or could not have happened in a lab. I don't buy it. We know virologists love to play games with accelerating evolution, so what a geneticist might or might not have done to add an FCS doesn't seem very relevant to what WIV researchers would have done.
6. Lab leaks aren't particularly rare. They happen quite frequently, especially if you consider accidents that are bad but don't manage to start pandemics.
Again, your entire argument here is that there is a cover-up that is effective enough to make zoonotic origin plausible enough to convince a lot of smart people who’ve looked deeply into it, but not so good that Saar et al couldn’t prove it to be false. That’s a fine line! Meanwhile you expect Peter to prove a negative (CCP didn’t lie in precisely that way).
I would also reiterate: a wet market outbreak still looks very bad for China, and we can see that the Party line is that *neither* theory is true. Why would the CCP stage an elaborate cover up that still paints them in a bad light, vs. cooking up some evidence that it was from a Sri Lankan tourist or whatever?
I don’t have infinite time but I’ll address two points where I think you’re being uncharitable to Peter’s argument.
2. Peter is not saying the WIV is “perfectly transparent”. He is however saying they were open about their research, and until the leak of Covid would have had no reason to cover up specifically the line of research and exact strain that would later lead to an outbreak, because they would have had no way to know beforehand that that strain and that research was the one that would leak.
Unless you believe it was a deliberate bioweapon, in which case you’d sort of expect them to be completely opaque rather than “very open and then opaque only after a possible screw up”.
The WIV going full CYA mode strikes me as likely whether or not they were at fault. There was uncertainty, the people involved are paranoid, and they’d know they were obvious suspects.
5. I agree that it’s possible a human scientist *could have* made the FCS as we see it. However, Saar’s initial position was that it was *basically impossible* for the FCS to be natural. When Peter gave several reasons why that type of FCS could occur naturally, Saar shifted to “well okay maybe the researchers were deliberately trying to mimic or boost natural evolution”. That strikes me as dishonest argumentation and kind of conspiratorial - it’s the exact type of epicycle conspiracy theorists use to dismiss contrary evidence.
Has it convinced a lot of smart people? I have to admit, my biggest surprise when I started reading this article was that there was someone smart who was willing to argue against the lab leak. I know a lot of smart people and not one believes in it was a natural event. And as Scott shows at the start, the vast majority of all people don't believe that either. Peter is certainly being very counter-cultural. And if you want to argue that "smart people" are in the minority, well that's an appeal to authority that academics no longer have.
Putting that aside, I don't think Saar retreated at all. Let's assume both argue in good faith, and I think Peter did even though I think he's wrong. Remember that language is very imprecise. This imprecision has already created a ton of confusion on this topic for several years already. If the WIV do something to speed up the natural rate of evolution, is that "natural"? Is it "engineering"? That's entirely subjective and you could easily argue for both words. Peter here starts by assuming that not natural means genetically engineered and then starts arguing against it being genetically engineered, which isn't a correct assumption given the nature of the WIV's investigation. They were interested in what a coronavirus pandemic might look like. They wanted to look into the future (at least, officially). Just directly engineering an FCS into a pre-existing virus can't possibly tell them that, so why would they have done so? Rapid forced evolution in humanized organisms on the other hand are a much closer way to "speed up time", and could easily have led to the evolution of a strange FCS under lab conditions. If they didn't expect it to happen because it's never known to have happened in nature before then they would not have been taking the necessary precautions, as indeed we know they were not. And bang, that's all it takes to start a pandemic for real.
So Saar didn't shift at all, in my perspective. He corrected a misunderstanding by Peter of what was being actually argued, an honest misunderstanding coming from the vagueness of terms like "natural" when talking about serial passaging experiments. But it does invalidate most of Peter's arguments about the exact genetic code of the FCS.
You are missing some important context which is that many of the allegations you mentioned (about a "severe situation" at WIV for instance) seem to be exaggerated or made up.
https://fallows.substack.com/p/on-that-propublica-chinese-lab-leak
The "severe situation" story by Katherine Eban and Propublica seems to be based on a bad translation, with China experts questioning it and Propublica eventually issuing an update on the article itself. And there are a number of stories about the lab like this that end up being the result of bad reporting, misinterpreted evidence etc.
When you look into who is behind these stories many of them seem to be coming from the Trump State Department under Mike Pompeo, which conducted an investigation into the lab back in 2020. "State Department investigators" is the source given for many of the key lab leak stories, such as the claims about sick scientists with COVID symptoms published by Michael Shellenberger and the Wall Street Journal or the Sunday Times story including the claim that a Chinese scientist might have been executed by being thrown off the roof of the lab.
There's actually an account by another State Department official who was involved in the investigation about what happened:
https://christopherashleyford.medium.com/the-lab-leak-inquiry-at-the-state-department-96973cff3a65
He says he was open to the lab leak but that certain officials like David Asher and Miles Yu became very fanatical about it. According to him they were already believing unhinged conspiracy theories like the virus being a racially targeted bioweapon and wanted to charge China under the Biological Weapons Convention, which he thought was a bad idea. Eventually they ended up accusing him of being part of the conspiracy and trying to cover up their work. The same people, along with Trump, Pompeo and Richard Dearlove, the former head of MI6, appeared in Sherri Markson's documentary "What Really Happened in Wuhan."
There's some interesting background to all this which is that both Miles Yu and Mike Pompeo have connections to the New Federal State of China, a weird regime change group that says it wants to overthrow the Chinese government. This group was the first to promote the lab leak theory in a major way starting all the way back in mid January 2020.
https://twitter.com/wanderer_jasnah/status/1631438683688648705
https://mediamanipulation.org/case-studies/cloaked-science-yan-reports
https://www.nytimes.com/2020/11/20/business/media/steve-bannon-china.html
Mike Pompeo received an award from the group and Miles Yu has spoken glowingly of its founder Guo Wengui, who is now facing trial for defrauding the overseas Chinese community out of $1 billion, in speeches he's given. So it seems quite possible this group and others in the anti-CCP Chinese diaspora are influencing the lab leak narrative to some extent.
https://twitter.com/dfriedman33/status/1631346935905886224
https://twitter.com/mrspanstreppon/status/1745855771005317279
https://twitter.com/gujingc/status/1707281401857663145
Only problem is that when the ODNI issued a report on the lab last year, it seemed to greatly downplay the significance of many of these allegations. Turns out there was no specific biosafety incident known about, the people at the lab who got sick probably just had normal seasonal illness and there was no indication they were the ones working on coronavirus research (discussed towards the end of this article).
https://archive.is/BEwU2
All of this therefore needs to obviously be taken with a lot of skepticism. There are geopolitical motives, propaganda and BS stories on both sides unfortunately, not just from the Chinese government.
You have arbitrarily excluded the third possibility that your friend is simply mistaken. His conclusion remains extremely unlikely regardless of (your judgment of) his personal honesty.
I was being facetious. If Peter is allowed to decide large numbers of random people are all wrong and liars to boot, I'm allowed to assume my friend is correct. You might claim that no, the other evidence means I shouldn't but that's circular
As I've been saying since the whole debate started, whether or not it was a lab leak is really irrelevant. What we know for certain, with no actual controversy surrounding it, is far more interesting: China did not shut down international air travel out of Wuhan until significantly after they knew the virus was circulating.
While they were covering up the origins of the virus, and locking down Wuhan to try to keep it from spreading further internally, the CCP was knowingly, deliberately exporting this deadly plague to the rest of the world. And that is far more damning than any question of its origin. The origin of the virus is China. It deliberately inflicted it upon the rest of the world, which makes the pandemic a *de facto* bioweapon attack, whether or not the virus was developed in a lab with the intention of building a bioweapon.
Closing down airspace was considered a pretty radical thing to do back then - it’s only in retrospect that it seems obvious. Had China stopped Americans leaving, for instance, there would have been an outcry.
And of course flights from China could have been banned at entry, but the west was clowning around with either calling that racist or denying that that whole thing was fake - which I believe was the position of Fox News and Trump for months.
I have a memory of Trump barring travel from china fairly early, while also saying it's probably fake, while also calling it the China Virus, and while also saying that it'll blow over soon. This is consistent with my model of Trump saying lots of things for various reasons, none of which are primarily anything like "provide parsimoniously accurate information". I'm open to seeing a well-cited timeline of his actions and statements, but I don't think any of us care enough to create or find such a thing.
The Chinese travel ban occurred in late January of 2020 after an increasing number of cases were discovered in the US. In mid march San Franscisco and the Bay area imposed the first shelter in place order (although behaviors were changing in the bay area for the prior two weeks and on March 20th NY declared a stay at home order (effective march 22) https://en.wikipedia.org/wiki/Timeline_of_the_COVID-19_pandemic_in_the_United_States_(2020). One thing is that the behavior in early 2020 involved a series of relatively (within the lifetime of most people) unprecedented steps and while it is easy in hindsight to argue about what should have been done, neither China nor the US really understood the consequences and in particular were constantly make decisions that were days to weeks behind where they should have been.
Yeah I personally have no blame for any decision-makers. That itself seems an unpopular opinion, but it's just incentives playing out. What's the age of the electorate? How scared of dying are first-world elderly? Boom, tells you everything you need to know.
I agree that it's unimportant per se. I'd argue that the other specifics throughout were also unimportant: masks, vaccine efficacy, vaccine safety, lockdown efficacy, ivermectin efficacy, vitamin d efficacy, etc. None of the mechanisms seemed to matter. What mattered more was "are people lying to us?" and the answer to that was "yes, unambiguously" (on both sides of each and every one of these debates).
The consensus prior to 2020 was that shutting borders would only serve to delay respiratory disease outbreaks by ~2-4 weeks because, under exponential growth, blocking 99% or even 99.99% of travel still means its not long until covid hops the border. For China to have acted maliciously in not closing the border would require that 2021 values and beliefs travelled back in time to 2019.
Then why did they lock down travel internally?
Locking down internal travel is just something communist regimes do. In China it is part of the Hukou system. In these sorts of regimes, internal travel restrictions are historically the norm, and permitting internal travel is the aberration. Therefore China reinforcing internal travel restrictions is far less interesting than it sounds.
They wanted to delay spread by 2-4 weeks.
They should have locked down the whole country at that time but that would have been too extreme in 2019 so they did an extreme but less extreme thing of locking down the city but of course it had already spread to all of China by that point so the lockdown was only moderately effective. Again everyone was learning and reacting in early 2019/2020 and typically they were making decisions that were weeks late due to not thinking about exponential curves.
If SARS-Cov-2 did not come from WIV then presumably WIV and the Chinese government would have had a very powerful motivation to open their records to outside investigators and make all of their people available for interviews. Instead they clammed up. Why?
The problem is that the Chinese government is also unhappy with zoonosis, because it's still somehow embarrassing. The narrative pushed in China now seems to be that the virus did not originate in China but was brought in to Wuhan in frozen food. This actually happened several times after China's zero-covid strategy started so it's a convenient theory to push the blame on.
Ok, but that still leaves them with a strong incentive to open the WIV records if they are consistent with the argument that WIV was not the origin of the virus.
At this point in time, anyway, would anyone believe opened WIV records that apparently exonerated them? The CCP would have had ample time to cook the books. ( I doubt that I would trust _any_ government's claims in an analogous situation. )
What do you mean, "somehow"? People have constantly pointed out that wet markets were a blatant health risk, and yet the CCP never never took action against them. If the wet markets were entirely responsible for COVID, the CCP would indirectly be responsible for it through their inaction. That would be bad for the CCP.
All the usual reasons that most virology institutes don't generally open up their records to outside investigators. Plus possibly whatever not-totally-kosher things they might have been doing there.
These kinds of arguments were used to devastating effect in the run up to the 2003 Iraq War. Saddam Hussein refused to allow inspectors to see his chemical/biological/nuclear weapons facilities. The only plausible reason for this was that he had secret weapons that he wanted to hide from the West.
As it turned out, he did not have secret weapons.
Instead, he wanted to appear like he had secret weapons to scare off Iran, a large hostile neighbor which he had fought with repeatedly in the past. A totally rational reason to want to appear to have secret weapons of mass destruction which did not occur to the American government before they made the fateful decision to invade.
Trying to ascertain the intentions of your enemies is something humans are particularly bad at. I wouldn't use it as a basis for making decisions, if it can be avoided.
It cannot be avoided.
...if you could avoid ascerting the intentions of other people (not only your enemies) in your everyday life (rather than only believing you do not have to do this), you could not walk down any street without being on constant alert and armed to the teeth.
...Because then you could not ascertain if the passers-by are unlikely to have the intention to suddenly and unprovokedly attack you.
Ascerting the intentions of other people, at least in the limited sense of being reasonably sure they have not the intention to suddenly attack you, is nothing less than THE fundamental feature that makes in possible for humans to live in societies consisting also of non-kin. It is the founding stone of sociology.
The Chinese government's default reaction is to cover up anything remotely controversial, even for stupid or irrational reasons, so failing to open one of their high-profile labs to foreign investigators is not really that surprising. They did allow the WHO to go there and interview the scientists for several hours about their work, and Zhengli Shi gave an interview to Science where she answered questions about specific lab leak allegations. From China's POV this probably represents more openness than usual (by comparison, the WHO did not get the chance to talk to animal traders at the market).
But ultimately the CCP cares much more about domestic public opinion than international. And within China itself, the widespread belief is that the virus came from overseas, possibly from a US lab.
I hadn't heard of Rootclaim before this, but that first screenshot which puts the odds of widespread fraud in the 2020 election at 8.7% tells me all I need to know -- that is orders of magnitude too high as anyone who understands how us elections are run would understand. Seems like there's something about this method, at least as practiced by Saar, that produces excessive probabilities of out of consensus or malicious explanations for things.
This is a known bias we have. We have several safeguards against reaching extreme probabilities, and it often causes us to overestimate low probabilities. We have improved this mechanism since, and would probably get a lower number for election fraud today.
I'm sure this is not a new argument to you, but if you admittedly exaggerate a say .0005 chance to .05 why not .05 to .5? And a number of exaggerated small risks will add up to a very high probability quite quickly.
because this mechanism specifically pushes posterior probabilities to the middle
On the subject of this showing up baysian analysis, since viewers all tried to track it and got different results anyway:
Quant trading also works like this - it's basically using baysian inference for the stock market - and also ends up with a lot of big discrepancies and requiring a lot of Metis to get right. But it's still consistently significantly outperformed (in terms of real money earned in the real market) over traditional non-quanty trading. To the degree that baysian analysis does work in real life, this is what I'd expect it to look like.
One thing that I'm confused about is that I often see the following two claims being made:
1. Quant traders / financial people in general are one of the most calibrated and truth tracking professions.
2. Basically all traders, after accounting for fees, risk, and long run performance, underperform the index.
Which seems contradictory. However, there are obvious ways in which these claims can be both true.
1. We're talking about different populations
A. People talking online are highly selected by the competence of their peers, so all the anecdotes online are selecting for say, the top1% of traders.
B. There are two criteria for success, marketing and actual financial talent. Funds that are mostly good at selling a story to retail investors "drag down" the average, but remain solvent because their customers do not understand "underperforming the index". Calibrated people come from the latter, or produce models for the former that are then organizationally neutered in some way.
C. This is just talking about two time periods. Kahneman released his analysis something like 20 years ago, and most of what I read about this was from that timeframe. 20 years is a very long time for a bunch of intelligent people to incorporate new information and develop good norms around them.
2. The second situation is that one of the premises above is flawed.
A. Individual models do perform better or worse, but they do so for significantly short periods of time, or the models themselves are calibrated, but the financial EVs aren't, so you end up with something like a black schols situation, where most calibrated models do well, then someone accidentally has a programming error flipping a sign, or something which wipes out a lot of value, and these unlikely incidents form a balancing factor for performance. So calibration does not necessarily track financial performance.
D. The "fees" part of the "outperforming the index fund" is where all the magic happens: the individual trader or firm wins by having good models, but they aren't good enough for the retail investor. So calibration does track financial performance, just not enough.
Hmm, maybe ai should talk to finance people more.
"All traders", taken literally, are the index, and therefore must underperform due to various frictions.
That's not it, I don't think, since the index is a specific algorithm weighting everything and "all traders" here specifically mean "like 200 firms that Kahneman looked at in Wall Street" or something like that.
I'm not sure if that's definitionally true or false, it is the case that every trade has a buyer and a seller, but I have an intuition that at least some movement of the stock market reflects "real underlying wealth" whatever that means, and that it would be possible to have economic growth such that indexes in fact do not perform well. Tricky, tricky, I feel like my intuitive model is wrong in some way, but I don't understand what.
If Kahneman looked at a representative sample (and the guy was meticulous, so I expect him to had done that), then their performance should closely match that of "all traders". The sense in which "all traders are the markets" is kind of definitionally true - "the market" is just the sum of all trading transactions. So the "market going up" means transactions overtime happen at higher and higher prices - the reasons for the uplift are not important for this specific understanding. Once we know the traders make up the market it becomes clear that they - collectively - underperform by the amount of "friction", i.e., costs of transactions.
There's a subtle but important complication though: what we call "the market" usually refers to some kind of a market index, e.g., Wilshire 5000 or some such. These indexes attempt to represent the market, but are not "the market", and therefore have their own deviations, or "tracking errors". So in reality, it is possible for anomalies to occur where, for a period of time, it would look like "all traders" outperform "the market". When in reality what this means is "a representative sample of traders" outperformed "a market index".
No, Kahneman was specifically looking at the trading activity of Wall Street firms, which isn't representative of all trades. Even then I don't think this would be definitionally true: for example, you could have one super sucker firm, who trades in the maximally ignorant way, and has deep pockets such that they don't stop trading. You'd get the result that most firms do outperform the market. I think you have to do the hard work of actually understanding how both statements came to be, sorry.
Oh, sure, "most" is not "all", it's entirely possible for there to be outliers with outsized impact. In fact, we know this has happened - most famously in 1998 with the LTCM collapse.
Another piece is that hedge funds are generally more consistent than index funds (even if they match them on over), and for a lot of people uncorrelated returns are more valuable (see e.g. herre https://capitalgains.thediff.co/p/naivecomparisons )
Also note that Quant traders vs finance people in general aren't exactly the same groups (the quant heavy places have generally done better over the last two decades or so)
Yeah, this looks like the "Micaiah is unreasonably conflating two groups, quants and non quants into one, and people are predictably talking about quants when he hears about it online" and "outperforming index, isn't the only way to be truth seeking, turns out having a model that doesn't rely in default economic growth and still earns money has to be more accurate since you don't have that as a crutch".
Feeling pretty not confused now. Whooooo.
There's one more potential point of explanation: exploitable mispricings are finite. Trading itself moves the market, so if you have a better model than the market consensus the absolute amount of money you can gain from trading upon it is still quite limited. Large funds – ones open to new investors and large enough to move the statistics of "beating the market" – are too large to meaningfully exploit these opportunities.
Hence, you see hedge funds and other large investment funds (pensions, for example) look at less liquid markets like private companies and privately-placed loans. In these markets, there's less public trading and thus less market consensus (thus more potential for mispricing) and without public quotes individual trading is less likely to move the average price for everyone else. If I'm a hedge fund and I know I can give some set of private companies loans at 0.1% less than my competitors, issuing that first loan isn't an immediate, obvious signal to my competitors that they have something wrong.
This doesn't roll off the tongue at all when I reply to "If you're so rational why ain't ya rich"!!
Thanks. That might also explain another flaw in my thinking about the study of below market index returns: if the firm's base rate isn't market returns but some different distribution, yeah, sure it's technically bad to give them your money, but that doesn't mean they're inaccurate. They're playing a potentially harder game.
"1/10,000 Wuhan citizens work at the wet market. So if a lab leak was going to show up somewhere random, the wet market was a 1/10,000 chance."
Why are we only counting the people who *worked* there? Wouldn't the relevant number be all the people who visited there (a much higher proportion of the population)?
The exact claim was about the index case being a market worker. That's generally OK. The main mistake is underestimating all the factors that would make that happen even if there was no spillover there.
First image: "The election was no different from previous elections" is not inconsistent with the claim that there was widespread fraud. As stated, none of the options are mutually exclusive.
(this is true regardless of the actual answer to the question)
I disagree with some of the arguments. I think COVID originally started spreading in rural villages. It didn't spread fast, because there wasn't that much unventilated space, and it wasn't well adapted to humans. And it wasn't detected, because until special tests were developed, COVID just looked like "atypical pneumonia". That is where it adapted to humans. Then it was carried into the wet market by a vendor. So it did spend some time learning how to spread in humans, but it wasn't by gain-of-function research.
How does this theory address the fact that the nearest comparable virus was from a cave 1,000 miles from Wuhan?
It doesn't directly address it. But if I am correct, our knowledge of what viruses are circulating where is quite patchy and incomplete. I've heard that in some cases of xenonotic diseases it's taken multiple decades to track down the "original source"...with nobody trying to make excuses or prove that it isn't their fault.
Also, if the virus was originally a bat virus, bats fly, and some are migratory. So perhaps they just haven't been looking in the proper caves to find a closer example.
FWIW, I don't think the evidence is sufficient to invest ANY of the theories assigning blame with much belief. People can't even agree on what was the most probable non-human host. (Why not some weasel...we know it spread well among mink.)
The main challenge to this idea would probably be the fact that the pandemic seems to be descended from the strain of the virus that emerged in Wuhan, with scientists arguing it traces back in terms of timing to late November, right around the time the Huanan Market outbreak would have begun. This is why the scenario of spillover at the market seems to make sense, especially given there were multiple potential host species sold and slaughtered there. As it happens one of the stores at the market had a website that was archived on the Wayback Machine and you can get an idea of the range of species that were being sold:
https://web.archive.org/web/20190707014745/http://whdaz.com/
Some of the animals were being imported from the part of China where similar viruses circulate in bats, while others are the same species found infected during SARS-CoV-1.
But all of that could be wrong of course. Apparently there was a wave of influenza spreading through Hubei province right before the pandemic began so you could speculate that this masked smaller-scale rural outbreaks. There's never been retrospective case-finding outside of Wuhan that's been publicly released. There's a report in a Hubei paper that markets in Enshi, where civets and raccoon dogs sold in Wuhan were raised, were shut down on December 23, 2019 -- a week before the Huanan Market in Wuhan was closed.
https://www.independent.co.uk/asia/china/covid-coronavirus-bats-caves-hubei-b1940443.html
China has also failed to release testing of blood banks outside of Wuhan, and testing of the animal trade in general has been limited. Officially the government still claims that zero animals within their borders have tested positive at any point during the pandemic and there have been no outbreaks on fur farms in spite of tens of millions of susceptible species being raised there. So it's difficult to draw conclusions but the genetic history seems to point to the outbreak in Wuhan being caused by spillover to humans either at the market itself, or at a closely linked location such as a farm that supplied the market.
Is 1000 miles nearer or farther than you expect for a virus that is separated by several years or decades of evolution from the virus we're talking about?
According to Saar's argument, which Peter did not refute, the nearest example we've found was located 1,000 miles away. If there are no intermediaries between that sample and what we later saw in COVID, then that seems to increase the chances that this was a modified strain. If there are intermediaries, then we would need to know where those were located in order to determine how their descendants ended up in Wuhan.
Whichever debate participant is most well-versed is most likely to have thoroughly understood the topic and thus is most likely to have a correct intuition on the subject. Thus, we should just decide on these issues based on our subjective intuition about which debater seems to be most knowledgeable and rational, and not reason through the content of the debate except to fact-check the debaters and ensure their reasoning follows so they can't get away with intelligent-sounding bullshit. For example, instead of checking every single fact, just check a simple random sample of them. This approach may seem objectionable to some people, but it's probably what we're all subconsciously doing anyways, and it's probably approximately correct.
>Whichever debate participant is most well-versed is most likely to have thoroughly understood the topic and thus is most likely to have a correct intuition on the subject.
I just don't think this is true. "Crank that goes down an obsessive rabbit hole researching conspiracies" is an archetype that will tend to end up (a) having a lot of domain knowledge, and (b) being completely incorrect.
Like, I'm pretty sure if you took a flat-earth nut and put them in a debate against a random person chosen off the street, the flat-Earther is going to be far more well-versed. I've spoken to flat-Earthers! They love getting into minutia about discrepancies in NASA image processing, and flaws with Newtonian physics, and Ancient Greek historiography, and so on. The average person on the street is going to be a lot less knowledgeable on all of these things!
Being "well-versed" just selects for obsessiveness, not correctness.
That was exactly the example I was going to come up with. I do enjoy learning from flat-earthers, because they dig up interesting anomalies, even though I think they've fundamentally learned the wrong lesson from them.
(To be fair, the same is probably true for most real experts for at least some of the things that they have real expertise on.)
> this disease that came out of nowhere and ruined all of our lives for a few years
I wish we could have a debate on what the proper response to COVID should have been. In theory, if governments refused to do much other than developing vaccines, COVID would've had a significantly smaller impact on our lives. But also this would've also had significant costs to many people in the form of losing their elderly relatives and more people would've had long COVID.
I wish we could have a Rootclaim debate on what the Western countries should've done in late February/early March as a response to the virus. I think there's a large degree of disagreement about this question to this very day and it's definitely not been discussed enough in retrospective.
Asking the real questions! Lab leak vs zoonosis strikes me as mostly a culture war question more than anything meaningful, and serves a distraction from these brass tacks of: okay, whatever, what do we do *next time*? I fear we'll repeat the 2020 psychosis instead of having a better response.
I recognize some people are working on this- biosecurity has become a larger EA cause, as one example- but I suspect that the "people at the top" will be just as ineffectual, inattentive, and political as last time.
With regards to the culture war aspect, I collected a (possibly relevant) list of headlines on the lab leak hypothesis over time. My intent was to try to see how normal people might see the news over time on the lab leak hypothesis so I pretty much just focused on headlines and restricted it to fairly mainstream sources (I'm sure it would look different if I included, for example, Daily Mail).
If we see this sort of thing the "next time" then I expect we will see similar belief splits.
Part of the problem is that I don't think most people are very good at treating headlines as "this might be true ... but possibly not" instead seeing things as "the experts have changed their minds ... again"
http://mistybeach.com/mark/LabLeak.html
I don't have any concrete suggestions.
There isn't any clear correlation between governments doing stuff and less covid happening. Sweden is the obvious go-to as a country that did little and got same or better results than it's peers, but Peru is also an example of a country that did a lot and got an unparalleled catastrophe for it.
Sweden got much worse results than Norway for example. I think this comparison is more informative than with countries such as France/UK/Italy
Why exclusively compare Sweden with another country that had relatively lax restrictions? Regardless, middling short term results and brilliant long-term results (lowest or near-lowest excess deaths in Europe) for the one country that notably rejected all but the most modest of restrictions is hardly a ringing endorsement of restrictions doing anything.
Because they both share similar climate, culture, and density of population. I'd assume there's also no big difference in term of popularity as a touristic destination between the two. Norway had a stricter policy than Sweden, schools and several business were closed. This was not the case in Sweden.
How confident are we that climate, culture and population density are even relevant? Relevant enough that we should ignore Peru's terrible results with restrictions, and Sweden's good results with restrictions, because one neighbour of Sweden slightly outperformed it in the short term but not in the long term? Does the Scandinavian Taiga perform spooky action at a distance by affecting transmission in Stockholm?
Stringency of restrictions and covid deaths looks like a scatter plot with no correlation. This is relevant, and can't be excused away with culture.
Regardless, this is a post-hoc epicycle. When lockdowns were brought in, they were advertised as necessary everywhere to avert megadeaths. Not that some countries didn't have to do them because culture or weather. Not that rural areas didn't need to because population density. And not that the outcome would instead be going from Low Norway deaths to Middling Sweden deaths. If you believe that certain population densities, cultures and weathers mean you don't need to lock down, you're already 90% the way to opposing lockdowns, you just differ in the universality of opposing them.
I agree Norway had a stricter policy than Sweden. It was, however, still more lax than Europe.
Awesome analysis. And really good that you picked up on the personality factor at the end. Kinda odd that Rootclaim do deep Bayesian stuff on the arguments, but do no personality screening for willingness to be objective / fear of losing a debate. The adversarial system is super-Western, I guess, with two dudes fighting their corners,. But in this instance it seems to me like it could be huge factor to swing things.
The comment at the beginning suggesting that it was too bad Putin doesn't have cancer felt a little odd to me, BTW.
The snide comment about Putin also grated with me. But I remember Scott's thing about Bin Laden and Thatcher - if Putin was assassinated (for e.g. by the CIA) what would my reaction be?
I think the human tendency to try to anthropomorphise our way out of problems can be an issue. Bad guy here, bad guy here...
You know, I looked at the "widespread fraud" one at the top, and it struck me that it didn't even consider that I would think was a much more likely hypothesis for it: That several big city political machines, acting independently and in parallel, jumped in an applied diverse methods of falsifying the outcomes, in just those places where such falsification was both (a) readily manageable and (b) critical to their respective states' electoral votes (giving motivation). It makes me suspect that you can get the results you want from Bayesian methods by carefully curating the initial hypotheses to compare. Not very different, perhaps, from C.S. Lewis's trichotomy about Jesus.
Hot take: Peter clearly failed to convince anyone.
The lab leak odds, in log10 (i.e. orders of magnitude are):
Peter -20.7
Saar 2.7
Eric -3.1
Will -2.5
Scott -1.2
Daniel -1.4
One of these numbers is clearly an outlier. Scott mentions it and calls it "trolling", I would argue that it is debating in bad faith. 2e-21 is a ratio which is just silly. For one thing, the gain of function at WiV pathway is not the only pathway towards a lab leak. The WIV could also have released a naturally occurring coronavirus at the wet market. At 2e-21 odds, we would probably have to consider the possibility that the WIV built a time machine and went back in time to infect the wet market.
Apart from the pangolin which got retconned out in favor of the raccoon dog, one thing I remember from the earlier discussions was that proponents of the lab leak pointed out that the bat with the virus likely did not come from Wuhan. Given that Saar has not made that argument, is that no longer the case? I imagine that by now we have a much better understanding of naturally occurring bat corona viruses in China.
Scott included several arguments where both sides agree that the source virus was from bats in a different region than Wuhan.
Hi Scott,
thank you for mentioning my "preachy and annoying" blog. :P
Can't do much about my eyes rolling backward on this topic, but since you are all about evidence and probabilities; you might want to keep your eyes peeled for my next 15000-word sci comm article that comes out in two weeks and will bring people to the cutting edge of origin science, as well as put the "gain-of-function research" narrative to bed once and for all.
Promised :)
Cheers
>preachy, annoying
I just read your piece. He's not wrong. If your goal is a broader audience you may want to reflect on it.
There are plenty of people who love tone like this. He's solidly in the middle of his lane, and will do fine.
Sure, the lane is a circlejerk of signaling and dopamine hits from feeling superior. No real persuasion or work is being accomplished.
Hihi, opinions differ, so do goals.
Maybe my goal is to accurately reflect reality, educate, and have fun writing, rather than catering to audience demand by constantly upholding the false equivalency spiel of influencers that care more about popularity than science :)
Scotts friendly bothsidism here towards a fanatic multimillionaire is what makes me cringe... guess that is what influencers do to be popular with the masses... "don't make enemies of powerful people, don't rock the boat, give your audience what it wants from you, everybody's opinion - no matter how unfounded - has value and must be taken seriously, isn't it that both sides have good points... maybe they might not be entirely correct, but those pesky Zoonati scientists certainly have their fair share of problems. too.." *yawn*
Seen it a hundred times, but a spade remains a spade no matter how many charitable hot takes are written about how it can also be seen as a fork
I think your writing is good and your videos are as well.
But I also think that the tone would put off anyone who's on the fence about the issue, because you lump together lab leak believers with other conspiracy theorists, and it makes it sounds like everyone who's taken in by the theory is just an idiot.
But support for lab leak runs around 60%, and plenty of smart people believe it. They aren't all anti-vaxxers or flat Earthers of UFO cultists.
I got taken in by it for a while, at least enough to say it was 50/50. You even did, yourself, for a briefer time (like 1 month in 2021, or something?)
I will admit that it's a struggle to write an article that simultaneously preserves the ego of the person who fell for a bad theory and also debunks that theory effectively. So you have to pick some approach there, and yours is just different than mine.
Maybe I should have a disclaimer somewhere:
"People often frown when I call the currently advanced "man-made" speculations conspiracy myths, because "how can you stigmatize 2/3 people as conspiracy theorists?". So let me quickly clarify: Conspiracy myth is an accurate, functional term for these sets of ideas and attached worldviews. However, believing a lab leak is likely, or having fallen for one or the other false myth does not turn ordinary citizens into conspiracy theorists, nor does it make them gullible or stupid, just incorrect. We all hold wrong, uninformed, ignorant, or shortsighted beliefs on many issues, make mistakes, or just get unlucky with our information diet. Yet we are usually able to update our beliefs and correct course by ourselves when such opportunities present themselves, while conspiracy theorists usually can not. The term "conspiracy theorist" should - in my opinion - be reserved for the accurate description of participants and activists in closely-knit conspiratorial communities that are hyper-engaged and trapped in these all-encompassing worldviews, even victims of it. They need help and deprogramming; they do not deserve our attention, but also not our scorn. All my ire goes towards those who wield and amplify conspiracy myths as a tool for popularity, persuasion, profit, or power. Self-serving influencers, populist commentators, unethical journalists, calculating politicians, fanatic activists, power brokers, media manipulators, conflict entrepreneurs, contrarian grifters, and their ilk would probably fall into that group. What they have in common is that they are a type of social elite with an outsized impact on the discourse that they poison for everybody else."
Yeah, it's a hard balance to strike. I loathe a few of the people that professionally lie and manipulate the public on this issue and a few others (like vaccines or climate change).
But I have no ill will towards the average person that just thinks it's a lab leak because it's all really confusing and there are more newspaper articles about lab leak than about zoonosis. To them, it's more like:
"I've been there! I was there! I'm trying to save you a year of your life so you don't have to spend as much time as I did, sorting it all out."
Yeah, no disagreement here.
That, in its entirety, would be a better title for your article.
In the expert poll, one fifth of respondents thought a lab leak more likely. The median expert puts it at 10%. This is, by definition, not a conspiracy theory
>However, believing a lab leak is likely, or having fallen for one or the other false myth does not turn ordinary citizens into conspiracy theorists, nor does it make them gullible or stupid, just incorrect.
I agree. So why take such a derisive and condescending tone? There is real work to be done to dispel the lab leak theory. You seem to be more interested in signaling than actually changing people's minds. Because your writing is laundry list of things not to do when trying to persuade.
I had to ctrl-f to find his mention of you after stumbling over your comment here, which stood out visually due to the annoying use of emoticons.
Does your blog contain a post about how you so fluently acquired journalism-speak? Did you really start out in 'science' without a journalism degree? Can you really write 15,000 words in that breathless, snarky tone on a regular basis? While you'll clearly never write for the NYT, this is more likely due to discrimination than lack of ability. You seem lke you'd be a great fit.
Is that a backhanded compliment or a diss? xD
FWIW I read it as a compliment.
Yeah, I am about 50/50 on this. Maybe we should set up a rootclaim debate to figure it out
Mostly a compliment. You're in a comment section famously inimical to the NYT, but we'd all agree there's a hierarchy to journos. That is, while they're all bad for society, only most of them are bad at their jobs. Some are talented and competent. The dig is the implication that if it weren't for anti-white, anti-male bias you'd already be a big deal at the NYT. I have no idea if I'm joking when I say that part.
*edit: were[n't]
Separately, I'm bothering to write the compliment because I work hard at good mental hygiene. I hope never to lose the ability to give credit where it is due. The annoying emoticons and snarky tone probably activated some prey-drive-adjacent subroutine in my brain that looks for easy dunks, but you don't provide any.
I enjoyed your blog post and am grateful you chose to cover this very interesting event. I drew from your post when writing mine and hope I gave you enough credit.
But yeah, I'm pretty jaded about the skeptic community, for reasons I've talked about at https://www.astralcodexten.com/p/contra-kavanaugh-on-fideism and https://slatestarcodex.com/2014/04/15/the-cowpox-of-doubt/, and fairly or not your blog post reminded me of them.
I don't know if you're joking when you say you'll put the gain-of-function research narrative to bed once and for all, but I look forward to reading what you've got to say.
Tongue-in-cheek of course, my whole point is that the lab leak theory is dead, but the lableak myth will never die for factors that go beyond scientific uncertainty.
No matter what I write, media manipulators will not stop pushing a false narrative.
However, for those with an open mind and a real interest, yes, the next article will reduce scientific uncertainty to a point where no "gain-of-function" scenario remains possible.
Looking forward to reading your comment then!
cheers
Huh, for me the real "gain-of-function narrative" is that regardless of whether it's responsible for COVID in particular, it might very well become responsible for the next pandemic, possibly an even worse one, and whatever dubious benefits this research brings can't pass any remotely reasonable cost-benefit analysis. I'd be much more interested to see a discussion of that.
Will be extensively covered.
But tl:dr is: GoF is not a precise term, but has both risks as well as benefits that are outsized. Way forward is that cost:benefit needs to be assessed based on individual experiments and by experts who understand what they are talking about.
Hi Scott, here is as promised, the investigation that will put the "gain-of-function virus" narrative to bed. Enjoy and let me know what you think. https://www.protagonist-science.com/p/treacherous-ancestry
> Fourth, for the first time it made me see the coronavirus as one of God’s biggest and funniest jokes.
The punchline is "hindsight is always 2020"
Grin, thanks.
I was pretty sure about lab leak, now I'm 50/50. Normally I should be frustrated to end up 50/50 after reading all this but I'm filled with a feeling of satisfaction. Great job for all involved.
Similar for me. This nudged me from 85% lab leak to maybe 60%. I do wish the policy conclusions section included some reiteration of the lesson that clearly we need to clamp down on GoF research and generally need a ton of better, independent oversight of labs handling/creating dangerous pathogens.
I went from 85% lab leak to 80% natural origin. The virology stuff seems pretty convincing.
His talk about the databases doesn’t seem to line up with what I remember reading from Alina Chan however. I’d be interested to see them debate.
I am WAY out over my skis here…but it just seems like there is no objective way to actually TEST any of your priors in a scenario like this. So you’re just giving a finite number (and “false” sense of precision) to what actually remains a messy and murky quandary.
Priors aren't the sort of thing that can be tested. Scott thinks that some priors are better than others, and that there's an objectively best prior out there, but I don't think anyone has made any better progress on this than "just assume everything is equally likely to begin with, and don't worry too much about what "everything" means". (Maybe that was a bit too snarky.)
But that’s what I mean. These guys calculated posterior probabilities to 6 or 7 decimal places. That’s an incredible amount of precision to be claiming when your starting point is an unquantified, unverified, and seemingly unverifiable prior.
So this debate brought out a lot of facts and information I wasn’t previously aware of. And it was presented by both sides incredibly well. This would be close to a best-case example of the concept, i would think. And it appears “zoonosis” “won”. Yet we are in fact still no closer to knowing who is right. That’s a fairly circuitous 15 + hr journey for that kind of payoff.
It depends on what you mean by "right". If you mean "who supports the theory that is true", I think we are closer to knowing that. If you mean "whose degree of belief is closer to the objectively correct degree of belief", well I don't think that exists.
I agree (to the minimal extent I understand this stuff). It seems, as you say, that there is no way to test how accurate your prior (or your belief) actually is.
And when I say “right”, I do mean “true” or “objectively correct”.
But this is what I’m referring to: without being able to say your prior (or starting belief) is correct (or how objectively relatively true your starting belief might be), how confident can you be that any reasoning predicated on that starting point will lead you to the correct place/conclusion? How confident can you be that your perfectly constructed and engineered thing will remain standing, when you’ve built it on sand?
If the conclusion is a probability as well, then there’s again nothing it means for that to be “right” or “correct”.
What we do have a guarantee of is that if something is true, then pieces of evidence that increase your probability for it will be more plentiful, and if it is false, then pieces of evidence that decrease your probability for it will be more plentiful. No matter what prior you start with, an investigation that continues finding new pieces of relevant evidence will in the limit converge to 100% if the claim is true and to 0% if it is false.
We can’t put bounds on that to guarantee that at any particular finite point in the investigation you will be particularly close to the truth, but we can guarantee that in the limit you will.
To ask for something much stronger than that is to ask for a guarantee of omniscience.
I agree that the objective is to get closer to knowing the true answer, without necessarily reaching a point of certainty. But this debate to me exposes some serious limitations to Bayesian probability.
Both sides were well researched and seemingly agreed to most of the facts. However, they had widely disparate priors that affected their interpretation of those facts…which then in turn affected their priors for other or subsequent facts. And they ultimately came to wildly and widely disparate results. And there’s no objective test of their priors. So it seems their conclusions are driven by their views of facts which in turn are driven by their prior biases. This does not give me confidence in the objectivity of the process or the accuracy of any result based on Bayes.
i "like" the Cosmic Joker hypothesis, meaning it's the most *interesting* one to me. i also like the idea of the Covid virus as a virus in a superposition, taking inspiration from Scott Aaronson's comment in a podcast that "There is a serious prospect that in our lifetimes people will be able to create superposition States of, let's say, a virus". (dunc tank podcast, July 11, 2020, with minor editing from his improv oral) in this krazy view, the two states of the virus are: (i) being from the Wuhan virology lab: and (ii) being from the wet market. the evidence never resolves decisively to one of them. to put it another way, a meta-version of the lab leak hypothesis is true: the Lab is some Schrodinger's Box, and the Covid virus escaped while still in its superpositional state. it moves though our world leaving a neat two-path evidential trace.
https://philarchive.org/rec/BALCDT-5
Very nice, thanks for watching, and reporting. I want to make an argument for agnosticism. It seems like we are all inclined to think we must have an opinion on almost everything. What is wrong with saying, "I don't know"? Being trained in physics, I just want more good data before I try and make any sort of decision. Eventually I hope we will hear news of some intermediate animal host, or news from some researcher that says, yeah we were doing GOF on those viruses. Until then I don't see how the odds (for me) can be anything other than 50/50. A sign of my almost complete ignorance*, and an unwillingness to give anymore credit to one sides 'experts' vs the others. I think the best path for humanity is to assume both things are true. And try to make both forms of possible virus mutation less likely to happen in the future.
Oh, and why are there no error bars in all of this analysis. I'm thinking the error bars should be huge.
*I read a lot of stuff about this in 2020, said then we need more good data, and stopped spending time thinking about it.
> What is wrong with saying, "I don't know"?
Because it was turned into a political issue (very early on) so you have had two choices since roughly February 2020: loudly proclaim the party line (of the tribe you belong to), or be othered. This includes loudly proclaiming masks don't work and being a derisive asshole to anyone who disagrees, then loudly proclaiming the very next week masks are absolutely required and being a derisive asshole to anyone who disagrees.
Saying "I don't know" is dog-whistling for the other side.
Yes, that does mean we on-the-spectrum get kinda fucked because we speak in terms of facts and objective truths, but this is nothing new. History is full of this.
You are saying it as a joke, but it is a very important point. Why do we care if the probability of someting is this or that? We often care because the probability we assign is a signal to others about who we regard ourselves to be, including who we regard as friends and who we regard as enemies.
...and even if we do not care, we overlook that others will interpret our assigned probability this way at our own peril.
Thanks for the write up Scott! This is definitely very interesting.
I find myself walking away with a couple of thoughts. First, about the debate itself:
1) I started out pretty strongly leaning towards but far from certain about Lab Leak being correct. A lot of this I think stems from finding out that the anti-lab-leak side was a literal conspiracy that had a lot of institutional support to try to shut down the debate. Even those shouting down the theory secretly thought it might be true (at least early on)!
2) Likely related, I could feel learned helplessness setting in while reading the back and forth descriptions of the debate. With no priors on the believability of either participant, and seeing that both were making very strong claims that could reasonably answer their opponents, I want to mentally shut down and hold my existing perspective.
3) I end the debate slightly closer to zoonosis/no opinion, but still leaning towards lab leak.
Meta-discussion I feel much worse about Bayesian reasoning, especially the hard version with math. Peter clearly has a number that's beyond implausible, but the error bars on all of the numbers seem unworkably large. You could adjust all of the factors up and down by relatively small amounts and come up with any interpretation.
Even more, sometimes really really implausible things happen. As Scott notes here, something really implausible definitely did happen! In fact, across the world, implausible things happen quite often. People die of incredibly unlikely events all the time - an average of about 7 kids per year in the US have drowned in buckets since 1984. If you had never heard that number, what probability would you put on the chances of a child drowning in a bucket? There are 22.4 million kids 0-5 in the US right now, with seven per year drowning in a bucket, which suggests about 0.0000003125 chance of a child drowning in a bucket in a given year. No matter how you do the math, the chances are very truly low. But it happens seven times a year.
How would we model COVID's origins if something incredibly unlikely happened? It's not impossible that it originated in neither the Wuhan Institute nor the wet market. In the realm of possible, it may not even have started in Wuhan. Saying that something is *more* or *less* likely is all well and good, but reality happens in exactly one way, which can be the very rare event. I think 1/10,000 is incorrect for the likelihood here, but sometimes 1/10,000 things happen. Just because we say something is unlikely doesn't mean it's untrue. Saar implicitly believes that the wet market theory is ~6% likely to be true. 6% is far from ruling something out, even if he's arguing as if it's closer to 99.99% the other way.
Interesting debate, well covered.
I updated my lab leak confidence from 99% to 90%. It swayed but did not convince me due to the somewhat sketchy logic about wet market probabilities.
#1 I really enjoyed reading this.
#2 My beliefs shifted from pretty confident in lab leak to 50-50
#3 Reading through your summary, I often got this horrific feeling of college debate and dropping points. College debate has this concept of dropped arguments, where the worst thing you can do is not respond to a specific argument your opponent makes, because they automatically win that argument. That's what this felt like, I'm not actually convinced by Peter's overall story, I'm not even sure there's a coherent story there, but Peter kept coming up with specific situations where Saar was wrong and that convinced me.
I'm reading back through the lineages thing and, just reading through, Peters positive argument for why the older lineage started spreading later is pretty handwavy. Compare that to the next section where he goes through why the intermediates are a programming error, he's on point. But I come back to the core question of why the older lineage started spreading later than the new lineage and I don't have a great answer to the main point, and I think Saar's is better here, but Peter dominates on the subpoint.
I'm just scared of focusing too much on the subpoint and irrationally weighing that too heavily. I dunno? For people who watched it, did you get "debate bro"/Destiny vibes from Peter? That's what this kinda smells like, I notice intense debates should work but I could also watch 50 debate streams and at the end of it be dumber than I started and I felt this way a lot in college debate as well.
Mine shifted from 60-40 lab leak to 75-25 zoonosis.
Incidentally, my probabilities were and now are identical to yours. Also, Peter is astonishing, that is a wild amount of effort put into something.
This is probably not the important issue here, but as a former university debater I feel compelled to note that the 'dropped arguments' thing is specific to certain formats of competitive debating and far from universal.
This was a wonderful read and personally seems like a great experiment. A bunch of credit goes to Saar and Rootclaim for courtesy and making this public and educating all of us, whatever position you hold.
> along with their $100,000 table stakes
I thought the judges fees came out of the stakes, rather than being in addition?
I had us both bet 105k so that someone could still say they won 100k.
An insanely good debate in many ways, and congratulations to Scott on a most excellent summary. This reads like a communication from an alternate universe. Seriously, since when are there 15-hour highly structured debates at the object level with this level of intelligence brought to bear? That just does not happen in this world, except it just did.
Despite the outcome not being what Saar wanted, he has ably promoted Rootclaim as a useful tool. I want a lot more of whatever this is.
Despite Peter being one of the smartest people I've ever encountered online, arriving at an odds ratio of ~10^21 against lab leak is prima facie evidence that he has not correctly employed Bayesian reasoning. That sort of ratio is where you start getting "simultaneous miracles". I estimate that the chances of a *natural* zoonotic event, meaning one in which human action was not a causal factor, infecting its first patient *inside the virology lab*, as being *considerably* more likely than 10^21 to 1 against. I possibly could estimate the chances that I, personally, created COVID, had it released at the Huanan Seafood Market, and suppressed my memory of doing so, as 10^21 against. (By this I mean if I sat down and calculated how likely such an event was, 10^-21 is an answer I might obtain. In reality I would throw out any probability that low and group the event with "weird shit happens".) 10^21 is a sextillion. I can't count that high and neither can Peter.
Not getting answers like 10^21 touches on a real problem with practical Bayesian reasoning, which is that almost all independent events are only very nearly independent. Clinamina like the Self Memory Wiping Garage Biolab Operator Who Is Me exist in enough possible worlds to start having an effect if you multiply together enough large Bayes factors. This is far from disastrous to the project of being better Bayesian reasoners but there should be more effort going into developing and teaching the techniques that deal with it. Rootclaim looks like one of the places that might happen.
>Despite Peter being one of the smartest people I've ever encountered online, arriving at an odds ratio of ~10^21 against lab leak is prima facie evidence that he has not correctly employed Bayesian reasoning.
Agreed. 10^21 is an absurd number. There have only been 10^17 seconds since the Big Bang!
TBF I think this just is a matter of approach. Peter is reporting the nominal calculation on the assumption that the theory is correct, the evidence is what it's claimed to be etc etc..
I think it's clear Peter doesn't actually believe those are the betting odds that one should use on the subject.
So I think it's less of an error and more a notational difference.
"Seriously, since when are there 15-hour highly structured debates at the object level with this level of intelligence..."
I just wanted to suggest that actually, they happen all the time: in courts of law. And it's instructive to think how that comes about. It requires wigs and centuries of tradition; and really strict rules that get actually get enforced about what you can and can't say; and real consequences at the end.
And there is an obvious legal correlate to this error that Peter was making with his 10^21 claim: that poor woman who suffered successive cot deaths and was put in jail for them, because a statistics expert multiplied the probabilities of cot deaths together and worked out there was only a 1 in a million chance she hadn't murdered her own children. She was eventually released.
I was also thinking of that story related to the dangers of professionals tasked with making difficult decisions becoming (amateur) Bayesians. Could not find the proper reference, though. If you have it, I'd be interested.
One thing worth noting is that the odds Rootclaim gave at the debate are considerably higher, their central estimate was 1 in 180,000 for zoonosis and their high end estimate was 1 in 700 million.
To a large extent, I decided to just mock the process and make up even bigger numbers on the other-side. (more specifically, I gave fair justifications for many of the numbers, but the idea that you can just multiply them all together is not fair, because you might have picked the factors in a biased way, and in the end the argument should be bounded by whatever confidence you have in your own abilities)
For this write-up, Rootclaim asked Scott to change all their debate numbers to smaller numbers. The furin cleavage site that was 1 in 100 at the debate became 1 in 20, some other factors just got dropped entirely, etc.
Charitably, I could assume they asked Scott to make the change because they have since become less confident in their arguments.
But, cynically, I might think they did this to make it look like, "they were reasonable, I made up numbers that were too large, and the judges were just too stupid to tell the difference so they came to the wrong answer"
"I decided to just mock the process and make up even bigger numbers on the other-side" and people here are updating based on the result.
So do your own bayesian analysis with whatever numbers you find most reasonable, if you think bayesian analysis is the right way to solve the problem.
I think Michael Weissman's robust Bayesian analysis is dramatically better than anything I or anyone else has come up with on the quantifiable subset of the evidence: https://michaelweissman.substack.com/p/an-inconvenient-probability-v57
It is described as "excellent" in the post here but, despite being the most comprehensive and sophisticated analysis to date, its aggregated results aren't included.
I asked Weissman yesterday for an updated summarized table but he told me he's against tables and prefers people read the entire text...
His is a bit like Rootclaim, where he gives the market data zero weight, the lineage A/lineage B data zero weight, and so on.
He explains those choices in detail though. Many people would benefit from reading his explanation of why the market data doesn't support market spillover.
I know you know all this, but other readers here should note that of the three big papers arguing for market origin:
Worobey et al 2022's conclusion is under heavy fire in the statistics journals (Stoyan and Chiu 2024, Weissman 2024).
Pekar et al 2022 already had to correct their code ... and change their text to adopt a more lenient significance threshold. The same guy who found those problems has found further issues that invalidate the conclusion completely, though the authors have not had a chance to respond fully yet. Some ACX readers will be amused to hear that the four errors found in Pekar et al 2022 all had the same sign.
Crits-Christoph et al was debunked before it was even published.
I give A/B zero weight in order to be generous to the zoonotic case. Standing by itself, without the A/B data, without the missing RNA/hostDNA correlations, and without data on Wuhan markets as a fraction of China markets, then empirically the HSM cluster would get around a factor of 10 favoring zoo. Putting all those factors together tends to weigh against the market case, but a priori markets gave only a fraction of the zoo priors, so it's not worth penalizing zoo for the collapse of the market case.
Here's an 11.5 min talk that may help put the JRSSA ascertainment bias case in context. It's not a side quibble. https://docs.google.com/presentation/d/1y3_KYMnU7YyMjkMvIWcpOxGZkWTfQ6b2/edit?usp=drive_link&ouid=106543307891005183746&rtpof=true&sd=true
It plays in ppt if downloaded, but not in google docs.
OK, here's a table, for what little it's worth. Without the explanations, not much. Odds are for general zoo (ZW) vs. DEFUSE-style LL.
Priors: some pandemic start in 2019 1/70 ±x10
Sarbeco: 10/1
Wuhan: 80/1 (would be roughly 800/1 for market)
No-host: 3.3/1
Pre-adapt: 2.7/1
Has FCS: 3.3/1
CGGCGG: 7.3/1
Restriction site pattern 70/1
Each of the factors has already been discounted by a hierarchical inclusion of uncertain nuisance parameters. E.g. the FCS factor allows for selection by using only successful non-bat sarbecoviruses for comparison. I include a nuisance parameter for possible irrelevance and end up with only the 3.3 factor although there are 27 non-bat species with SC's detected and zero of them have an FCS.
The final odds are calculated by integrating over a fat-tailed version of the highly uncertain priors.
During the week of January 23rd, 2020, the day lockdowns were enforced, there was somewhere between 175-225 excess pneumonia deaths in the Wuhan DSP per 100K. This is eyeballable in figure 1 here.
https://www.bmj.com/content/372/bmj.n415
They report 243 per 100K for the entire first quarter of 2020. That's an *enormous* number.
Smooth this out over a week and say that during the week of 1/23/2020, ~30 more people than expected were dying from pneumonia in the Wuhan DSP per 100K.
The cool thing about the beginning of pandemics is that there's no immunity and the mixing assumptions of SIR models probably work well in the beginning of a respiratory outbreak. You can play with this one:
https://gabgoh.github.io/COVID/index.html
I couldn't find good information on the size of the Wuhan DSP, but allow it to vary between 2M and 10M people. Assume the R0 was between 3-4, the CFR was .5%-1%, and a few other now fairly well known things about the course of the disease. And look at the fatality numbers.
Using these ranges and standard compartment model logic, I don't see how you get such an enormous increase in excess pneumonia deaths consistent with an early December outbreak in the HSM. It doesn't make any sense at all. The best case scenario from these simulations is sometime early in November 2019.
Several of Peter's arguments treat the doubling rate as a real property at low populations. That doesn't seem right.
If patent zero only infects one person, the pandemic is delayed by one doubling period. If they infect 4, it arrives one period sooner. What are the relative odds between someone infecting zero, one, two, four, or fifty? When there are enough infected this kind of effect averages out and we can just say it doubles at some rate. At small numbers you get swings in infection progress based on how each person rolls.
Because of that attempts to extrapolate infected populations backward using the doubling rate seem doomed once you get into single digits. The error bars on that kind of estimate should be large. It's noisy.
So I don't think you can confidently say things like this
> The COVID pandemic doubles every 3.5 days. So if the first infection was much earlier - let’s say November 11 - we would expect 256x as much COVID as we actually saw
Or this
> Although Lineage A is evolutionarily older, Lineage B started spreading in humans first.
> We know this because Lineage B is more common. Throughout the early pandemic, until the D614G variant drove all other strains extinct, a consistent 2/3 of the cases were B, compared to 1/3 A. Both strains spread at the same rate, so the best explanation is that B started earlier than A. Since COVID doubles every 3-4 days, probably Lineage B started 3-4 days earlier than Lineage A, which explains why it’s always been twice as many cases.
Yes - the Skagit chorale is a great example of the superspreader dynamics at play. R0 is a post-hoc measure for clowns when it's as k-dominated as COVID seems to be. The problem with "COVID spread in the wet market at exactly its normal spread rate, doubling about once every 3.5 days. Stop calling the wet market a super-spreader event." is that it's "normal spread rate" is an average that is lumping in **the fact that COVID has super-spreader events**.
I find it curious this comment sits here unaddressed.
I think it sits here unaddressed not because a rebuttal is easy and people are bored, but because a rebuttal is hard, and unless rebutted, it seems a fairly devastating point. To elaborate a bit - and to integrate over a few strands of ~independent evidence:
a) We have two lineages, Lineage A appears (with v. high probability) older, and the wet market cases are lineage B (and iirc, no original Lineage A cases at all).
b) We have other relatively early cases that don't have a known market link
c) None of the early wet market cases look remotely like Patient Zero smoking guns (none of the animal trade workers, no infected animals detected at the market etc).
d) We see superspreading events at other wet/seafood markets in other cities/countries.
These fit quite well with a story of "the outbreak starts at unknown location, the first detected cases we see at the wet market are several generations in [something that in the abstract is pretty likely - what are the odds of observing Patient Zero?], a superspreader event happens at the wet market". This story is bad for zoonotic proponents because it doesn't have a high odds ratio of P(.|zoonotic)/P(.|LL)
The response is "well, we can't be missing the first few generations because we know the doubling rate and if were missing the first few generations then by January there would have been far more deaths/cases". And then conclude:
(i) "well Covid can't have started earlier, so there can't be meaningful early non-market spread, so anyone who looks early and is unlinked must have had a really short incubation period", [even though ex ante that's less likely and should be penalised].
(ii) "well, Covid can't have started earlier, so even though Lineage A looks older, since Lineage B is more common, it must have started spreading first"
But these deterministic timing arguments are nonsense because not only is stochasticity potentially important anytime you have small n, the over dispersion of Covid spread is huge so the variance of spread early on is very high. The most common number of people for someone to infect is 0. The second most common is 1. The vast majority of spread is driven by a small minority of infected people.
So the argument dismissing the Lineage A vs Lineage B point dies immediately. (Suppose Lineage A has a 1-2 generation head start. What is the probability that a month hence it has more people infected than Lineage B, if both have identical latent spread potential? It won't be far off 50-50, even if you condition on both surviving, which we should). You don't even need to penalise Lineage B being smaller as "improbable bad luck, what a coincidence that the more predominant lineage is the one at the market", because while there is probably lots of room for random luck, the most likely reason Lineage B "got lucky" and outgrew A early on is that it (by happenstance) spread at the market while A did not. E.g. if I tell you "Lineage X has a 1-2 generation head start, but the first known Lineage Y case is at the wet market, which lineage has larger share a few generations hence", isn't the smart money on Y?
And the point of "the total outbreak would be too big by date T if it started a couple generations earlier" also dies immediately because the 95% CI of outbreak size after 1-2 months given a t0 probably covers at least an order of magnitude.
It seems really bad for Peter's arguments in general that he would consistently make this obviously wrong argument, to argue against multiple lines of evidence, seemingly without realizing that it is wrong.
I can't find the ref, but there was a paper that looked at changes in reporting criteria during the outbreak. In addition to any differences due to the small population, the reporting bias also changed over time.
The doubling rate argument is, indeed, oversimplified and invalid in the earliest stages when you have single/double digits of cases. I have also programmed epidemic simulations that take that into account. Here's a recent write-up:
https://twitter.com/tgof137/status/1772417277670871113
How sensitive are the sim outcomes to the choice of distribution here? https://twitter.com/tgof137/status/1772417293181505543
I find that for the range of values that I find most plausible, it's hard to support any case that there was a large covid outbreak besides the market.
If I were really motivated to develop a lab leak case, I suppose the approach would either be to say that the outbreak in January is much larger than anyone thinks, or that the growth rate (R0) is slower than anyone thinks. I don't think that changing the dispersion/superspreading rate (k) would impact it as much.
Changing January size is hard because you can bound that by excess deaths and serology. And you can measure R0 in other cities besides Wuhan, so you can't make that so much smaller.
I am, of course, not sufficiently confident in this model to say, "there were exactly 70 cases in Wuhan, on December 10th". There's a range of values, both because the parameters aren't exact, because there's a statistical range even within given parameters, and because the model may not perfectly represent reality.
None of this absolutely disproves a lab leak, either. It just makes it very unlikely because the lab leak needs to cross town to the market within, say, the first 40 cases or less. And then both lineages need to show up there, and the clock needs to reverse by chance.
Just checking: the simulations you're referring to are your implementation of Hao et al 2020, which you said was too messy to put up on github?
It's just some code I cooked up from scratch in C++. Some of the input parameters are based on Hao et al 2020.
I'll post it at some point -- one scientist wants me to publish it, maybe I'll do it that way, maybe I'll just share it otherwise. I'm too busy with some other stuff for the next few weeks.
The original motivation was just that I didn't want to trust Pekar's paper on faith, so I wanted to see if the same results just fell naturally out of any reasonable model. I tried several models and they got similar results.
I think the person suggesting you publish it thought that peer review would be a good learning experience for you, since you weren't accepting informal feedback. I'm afraid they weren't suggesting that your work was good enough that you should go public without anybody double-checking it. You didn't seem to understand why people were telling you to preregister either.
Pekar et al 2022 has serious errors despite having 29 authors and however many peer reviewers. @Nizzaneela has worked in the open so you can see how subtle the bugs are. And you think you can just hack up some C++ and settle the COVID origin question!
Have you thanked Rob Townley yet? If he hadn't taken pity and given you the Hao paper a couple days ago, you'd still be trying to reinvent SIR from scratch. You flat out insulted the expert who recommended the Epidemics on Networks book and python code for you.
I was worried about k in particular. The variance in the start time estimate should mostly come from kurtosis.
The original model I built used a poisson distribution of secondary infections and it still showed a similar result for 2 lineages vs 1. I didn't test that one for outbreak timing.
I can do a write up at some point to show how this all varies with each parameter. But I need to get a smaller search space before doing that.
The first thing I was trying to do is get a better answer from a few scientists about how to set the mutation rate. The answers I got were highly technical and will take me days of study to sort through. I could just linearly scale it to some value that I think is right, but I want a better justification than that.
Isn't Poisson completely wrong? That implies common latent mean across individuals, which is inconsistent with evidence of large over dispersion
>Several of Peter's arguments treat the doubling rate as a real property at low populations. That doesn't seem right.
True! At small numbers, the fractional scatter is large.
I find it somewhat amusing that in textbooks illustrating radioactive decay and the idea of half-lives, they always show curves that go e.g.
1024, 512, 256, ...
_not_
1024, 512+-sqrt(512), ...
they don't show the randomness of all the individual decays and how that scales.
This is silly but:
> This is a pointless too-clever-by-half “flourish” that there would be no reason for a human engineer to do.
Have you ever met an engineer? We do this kind of nonsense all the time! This particular argument cuts in exactly the opposite direction Peter intended.
I don't have any comment on the outcome, but this little detail drove me nuts.
I think it's different in biology (assuming you're speaking from a non-biological engineering point of view?). It's been a while since I did any myself, but making mutations in any organism is not something you mess around with. You insert your chosen mutation in the simplest possible well established way because even then most of the time it won't work. Messing around for the sake of it is best kept to other things, like giving genes and alleles funny names.
Peter claims that the WIV sampling stopped in 2015. This is not true. See page 60 downwards of this FOIA. https://s3.documentcloud.org/documents/21170561/536974886-gain-of-function-communications-between-ecohealth-alliance-and-niaid.pdf
EcoHealth say this to NIH in 2018:
"As per last year, we will not be subco ntractin g any funds to the intuiti ons in these
countries. All efforts expended in these co untries will be from collaborating
partners and not funded by our award. PI, Co-Investigators or other team
members may con duct short field ttips to assess markets, identify wild life in
them, and anange for shipment of samp les of bats and other high-risk host species
in co untri es that nei ghbor China (Burma, Vietnam, Cambodia, Laos) and that
supply wildlife to the internationa l trad e to China (Tha iland , Malay sia,
Indon esja)."
i.e. they say in 2017, they did sampling in these countries, and all samples would be sent to the WIV. Weirdly, all of the closest known relatives to SARS-CoV-2 have been found in places the WIV was actively sampling for more than a decade.
This is not even considering the sampling that the WIV would have done without EcoHealth funding.
The DEFUSE proposal proposes extensive bat harvesting in Yunnan in 2018. This is also evidence against them having given up on bat harvesting in 2015.
@Peter, what are your thoughts on this?
My intuition is that Rootclaim is correct that you should be able to use Bayesian spreadsheets--especially after some adversarial collaboration--and get a format that consistently outperforms intuition and even professionals informally weighing the evidence. It's sort of like how as an analyst I use fermi-calculations to get much more reasonable estimates of unknowns than just trying to estimate the unknown directly. I think Rootclaim should focus more on testing that.
I'm with you (after reading two long summaries of their debate) that zoological is more probable but also that Rootclaim is a worthy project. I wonder if the change in format that Saar should focus on though is more emphasis on the spreadsheet and adversarial collaboration to get both participants using the same spreadsheet. I'd love to see a version of the spreadsheet that two smart opposing viewpoints people agree is a reasonable setup (without any multi-stage fallacies, etc.) so we can see which line items are driving most of their disagreement free of any logical reasoning errors (or at least free of errors that adversarial smart people could identify). And then judges can focus more on the individual ratios (and any remaining spreadsheet steps that debaters couldn't adversarially reconcile)
The problem is that it's us people don't come pre-equipped with a distinction between priors and evidence and we can't unlearn facts. Thus, even if in theory it's a good system you can't get reliable data for it because you can't go back and figure out what your prior was. As such it's very easy to double count or undercount evidence.
It works great in situations with some obvious prior because it's a common situation (do I have cancer...well I know what the base rate for my demographic factors is and can adjust that by test results) but most problems aren't like that.
Doesn't normal human reasoning have the same problem but just isn't explicit about it?
The hope here though is the adversarial collaboration aspect of it would point out when double counting or undercounting is occurring and judges could focus in on those areas as driving the crux of the disagreement.
Ofc it does, but the point is that we don't have a reason to think this mechanism solves those kind of failures as it's being promoted to do.
It's like saying: I'm going to solve the political mess we are in by first figuring out what the government should do and then persuading the voters to support it.
I mean maybe I will but breaking it up into those two steps didn't help, it just gave names to a way you could imagine making things better.
Interesting debate with Mark here. Have you written anything else/longer on the cognitive & other problems of separating priors and evidence?
That was really interesting. My bayes argument would be something like "If covid WAS leaked from the Wuhan research lab, how plausible is that picture of events?" If so I'd expect an average amount of "well, we don't know" for scientific research. But each thing that needs to be explained away rapidly adds to "that doesn't sound right". Today I learned that they did do modifications like this. But I also learned that they were usually publishing their research like any academic institution, so you need coincidences or cover-up to suppress anything they'd done on COVID specifically.
This includes a lot of things that I don't know. But I still like my reasoning, because I know there's lots of biologists out there who have a better picture than me. And I haven't heard of any reason for biologists to avoid thinking about a lab leak, and don't think it's likely there could be a reason that wouldn't be well-known. (Whereas there ARE reasons for physicists to dislike many worlds, or doctors to dislike "chronic fatigue exists and easy weight loss doesn't") But what I usually hear from random biologists is "well, yeah, it could be, but I don't have any particular reason to think it is". And so I think that's probably right.
I'm interested in the ways of putting more specific numbers into reasoning but I think you can rapidly get into territory where it's hard to tell if the reasoning is sound. E.g. trying to describe how much of a coincidence each theory is raises questions like, if it was zoontic, how many other places would have sounded as suspicious as the wuhan research lab? Restricted secret areas? Other kinds of labs? How unique is the nearby lab, in advance? If it was a leak, how many places would have seemed as suspicious as a wet market? Nowhere else seems very suspicious, but "coming out of nowhere" would have seemed completely plausible for zoonotism. Or how much are the lab and the pandemic both there because there's convenient reservoirs of coronoviruses nearby, that could be, I've no idea. I think trying to make these precise leads to "things we didn't know we needed to account for" swamping the analysis.
I ought to read how the probability calculation is actually working. I would be really interested. Although I'm not sure it would help: e.g. I'm more persuaded by my argument of listening to biologists, so any attempt to quantify likelihoods for genetic arguments will be swamped by the chance of something I don't know mattering.
I should also add, I don't really know how plausible it is that COVID was circulating a few months earlier and just happened to take off at the time we identified. That seems pretty plausible from a general "we can't usually trace outbreaks back to an individual person" expectation. This debate showed some argument why it's plausible it started at the market, but I don't understand the biology well enough to really know. But it doesn't really affect how likely I think a lab leak is, that's more based on the default in the absence of evidence being spontaneous mutation, not the specific evidence for the market origin.
Perhaps this description of the debate wrongly magnifies this, but it struck me as really discrediting to the lab leak side that they kept relying on debunked claims. The LL side would bring up things lime the 90 missing cases, the daily mail guy, or the idea that the market was a superspreader event, and Peter would describe that evidence in greater depth and be able to pinpoint why that evidence was faulty, which was not rebutted by the LL side. One of the most annoying habits conspiracy-adjacent thinkers online have is to gish-gallop between weak evidence they're only lightly acquainted with, and then not change anything about their thinking when a piece of evidence is disproven. I generally find the meta-bayes "what about the odds your argument is secretly bad and mine is secretly good" mostly annoying, but how can you not apply a pretty heavy discount to the credibility to a side that is frequently trotting out rather flimsy evidence?
Agreed, and this is a significant part of what updated me heavily towards zoonosis. I didn't get the impression the lab leak folks updated at all on Peter's comprehensive dismantling of a large number of arguments.
Yes, same.
The $100K prize introduces a huge signaling hazard, mainly because Saar is rich and Peter isn't. I'm sure Saar's lost that much playing poker. For Peter, this is a sizable portion of his net worth — so he plays to win (and, as other commenters have pointed out, may be playing a bit dirty).
That said, the biggest issue with this debate format is that it's about an *incredibly important topic* where several major players have *a lot* of skin in the game. As much as this is a battle of evidence, it's also a battle for narrative (see Peter's rationale for wanting to get one up on the conspiracy theorists). And the "Official" narrative, that all good, sensible, educated people should agree on (you don't want to be a conspiracy theorist, do you?) is zoonosis. There are probably better, less charged questions where the Rootclaim Bayesian approach would work.
Not that I'm entirely convinced by Lab Leak either - maybe 60% odds? But given that I read ACX while goofing off at work, I don't have time to watch 15 hours of debates. Gotta go with my gut at some point.
I don't think that's really a problem. Saar has an idea he wants people to take a look at so he's paying to get eyeballs. It's just like Pepsi advertising on TV with more social benefit.
I actually underestimated the signalling value of it since I'm so used to rich people (my pillow guy) announcing this kind of shit but then adding fine print that makes it impossible to win.
Maybe signaling hazard is the wrong way to put it. The $100K means there are very strong incentives for Peter to *win the debate*, rather than find the truth. Saar is rich af and it's marketing dollars to him.
Certainly, but that's not bad for the listener. Indeed, I'm well served if both sides are strongly incentivized to win (and Saar is motivated here by wanting his idea to look good).
Two big issues with trying to formalize Bayesian reasoning is to properly identify what is already part of your priors and what isn't. Second, using it at scale in a systematic way so that you actually benefit from the correct updating mechanism. After all, the more you just rely on the prior the less it makes a difference.
The problem is that we only pay attention to issues once something makes us think there is an interesting question there and often these issues are relatively unique so we have no obvious choice of base rate to use as our prior. And there is the very real danger of pulling in information that is really only a result of the update.
I think distinguishing the prior is one place where AI can sorta help. Maybe not yet, but a great feature with AI is that you can remove data from it's knowledge and see how it would judge things. I'm particularly hopeful for this in politics (this AI predicts your judgement in these political issues 90% of the time and when it doesn't know that Trump supports this policy it predicts you will too).
I'm more skeptical about the ability to do this systemically at scale in a way that uses enough interconnections to improve judgement.
Thanks for this add-on to your comment further up, on the problem of differentiating information used to form priors and information used as evidence.
Also a statistical analysis that favors bass sounds awesome!
Some people talk about mammal origin at the seafood market, some talk about accidental infection from a bat at the virology lab, some people talk about gain-of-function gone wrong - why does no one consider the hypothesis that it was actually the seafood at the seafood market?!
I'm absolutely convinced the first mammals originated at a seafood market!
I think people get caught up about the furin cleavage site mutation because they're looking at the wrong base rates. To think about the furin cleavage motif, you need to ask (I'm phrasing these questions a bit imprecisely):
- How often are furin cleavage motifs studied in places other than Wuhan? How often are they studied in Wuhan?
- How often are furin cleavage site insertions the cause of viruses learning how to become pandemic? How often are other mutations the cause of viruses learning how to become pandemic?
I don't see people asking those questions. The fact that people don't ask those questions makes me suspicious of everyone involved, honestly. These appear to me to be the questions that arise out of applying Bayesian reasoning straightforwardly. (I'm also suspicious, as a sidenote, because the question overall is kind of irrelevant: everyone agrees lab leaks can cause pandemics, and everyone agrees zoonosis can cause pandemics, and everyone agrees zoonosis is more common than lab leaks, so we should just try to prevent both, focusing more on zoonosis than lab leaks.)
Also, WTF do people care so much if this particular disease originated from a lab leak.
We have a pretty good grip on the problem of biolab safety (eg the leaks in the UK) and the existence of gain of function research. So whether or not COVID itself happened to be a lab leak isn't reason to update too much on the future probability. Any forward looking policy question shouldn't really depend much on whether COVID was or wasn't a lab leak.
I get the debate is interesting but it's ironic that a claim to be more rational is being tested via a subject that it's irrational to be so concerned about.
If we determine someone to be guilty of something bad regarding this case, we can punish them.
But people aren't more morally blameworthy because they were unlucky. I don't think we have any reason to believe that the people in Wuhan virology were behaving particularly badly so why go looking to punish someone?
If you want to punish people who behave badly we have a great way to do that: look for the evidence of bad behavior and investigate those cases not the ones where things just happened to go badly.
Sure, for some types of behavior that's really the best we can do. The primary evidence we get that someone might have been shaking their baby is when it goes wrong and the baby ends up hurt/dead (tho still often wasn't shaking). But this is exactly the opposite situation.
We have piles of information about how biosafety labs are run and plenty of cases they behave incredibly dangerously. We know what research gets done because it usually gets published etc. So we have plenty of leads about who is behaving in unsafe ways if we are intent on punishment.
Focusing on the time shit went just seems like hurting someone to give us emotional satisfaction about a sucky thing we experienced and that's awful.
>I don't think we have any reason to believe that the people in Wuhan virology were behaving particularly badly so why go looking to punish someone?
Well, the talk about doing dangerous research in an insufficiently secure lab does sound pretty bad, and that the US were willing to collaborate with them also isn't a good look.
A general program to go after people doing what you just described seems reasonable but it doesn't seem we have particularly strong evidence it happened here.
If it shows up fine. But there isn't any reason to particularly focus on this case because bad shit happened. If this was the cause it would be crazy unlikely this was the first time someone was breaking the rules like this.
Most people who care probably have political or emotional reasons, whether it be a desire to blame China, or anger at the US health institutions and distrust of the mainstream media.
But ignoring that, if the risk of pandemic from zoonosis is high and the risk from lab leak is low, we probably want more gain-of-function research so we can prevent the former.
My favorite post in a long time. Thanks for the recap Scott.
"Either a zoonotic virus crossed over to humans fifteen miles from the biggest coronavirus laboratory in the Eastern Hemisphere. Or a lab leak virus first rose to public attention right near a raccoon-dog stall in a wet market."
One thing that strikes me is that there is an asymmetry here. The first part of the above does indeed sound like a grand coincidence. It's not like the mutating virus would know that it happened to mutate near a lab. However, the second part is only a grand coincidence if we assume that if there was a lab leak, that it was a) fully accidental, and that b) no one acting in bad faith had the ability to distort data from early testing, etc..
I'm not sure I'm of the opinion that lab leak is thus more likely. But I can't help but think: it sure would be convenient if you were a lab researcher/administrator/related government official and you didn't want to be blamed for something you negligently made possible - then it sure is convenient that there is that nice big wet market nearby that perhaps you can shift the blame toward.
My current thoughts: though I didn't sit through the 15 hours - I am now of the opinion that the zoonotic hypothesis is stronger than I had thought before. I am updating to be less confident in the lab leak. But I'm probably now to about even odds, fwiw.
Exactly my thought. I don’t understand why this argument wasn’t presented or discussed. I know nothing about it, but why would we trust any data or research coming from China on this? My default prior would be that the ccp goes to extraordinary lengths to fabricate an alternate reality in many other places, and they have reasonably strong incentives to do so here, so why wouldn’t they? I’m assuming this presents as obviously not the case to the people on your ground otherwise it would have been brought up?
Well, the best argument here is that the CCP likely didn't know exactly what, if anything, they needed to fake in the earliest days, so there's no particular reason to suspect the initial data to be (competently, non-obviously) tampered with.
There are very few coronavirus laboratories, but there are many wet markets. We don't really get to condition on "raccoon-dog stall", but rather the set of all animal stalls such that they are plausible carriers such that a zoonotic origin could be posited - so:
Bats
Pangolins
Mustelids
Felines
Dogs
Deer
Rodents
Marsupials
And at this point I'm just tired of listing the animals that COVID-19 or sufficiently close diseases have been observed in that somebody could posit a zoonotic origin, given COVID-19.
But it's worse than that, because there's nothing preventing a zoonotic disease from appearing in places other than wet markets. That it appeared in a wet market near a particular stall feels like strong evidence because it's so specific, but -any- real world-case would be equally specific - what we actually need to consider is the full reference class of places which might -feel- weirdly specific. Like, it had first appeared in a zoo with an on-site veterinary clinic which regularly treats wild animals. My local zoo does this, although I can't find statistics on how common this is. Or maybe the zoo just got a shipment of wild animals, which isn't that unusual, particularly in China. (Sure, none of the shipment of animals tested positive, but did any of the raccoon-dogs test positive?) Or maybe it first showed up in a restaurant who had a staff member test positive, and who had recently been at a wet market.
That it first showed up where it did, then, doesn't actually give us very much evidence for or against a zoonotic origin - the reference class for places that intuitively feel weirdly coincidental/specific is huge, because the specificity is post-hoc.
However, that it showed up where it did relative to a coronavirus research laboratory researching not just the disease, but the specific thing that made the disease such a problem, is much more unusual; there are only a handful of places in the world which satisfy even a broadened criteria.
You'd have to have quite high priors against a laboratory leak happening at all to arrive at a conclusion, based purely on the location of first observation, to think that it is zoonotic based on that information. The location is strong evidence in favor of the lab leak hypothesis.
There's probably under a billion people who shop at wet markets with this sort of crossover potential. Something like 1% of them are in Wuhan.
I think you are overstating the numbers here. 38 raccoon dogs were sold per month at four markets in Wuhan. So, maybe around 20 raccoon dogs were sold at Huanan in the month of November 2019. The zoonosis hypothesis outlined by Miller is the following specific claim: the 20 raccoon dogs sold at the Huanan market were infected with covid and caused 8 spillovers, 1-2 of which sustained an ongoing outbreak. This is an extremely small fraction of the relevant wildlife trade in China, <<<1%.
Despite this, there are various points of counter-evidence against the raccoon dogs:
Negative correlation with covid in the environmental sampling, but positive correlation for animal viruses the animals were actually infected with.
No raccoon dog ever found infected with the progenitor virus
No raccoon dog ever infected by humans in later outbreaks, despite this happening several times with other animals such as mink.
The raccoon dogs were wild caught in Hubei, not farmed
Michael Weissman's analysis is by far the most principled and persuasive thing I've read on the topic, it deserves much more discussion than just the link.
Some examples of things mentioned in Weissman's analysis that don't show up here:
* the initial outbreak at a market proves very little: the initial Beijing superspreader event was in a wet market, and the largest outbreaks in many countries were at markets. [EDIT: actually this is mentioned above]
* lab accidents are much more common than you think
* Pekar 2022 looks like it's failing replication: https://pubpeer.com/publications/3FB983CC74C0A93394568A373167CE#12
* spatial statistics experts dismiss the methods of Worobey 2022
* the 2018 DEFUSE proposal really predicted some highly unlikely features of the virus
This was a debate between a smart autodidact and an unprepared multimillionaire, in a format which prevents easy annotation and rebuttal. It's maddening to watch the debate because Peter gets away with so much. I'm really surprised how much people are updating based on it.
Do you think someone reading Scott's summary and someone watching the debate video would come away with very different impressions? It sounds like you did?
I think this post goes beyond being a summary of the debate.
In a tendentious way?
I think I basically agree with you on all the substance - I was frustrated by some of Peter's questionable claims that were allowed through as fact, and the Weissman post you linked to was very helpful (and does a much better job of discussing what the conditioning set is, and how that affects what different bits of evidence mean). But beyond everything else, I agree with you on "why are people updating". If one person is much slicker, and is much more prepared, and has nice "just so stories" to explain away certain bits of evidence, then ***holding everything else constant*** this should reduce how much you update in their favour. A more prepared person is more likely to win a debate on points/style even when they are wrong. Verbal debates just aren't a very good mechanism of accurate determination of truth, they have a lot of noise/aesthetic factors seem to have great influence on perceptions! You just shouldn't update that much when a priori there's no reason to think the best arguments on both sides will have been presented / equal scrutiny been applied to claims on both sides. (This of course blows up 100x when you see the slicker person getting away with clearly erroneous claims).
The claim that you shouldn't have high Bayes factors on both sides of a debate is just false.
Every event with sufficient detail involves a ton of unlikely stuff happening. Every time you flip 30 coins you get an exceedingly unlikely list of heads and tails.
Sure, if you understand the data to be the **complete** specification of the even then yes you don't get high Bayes factors on both sides because you only get one. But that's not what's going on here. They are both cherry picking the aspects of the situation that most favor their hypothesis so of course you get high Bayes factors on both sides.
I don't have anything to add on the COVID origins topic, but I do want to register some (possible?) disagreement with this bit:
"For example, suppose I win the lottery, I’m told I win the lottery, the lottery company gives me a big check, I cash the check, and I become rich. Given that there were 1-in-100-million odds against me winning the lottery, the lottery company giving me the check and so on must be at least 100-million-to-1-level evidence - otherwise I should refuse to believe I really won the lottery, even as I enjoy my newfound wealth!"
The quip-level counterpoint: You *shouldn't* believe that you won the lottery, because which is more likely -- that I won the lottery and cashed the check and got all this claimed evidence, or that I'm *dreaming* that I won the lottery and cashed the check and all the rest? After all, I've had several dreams where I won the lottery before, and I can get just about any evidence you like in a dream (all made up by my brain, of course). Sure, a 10^8:1 update is possible in this case (er, maybe, I'm not sure because it sort of hinges on things I think I know about dreams but which might be false -- after all, I might be dreaming I know them), but the strongest counter hypothesis is this sort of out of band thing (dreaming, or maybe something like solipsism or a "living in a simulation" or "being in The Matrix" or similar roughly-equivalent thing) rather than something like an elaborate prank or a mistake.
Yes, on at least one occasion I have realized that I was dreaming by applying this logic.
This might seem academic or silly -- how do I **really know** that I'm not dreaming my whole life, being continually deceived by Descartes's demon, etc. -- but I'm pretty sure that similar caveats ought to apply to trying to reason about some things that people care about, like "Is there a God?". If you think you are more than 99% sure you know the answer based on Bayesian reasoning, you are being overconfident that you (a) understand the various conditional probabilities involved and/or (b) have properly considered all the relevant hypotheses.
Well, it's pretty obvious that I haven't just been dreaming my whole life because the human brain is simply not powerful enough to simulate the level of consistency that "reality" offers. Of course, that would be making a pretty big assumption that it's a human brain that's producing the dream, but there just isn't any rational incentive for any potential beings to produce such a simulation. ...If they wanted to torture me, they could certainly do far worse.
> Well, it's pretty obvious that I haven't just been dreaming my whole life because the human brain is simply not powerful enough to simulate the level of consistency that "reality" offers.
But within a dream (or within a dream-within-the-dream, if we accept the hypothesis) we generally don't take notice of the inconsistencies.*
So if you were dreaming your whole life, it may well be full of inconsistencies which you simply haven't noticed - but likely will once you wake up.
*Sometimes we perceive noticing the inconsistencies as causing us to wake up (as frequently depicted in movies and television - there's a House M.D. episode that immediately springs to mind, and I vaguely remember the Sopranos using the motif as well). In reality, though, it's the opposite - our conscious mind waking up causes us to notice the dream's inconsistencies.
I think that most people overestimate how confident 99% is, or even 99,999,999-to-1. Those are both big, but you can reverse the first by observing 7 coin flips all go the same way, and even the latter just needs 27 coin flips. You're unlikely to find this evidence if the thing you thought was likely is in fact true, but it's actually not all that much evidence you'd need to see to convince you of the a priori unlikely thing.
Great example/illustration!
I think this is a general case of the principle that anyone in a weird enough situation should think their whole life is a dream or simulation. For example, I think all heads of state and billionaires should think this (though luckily most of them haven't reasoned it through, or else politics would get very weird).
I'm not sure if my situation (famous blogger) is enough to get me in this position; I'm about 50-50 on this one.
>or else politics would get very weird
Is this tongue-in-cheek? I've long thought that we were transplanted into an alternative clown reality around 2015.
Isn't just being a human after the industrial revolution (and just before the singulairty maybe?) already a very strange situation, rather than being a trilobite during the Cambrian or something?
What’s the Bayesian probability that Scott wrote this entire article to make the pun in the first section heading?
> So probably we should just accept that the first reported case - a wet market vendor, December 11
So, after searching a bit, the closest article I found is this : https://www.nydailynews.com/2020/03/27/shrimp-vendor-identified-as-possible-coronavirus-patient-zero-leaked-document-says/
A shrimp vendor, not raccoon dogs vendor. This is to me an important piece of information and leaving it out feels very misleading.
Was her stall next to the raccoon dog stall? It looks like most of the early cases were.
I don't want to divert attention from the debate, but since it seems appropriate, I'd like to share a post I wrote nearly three years ago proposing an hypothesis that is not exactly lab-leak, though not incompatible with it, and not zoonosis (at least as I've seen it stated):
https://medium.com/analytics-vidhya/a-novel-hypothesis-previous-ancestor-virus-ba2ae7b10062
Short version:
The propagation of Covid-19 seems to point to certain areas of the world being less affected by the disease for no particular reason (no younger population structure, no less obesity, etc...). The main area being Southern China and Northern SEA. This is precisely the area where you would expect zoonosis for a bat coronavirus to happen. Other geographic areas correspond to places where infected people might have first, and in greater numbers, arrived if they had been infected by a previous virus in the aforementioned main area. This and other evidence (see full post) point to the possibility that a previous ancestor (to SARSCov2) virus emerged and propagated in that area, before mutating into SARSCoV2.
The competing 1/10,000 claims about the wet market are interesting, but feel too speculative without enough hard evidence for me to know what to do with. Far more salient and intriguing to me are:
Saar's claim that the 12-nucleotide furin site comes out of nowhere...
> COVID - which mostly just resembles BANAL-52 with a few scattered single-point mutations - has twelve completely new nucleotides in a row - a fully formed furin cleavage site that came out of nowhere.
vs. Peter's claim that the the furin site doesn't use the expected sequences for such a site and actually uses sequences that wouldn't be expected to work. Plus Peter's claim that WIV didn't have the technology or expertise to insert such a site anyway
> COVID’s furin cleavage site is admittedly unusual. But it’s unusual in a way that looks natural rather than man-made. Labs don’t usually add furin cleavage sites through nucleotide insertions (they usually mutate what’s already there).... COVID’s furin cleavage site is a mess. When humans are inserting furin cleavage sites into viruses for gain-of-function, the standard practice is RRKR, a very nice and simple furin cleavage site which works well. COVID uses PRRAR, a bizarre furin cleavage site which no human has ever used before, and which virologists expected to work poorly. They later found that an adjacent part of COVID’s genome twisted the protein in an unusual way that allowed PRRAR to be a viable furin cleavage site, but this discovery took a lot of computer power, and was only made after COVID became important. The Wuhan virologists supposedly doing gain-of-function research on COVID shouldn’t have known this would work. Why didn’t they just use the standard RRKR site, which would have worked better? Everyone thinks it works better! Even the virus eventually decided it worked better - sometime during the course of the pandemic, it mutated away from its weird PRRAR furin cleavage site towards a more normal form.
Peter's claims seem stronger to me, and I didn't see any response to these specific claims in the summary (although I may have missed), but I also can't explain Saar's claim, nor do I have the expertise to assign probabilities to any of this. But these claims seem to me to be the most important and what I would love to see further discussion about.
I'd put basic internet research on the subject as "very low reliability", so, take this with a grain of salt, but:
Some internet research suggests that the PRRAR furin cleavage site is better-adapted to animals (ferrets in particular) that were used as test subjects at the laboratory in question (but poorly adapted to bats and pangolins), and the RRKR cleavage site is better-adapted to humans.
Interesting, thanks! In which case, Peter's claim could possibly (accounting for salt grain) fall under the all-too-common "Simple Absolute Claim About How Humans Always Do A Thing Turns Out To Overlook Real-World Nuance"
It seems entirely plausible that the WIV was trying out a new type of furin cleavage site, to examine how it worked. It's evidence against SARS-CoV-2 being a deliberately-created bioweapon, not against it being part of a research project about how animal-borne diseases can jump to humans. In my own field I've often copied something new and interesting that I saw in the wild, just to see how it works and improve my own understanding, and I'm not even a research scientist.
typo: they a found
I just wanted to respond to this point in the conclusion:
"Cool Machiavellian plot you have there, but maybe the fact that you’re losing 16%-66% should make you question whether you’re really as smart as you think you are...At some point you have to start debating!"
The reason that people in the field might be reluctant to accept this point is that they don't, and shouldn't, give a shit what 66% of the American public think. And debating just gives the American public the impression that it has some kind of a right to have a say in what the truth is. If you're a virus researcher now, working on how to predict and prevent the next zoonotic infectious epidemic, you really really don't want the public deciding to shut you down or order you to focus on lab leak prevention.
This is obviously very problematic - I'm kinda making the argument for important things to be decided in smoke-filled back rooms. But it's also unavoidable, I think. Science must by its nature be difficult and cutting-edge and unknowable to most people; and scientific advances are by definition decided by something other than human opinion - ideally, by what the universe is really like. So there is just an inherent limit on the level to which science can be democratised. Which means that noting which way public opinion blows is unlikely to be very relevant to the people who want to advance science.
That said, I do think that science is getting more democratic, and that that's a good thing. It's just a messy process.
"If you're a virus researcher now, working on how to predict and prevent the next zoonotic infectious epidemic, you really really don't want the public deciding to shut you down or order you to focus on lab leak prevention."
A question in good faith: Is "working on how to predict and prevent the next zoonotic infectious epidemic" really what this field is doing? We just had a major infectious epidemic. Did work in this field help predict or prevent it?
At UC Berkeley in the 1980s (and so presumably lots of other places) there were loads of scientists talking about the need for pandemic prevention research. The people who have been doing that kind of work for decades have been working without adequate funding. None of these people are surprised this happened. The fact it happened as they predicted it would is not evidence for their work being useless.
Millions of people were killed or disabled by polio.... until they weren't. We could make a list of many other things like this in the realm of scientific research.
Was there any research going on at WIV that was useful in predicting, preventing, or combating COVID-19? I'm sure there are lots of smart people spending lots of money studying viruses there and elsewhere. But what good have they done anyone.
I specifically ask, because it does seem like there's a >0% chance that the research itself is actually increasing the risk of pandemics.
I feel like the best answer we can give to that is: look at the historical record. Are there in fact more pandemics now than there used to be, or fewer? I think it's fewer, because many serious diseases like polio have been wiped out in the developed world, and flu has been controlled. You're right about that non-zero probability of science making things worse, but the facts seem to suggest we're moving in the right direction.
There is always a nonzero chance that scientific research will make some things worse. We have a ton of examples of this. Thalidomide given to pregnant women causing tons of children with birth defects. There are so many examples.
Do you think vaccines have been not worth it? Antiobiotics? Everyone studying infectious disease treatment is working with diseases in labs. Would you have them stop?
We have a risk of pandemics built into our ecology on this planet. So far as we know. all the pandemics we've had historically were from nature. If this one wasn't, it was very likely the first one like that. We know up to this point that there have been lab leaks that did cause deaths, you can go read the wikipedia page about it. Those deaths are dwarfed by almost any other cause of death you can think of.
What would you like to see happen?
Right, but specifically WIH and GoF though? Your argument seems to be "science is mostly awesome, even science with bad side effects, therefore this specific scientific endeavor with bad side effects is awesome".
Can you point to one concrete way in which having GoF research helped us significantly in handling COVID-19? If not, then you better be damn sure lab leak not only didn't happen but is extremely unlikely to happen (almost certainly false).
What would one like to see happen? The possibility that GoF viral studies are strongly net negative seriously evaluated and if not unlikely enough, GoF studies stopped or much more tightly controlled.
One would also like the people in those "smoke-filled rooms" with their superior expertise or whatever not to engage in mass misinformation. If they do because they think that accurate information might lead to just the results above - restrictions on their research - who am I to argue with their expert opinion?.. Let's consider having those restrictions on their research.
Because I agree with what you're saying, I think I potentially misread the intent of DavidK74 and heard him asking a broader question like what's the use of research that happens in labs that have diseases and infectious agents in them.
I also agree with the need for oversight and accountability for research, absolutely. I just don't want that to happen via general public opinion. I can't remember but I think the original thread of this discussion was off a side comment Scott made about public opinion.
I think it's good to have lots of arguing with expert opinion, including multi-disciplinarily (if that's a word).
I'm not in this field, so I can't claim any deep knowledge. But yes, clearly some parts of the field do do that. I think they did indeed predict more SARS-like outbreaks following SARS in 2003, and I'm sure did some work to try to work against them (those models that they kept wheeling out early in the Covid pandemic were created long before it struck). I think the work on vaccines made a big difference in the Covid epidemic.
I was really making a more general point, that science and technology don't advance through democratic processes. Even the most perfect democratic process couldn't actually help scientists uncover truth more quickly. And real democratic processes are not perfect; they are political. If science starts worrying too much about what the public thinks, then it just opens itself up to political meddling.
FWIW I asked people with expertise in domains somewhat relevant to mRNA vaccines and their opinion was that no, prior research in institutes such as NIH did not materially help in designing those vaccines. I'm not an expert - I could have misunderstood. They're not literally the team/ teams involved in various stages - they could be wrong. Other objections might exist.
But it's not obvious that the burden of proof is not on you with the claim that they "clearly" helped. Said burden could be met of course :)
For the more general point - 60% doesn't mean the dumbest 60%. If so many perfectly intelligent and qualified people strongly disagree with you - I do think it's your duty to take care of the messaging and convincing. As opposed to, say, deliberately and consciously lying.
I guess I'm willing to believe that claim about the development of mRNA vaccines. But I don't think it's relevant to the point I was making?
I mean. The debate about usefulness of oversight over those "smoke-filled rooms" has to take consequences into account. If specifically GoF research has strictly and severely negative net value for society, would you expect this evaluation to be made, published and pushed by experts in the field? We know the answer to that. They explicitly lied about this.
If this is a line of research contaminated (har har) by conflict of interest, then public attention to and comprehension of key arguments seems essential.
Politics and meddling come in different forms. One is what you alluded to. Another is endemic to "smoke-filled rooms" - and against this one, just like for some viruses, sunlight is the best disinfectant.
I think there's two different issues here. My understanding is that a lot of gain of function research is basically about bioweapons. If that's right, then shining sunlight on virologists isn't going to make any difference at all. Virologists aren't the once who take decisions in bioweapons research. It's the army guys. And they don't accept public scrutiny.
My point was about what the field of virology should be doing right now. I assume that their research paths have been affected by Covid. Scott was suggesting that they should get out there and debate the public to make the case that it probably wasn't the lab leak. I think they have good reason not to do that: if they go out and start debating these points, they only risk increased political interference in their research decisions.
That increased political interference won't make any difference at all to gain-of-function research, which will or will not carry on under strict secrecy at the whim of the army guys. But political interference could make a difference to the work of mainstream virologists, particularly if public pressure ends up pushing grant money towards lab leak research projects. That seems to me by far the most likely political outcome. It wouldn't achieve the sunlight that you want (which is mostly achieved through open science publication, anyway); and it could send research up a bunch of blind alleys.
I think decisions about what research to fund needs to be made by people who are familiar with the domain of research in question. I don't think we need to think of this as smoke-filled backrooms. I don't want the public deciding which strand of breast cancer research shows most promise. I don't want decisions about research into the rise in colon cancer in young people being made by public opinion about whether genetics, environment, or having gotten the MMR vaccine is the more important line of research.
That is true under at least minimal assumptions of benevolence and lack of conflict of interest. Those aren't necessarily warranted in this case.
I'm not sure what you're proposing?
That this particular field of research has done much to lose its credibility to do what's right, probably doesn't have much in the way of practical benefits to show for it, and thus decisions regarding it should have a greater outside involvement. Which, in turn, requires a greater understanding of the underlying issues. Which, in turn, means at least that defending and explaining their point-of-view is the duty of scientists in the field, rather than shrugging at the deplorable ignorance of everyone who disagrees with their estimate of lab leak probability.
That all makes sense to me.
This is not remotely my field. I hear you speaking about arrogance in a sub-field that has questionable value. I can confirm I see this over in my corner of psychology.
This is not right. We can't assume *perfect* benevolence and *perfect* lack of conflict of interest. They're human, after all. But that isn't a minimal assumption. It's a strawman level assumption. Minimal benevolence and minimal lack of conflict do seem warranted - by, for example, all that paperwork from WIV that people have pored through. That paperwork exists because even in nasty old China, scientists who may or may not be interested in creating biological weapons still observe basic safety protocols. That shows minimal benevolence to the rest of us.
We know safety protocols were not WIH's strongest suite.
And you seem to be ignoring the whole "let's lie about any chance of lab leak, a lot" part.
My whole theme here is that information looks different inside a specialist field and outside a specialist field. I'm a teacher, and I "lie" to the children I teach all the time. I do it because they're not yet ready to process the full complexity of English grammar yet. I would expect virology experts to say different things to people inside the field than they do to people outside the field. So when we observe them doing that, I don't really see it as a "lie". For example, it would seem perfectly reasonable to me if they chose to communicate to the media the *most* likely origin story as understood by their field, i.e. wild animal origin; while internally discussing the less likely but more dramatic possibility of a lab leak. That doesn't sound mendacious, it sounds like normal professional behaviour.
But I could be wrong about that. Perhaps a significant number of virologists (not all, remember they're not a monolith) behaved badly in the immediate aftermath of the outbreak. Even if they did, that absolutely does not invalidate the whole field of virology, or suggest that virologists need to be put under greater public scrutiny. You'll recall that the president of the USA at that time suggested drinking bleach. This does not imply that all presidents should henceforth be subject to stronger rules on what they can and can't say.
The lying itself doesn't invalidate the field. The possibility that the field is responsible for COVID-19 does - unless there's proof of some benefits to outweigh this possibility. Which is why I brought up the field's questionable contribution.
The lying invalidates the a-priori trust in the decision-making that the field enjoyed beforehand. If there's a real chance the field is a net negative, and I know for sure that such possibilities are exactly what you might be lying about, you don't get to tell me "oh trust me I'm a scientist" without so much as the courtesy to even show up to debate something that tons of intelligent people disagree about.
Two more general points.
Firstly, I think you're under-playing the role and duty of science in education and enlightenment of society. If the origins of such a world-changing event are of interest to the public, and the opinions of the public (including the generally well-educated public) strongly disagree with scientific consensus - it is a failure.
Secondly, I think you're overestimating scientists. I did a PhD (in an entirely unrelated field) and am married to a post-doctoral researcher in a biology-flavored field (involving RNA, as it happens) and I had many opportunities to see how the sausage is made. It's not pretty. And it's not pretty exactly in a way pertinent to our discussion. Ego, desperate need to publish or to secure a grant at the expense of quality, integrity, basic decency. I absolutely do not believe that their internal and opaque communications are more likely to be truth-seeking and productive than anything exposed to broader scrutiny.
Some aspects of modern science are profoundly broken, and there's every reason to believe that viral GoF is not one of the healthier ones. So when there's a reasonable debate about its usefulness, they do not get to go "please play quietly in the corner while we brilliant grown-ups decide matters". They lost this right, if ever they had it. They get to show up and at least explain.
I'm against using a text format for future long-form debates like this.
I watched this debate an a second monitor while doing other non-cognitively-demanding things like working on design projects and playing video games- and I enjoyed it a lot more than the podcasts and audiobooks I usually use for that purpose. In contrast to those, nothing about the debate felt like dull filler, and yet, unlike most really demanding books, it was possible to miss bits here and there without losing the overall thread of the argument. It's almost a perfect format for content to listen to while slightly distracted, and I think that for that reason, debates like this can provide a ton of value to people beyond just shifting their opinions on the topic.
Granted, a text format would probably result in slightly better arguments- though I don't buy that Peter ambushing Saar with new claims was a major factor in his loss; the lab leak side had plenty of time to research claims between debates, and plenty of time to respond later. A book-length written debate, however, is something I expect almost nobody would be willing to read, since it doesn't provide that extra value as good content to consume while doing other things. It would also require a lot more effort on the part of the debators, so I'd expect there to be both fewer long-form debates and much less engagement- and I don't think a slight bump in quality would at all make up for that loss of value.
Oh interesting argument, I like that. You're saying it's worth losing some potential accuracy in outcome on the issue in question in exchange for the public educational value of it?
I don't think very many people are ever going to watch this kind of very long form debate and so I have a preference for thinking of it as a kind of science project on its own and that maximum accuracy is better for that. And then the NYT or Nature or bloggers or whoever can talk about how this very sophisticated well-run debate with very knowledgeable people happened and they concluded X instead of Y, and they basically do a book review of it.
Oh wow, I guess I really am an elitist, huh?
Education, definitely, but also entertainment- which is a less high-status sort of value, but still real value.
Well-produced podcasts and hours-long video essays will often be seen by very large numbers of people, even when they dive pretty deep into technical questions. Dwarkesh Patel's two-hour-long interview with the Anthropic CEO, for instance, has over two hundred thousand views, and wasn't much less technical than this debate. I also found it a bit less engaging- while a good interviewer will challenge an guest's ideas to some degree, they'll never do so as thoroughly as an opponent in a debate. Ideas are not only more relevant, but also more interesting when you have good reason to trust that they're not being presented misleadingly- and while in a polite conversation, everything has to be taken with a grain of salt for that reason, misleading statements in a long debate tend not to survive the response.
I have a feeling that the low view count of this debate has more to do with people not being familiar enough with the format to think of adding it to their usual rotation of long-form audio content- plus maybe the fact that it's usually been described as a single 18-hour debate rather than a series of 1-3 hour debates. If this sort of thing was a lot more common, I think we'd see some series with far higher view counts.
You create a compelling image of this, I'm sold.
Agree and I think Saar should take credit for coming up with a compelling format for debating complex issues like the origin of COVID. True he lost on this particular issue but the outcome was consistent with the majority view of experts in the field based on recent polling.
If there were a series of debates like this I'd expect Rootclaim to win some and lose others but either way it's much more compelling and will reach a wider audience when you get to actually watch the two sides debate.
Fantastic post - I've shifted my opinion quite a bit in response.
One thing that sticks for me is this: it didn't take long to find the civets responsible for the earlier SARS outbreaks. Why has this proved so challenging here?
The troll in me wants the answer to be that it was zoonosis that occurred in a lab, and then it leaked out of the lab.
"The troll in me wants the answer to be that it was zoonosis that occurred in a lab, and then it leaked out of the lab."
I've thought about that, too. If that happened, what would we call it - zoonotic or lab leak?
And what if the virus evolved through inadvertent natural passaging in the lab? Lots of weird animals in cages, lots of humans around with not much in the way of safety precautions, plus thousands virus samples collected from all over the place being handled. Sounds like a real witches brew.
Thus, a natural emergence, in the sense that it wasn't the result of any deliberate human manipulation. But also something that would not have happened if the lab hadn't been there doing the types of things they did.
I hope it turns out they were studying raccoon-dog behavior in a psych lab and got bitten.
The Chinese government prevented all research by outside investigators for a year, which made it very hard either way.
I'm a bit surprised that they didn't consider the zoonosis-in-a-lab hypothesis, because that's the one that I had thought was most probable before I went back to the wet market hypothesis.
Thank you for this excellent post. It's one of the best things I've read on the internet ever.
I remain in the Lab Leak camp, where I started. I'm of course not 100% certain, and I don't think it's possible evidence would ever emerge that would make me 100% certain. I really don't like people who claim Lab Leak is a "conspiracy theory" and am inclined to discount everything they say.
A question:
Should we accept claims by the Chinese Government and technical/academic/policy experts in this field at face value? Aren't they all hopelessly conflicted? Even if what they say is logically sound, aren't there myriad ways in which they could be fudging the analysis a bit, even unknowingly? What value have researchers doing gain-of-function work on viruses provided to the world? What is the likelihood someone in the field is going to publish findings saying the risk reward in their field just isn't worth it and all their skills are worthless because the work should be shut down?
Another question:
What do those maps of the outbreak in the wet market purport to show? The location where someone was exposed to COVID (how can that be known)? Some point that a person who got infected passed through (what about all the other locations they visited)? Whatever they claim to be showing, how accurate could their measure be? Based on interviews? How hard would it be to fudge the data a bit - for example, an infected person tells a researcher all the places they've been in the last week, and of the dozens of locations the researcher chooses the wet market to record.
I agree with you about this: " I really don't like people who claim Lab Leak is a "conspiracy theory" and am inclined to discount everything they say."
First of all- awesome post! I loved all of this. Saar for setting up this challenge, Peter for taking him up on it, the judges for being fair, and you for summarizing the results- all great stuff.
As for the weird coincidence of the Wuhan Institute of Virology being located so close to the epicenter of the breakout- is it? Everyone knew that wet markets were a potential risk center for spreading diseases. See for example: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7141584/ about the 2003 Avian Flue outbreak. That's the same year that China approved construction of a biosafety level 4 lab at WIV.
They built the lab there because it was a good place to gather research data, and maybe also because the local government was strongly interested in that sort of research to protect themselves. It's the same reason that the "Center for the Study of Active Volcanoes" is in Hawaii, the "Earthquake Engineering Research Institute" is in California, and the "Hurricane Resilience Research Institute" is in Houston. Of *course* they are there where the threat is greatest. And gain-of-function research was just the most effective means of researching that kind of virus, if you want to research it before it actually starts spreading.
You might ask, if they knew the threat of wet markets, why didn't they just shut them down? My understanding is that China doesn't have that kind of state power. There are too many poor, rural people who really want to eat meat, and still can't afford ranch-raised animals. If you try and shut down the wet markets they'll just start selling them in secret someplace even worse.
seems like we should put some thought into hypotheses that prevent the coincidence
could Chinese authorities manipulate the evidence to have a wet market excuse?
maybe the raccoon-dogs intentionally sent their trojan dog to Wuhan to frame the virologists
Nitpick on that poll asking Americans if they think lab leak is true. (Before this) I'd have had to answer this as "probably true" but my probably was that I was at like 60% lab leak and 40% zoonosis. 40% chance of zoonosis is still pretty dang high and doesn't really seem that indicative of acceptance of the bad epistemic edifice that accreted around lab leak. I mean 30%ish saying it's "definitely true" is a problem but "probably true" seems to be doing a lot of work.
Also, good post. Quite markedly shifted my opinion.
Another looooooong series of posts that markedly shifted my opinion on a controversy with a lot of noise recently was this series:https://radleybalko.substack.com/p/the-retconning-of-george-floyd
*Edit* Also, both these posts and that parliamentary procedure kid that's been mentioned so much recently has made me think that the best model for thinking we have isn't heuristics or Bayesian reasoning or the wisdom of crowds. It's probably just smart guy who does like a ludicrously unreasonable amount of homework.
This has been my thinking too: " It's probably just smart guy who does like a ludicrously unreasonable amount of homework."
Thank you for linking to the Radley Balko piece. I just read the whole thing and am so glad to know it's there.
I think this quote stood out to me as off model:
> Zero COVID era Chinese outbreaks were concentrated in wet markets because they received infected animal products. We know why there was an outbreak in the Xinfadi Market in Beijing: it was because the seafood stall got frozen fish from some non-Zero-COVID country, the fish had COVID particles on it, and the vendor got infected and spread it to everyone else. Something like this is true for the other Chinese wet market based outbreaks we know about it. So this makes the opposite point you think it does: wet markets start outbreaks because there are infected goods being sold there. Then the virus spreads through the wet market at a completely normal rate.
I definitely know that China ascribed COVID outbreaks to things like "the person in the plane on the previous flight left covid residue". China was desperate to explain every outbreak so they could continue attempting to achieve a horrible and impossible goal. I don't actually believe that is how the virus was spreading? In particular, surface contact spread and the idea that you should wet wipe your boxes quickly fell out of style. I think the ubiquity of spreads getting first observed at wet markets during covid zero is strong evidence that even if covid is zoonotic but first entered the population by like, bats during a vacation, you would still have first observed it in the wet market.
I filled this out with others' hidden, and ended up weighing that even less than Saar did, which probably means I'm overly dismissive.
https://docs.google.com/spreadsheets/d/1IOHp6PINEbqizpbO7liXLEQTsIuNRpqx/edit?usp=sharing&ouid=113816084183231439418&rtpof=true&sd=true
I think that surface contact spread and the idea that you should wipe your boxes very naturally fell out of style, because it's probably several orders of magnitude less likely a mode of transmission than airborne directly from a person. But when you're in covid-zero conditions, and the rest of the world is in full pandemic, then suddenly "several orders of magnitude less likely" is still the only way for anyone to get infected.
Sort of like how, to a first approximation, you can think of a gas stove, or a candle, as basically having zero respiratory issues compared to a wood burning indoor stove. But once you've eliminated the wood burning indoor stove, those last few indoor flames, even if they're very clean, may be your main respiratory risk.
"But when you're in covid-zero conditions... then suddenly "several orders of magnitude less likely" is still the only way for anyone to get infected."
Are there any clearly documented examples by a trustworthy source (e.g. not the CCP, i.e. in another country) of infection due to residue on frozen package after international transport? Just off the top of my head, Australia, New Zealand and Taiwan all had "zero-covid" regimes for an extended period and had multiple outbreaks. None afaik was due to frozen packages. They were all "failures in the quarantine process" of one sort or another. It's pretty suspicious that China would be the main source for "spread driven by imported frozen packages" claims given tha is reasonable to t this is the sort of thing that perfectly aligns with their desire to shift blame for Covid onto someone else (even more narrowly, ignoring that motive, "our quarantine of incoming travellers failed" is much more egg on face than "we were tricked by those dastardly frozen food packages coming from those imperialist bastards in XYZ").
And thus, like Robert Kennedy, I think it's reasonable to infer from this that large public markets are probably decent locations for superspreading. (Indoor location with a lot of foot traffic is a prime potential site for a virus outbreak - doesn't seem shocking).
Good point! The Australia and New Zealand comparisons seem important. If their outbreaks were all traceable to quarantine failures, then that’s pretty good evidence.
Thanks. To be clear / maintain epistemic humility,
(a) I may be misremembering - every outbreak I know of is either believed to be due to / was traced to some quarantine infection/escape, but my memory could be faulty and my recollection of outbreaks is not exhaustive. (e.g. worker in quarantine hotel caught it from infected quarantined traveller, then took it home with them etc).
(b) logically there should be outbreaks that happened that the authorities never noticed (e.g. one person is infected by some source, they never infect anyone else - which happens >50% of the time - so the outbreak dies before it is noticed), about which we naturally know little
(c) there was an outbreak in New Zealand that initially had some speculation about imported frozen food transmission (if I understand correctly, justified on "well, the Chinese are claiming this can happen), but this was ruled out within a few days.
I hope this isn't a derail: In one of the previous ACX posts' comments, someone posted a link to a page explaining the case for gain-of-function research. I'm embarrassed that I didn't either follow the link or bookmark it. Does anyone have that link? Many Thanks!
Way back in 2020 there was an article https://www.independentsciencenews.org/commentaries/a-proposed-origin-for-sars-cov-2-and-the-covid-19-pandemic/ , which I read after George Church tweeted it (!) (without comment or explanation). Their proposal (they call it "Mojiang Miner Passage" theory) in brief was that it WAS a lab leak but NOT gain-of-function. Rather, in April 2012, six workers in a "Mojiang mine fell ill from a mystery illness while removing bat faeces. Three of the six subsequently died." Their symptoms were a perfect match to COVID, and two were very sick for more than four months.
The proposal is that the virus spent those four months adapting to life in human lungs, including (presumably) evolving the furin cleavage site. And then (this is also well-documented) samples from these miners were sent to WIV. The proposed theory is that those samples were put in a freezer at WIV for a few years while WIV was constructing some new lab facilities, and then in 2019 researchers pulled out those samples for study and infected themselves.
I like that theory! I’ve like it ever since 2020! It seems to explain many of the contradictions brought up by both sides of this debate—it’s compaible with Saar’s claim that the furin cleavage site is very different from what’s in nature and seems specifically adapted to humans, but it’s also compatible with Peter’s claim that the furin cleavage site looks weird and evolved. It’s compatible with Saar’s claim that WIV is suspiciously close to the source of the outbreak, but it’s also compatible with Peter’s claim that WIV might not have been set up to do serious GoF studies. It’s compatible with the data on differences between COVID and other previously-known viruses. Etc. (As I understand it.)
Old as this theory is, the authors are still pushing it and they claim that it’s consistent with all the evidence that’s come out since then (see https://jonathanlatham.net/ ). But I’m sure not remotely an expert, and would be interested if anyone has opinions about this. I’m still confused why it’s never been much discussed.
It looks like it’s not even a coronavirus ? https://en.wikipedia.org/wiki/M%C3%B2ji%C4%81ng_virus
Good find! The wiki article is the product of raging 8000-word debate on the talk page: https://en.wikipedia.org/wiki/Talk:M%C3%B2ji%C4%81ng_virus
Update: I read the talk page! My understanding is that the miners got sick from Disease Agent A, and then later someone took a sample from the same cave (NOT from the miners) and found Disease Agent B. The Wikipedia article you found is about B, not A.
It’s POSSIBLE that A=B, but neither the Wikipedia article nor anyone else is claiming that A=B. There’s actually some evidence that A≠B, but that evidence didn’t make it into the main wiki article, partly because the WIV master’s thesis discussing the miners was declared not to be a “reliable source” per Wikipedia standards, and partly because the anonymous person who created the article was very aggressively opposed to providing fodder to what they called “baseless sinophobic paranoia” “conspiracy theories” 🙄
I think the pertinent article linked by wikipedia is https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9744119/, which seems to be an explicit rebuttal to that theory.
Thanks! That link isn’t much evidence though, right? Like if we’re gonna do the Bayes thing, I dunno, likelihood ratio 1.1 maybe? Like, if a biologist is accused of stealing a car, and then the biologist announces that he did his own DNA analysis in his own lab and found results that exonerate himself … well, we haven’t learned very much from that. :-P
I’m much more swayed by e.g. the fact that RaTG13 is from the Mojiang mine.
Let me know if I’m misunderstanding or if you find anything else :)
Yes, they could easily lie here. You have to decide for yourself what weight to put to those words. Johnathan Latham explicitly says they’re lying in https://jonathanlatham.net/faucis-covid-origin-swat-team-versus-the-mojiang-miner-passage-theory/ ("Other criticisms of the MMP theory"). I’m somewhat sympathetic to that : "they died of fungus" is clearly nonsense and can easily interpreted as the researchers saying to us "we have to lie to you but we’re not happy about that so we are going to pick the least believable lie".
The big gap in the MMP theory is BANAL-*. I hope you’ll forgive me if I can't find the exact quote, but in the initial MMP theory Johnathan explictly said something in the line of "one thing that will be a big problem for the MMP theory is if the virus hunt yields a genetically closer ancestor to SARS-CoV-2 than RaTG13". And we found one : BANAL-52. Johnathan handwave it by saying "we found it in north Laos, and north Laos is geographically the same thing as Mojiang".
But we don’t know where exactly the mine is in Mojiang county (eyeballing google maps, it’s 300 km from north Mojiang to south Mojiang). More problematic, the paper that announced the discovery of BANAL-52 never said it was found in north Laos ; indeed they say that all BANAL clades were found in the same site, "site 1", which was in east Laos, close to (again, eyeballing the map) Pak Lay, which, again by eyeballing Google Maps, is close to 800 km to Mojiang county.
This is a big problem for MMP theory, I haven’t seen a good answer to that.
But to be fair, it’s a big problem for the Wuhan wet market theory too. It’s mostly a point for the Chinese official theory, "it comes out of China".
This was incredibly informative. Thanks to the participants for doing it, and to Scott for summarizing and publicizing it. I hadn't thought about COVID origins in a while, but I would have told you it was in the tossup zone, and digging into this a bit puts me closer to Scott's 90-10.
I wish this was a highly replicable method for illuminating tricky topics, but I kind of suspect that the conditions here (two reasonable and sincere people who believed an impartial jury of smart, curious neutral observers would be won over to their side, and who were able to agree on who constituted a smart curious neutral observer) are tough to arrive at. I hope I'm wrong!
"some rando who nobody had ever heard of"
Not to be too facetious, but perhaps Rootclaim should have calculated the probability of the existence of a debate savant before building an attractor for one.
The big problem is that we still have very few details about all this. How is that possible after all these years?
The Chinese authorities haven't been very transparent. Other governments haven't demanded transparency from China. Why not? They just accept that China doesn't want to give any details.
The pandemic was a global catastrophe. It killed a ton of people and caused enormous economic losses.
The people deserve to know how it all happened. Secrecy is unacceptable.
The main reason why I'm leaning towards the lab leak is the severe censorship about the topic. That came from the individuals who strongly supported the zoonosis hypothesis. Usually people who desperately want to hide something have something to hide.
I think most leaders know it was a lab leak. They just don’t want to deal with the political shitshow that will come out of making it official. It has the potential to disrupt the fickle relations with China and as a result, significant amount of global trade
Whether or not it's zoonosis or a lab leak doesn't actually matter. Either way the CCP is indirectly responsible. The CCP will never admit wrongdoing as it has zero incentive to do so. Every other country is obviously upset about this, but not enough to stop doing buisness with China, as that would put them at a major disadvantage. While COVID did cause damage, and it damaged the entire world pretty much equally, and thus most countries are not in a worse position than they were before, relatively speaking. Economies can be restored, and human lives can be replaced.
As for the gain of function research... do you really think that even if there was conclusive evidence in favor of a lab leak, it would change anything? Bioresearch is just one of those things where the positive consequences are localized and the negative consequences are not, and thus there is no incentive for self-regulation. Any attempts at globally regulating it would be just as futile as nuclear disarmament.
Every government demanded transparency from China. They refused. What more do you want?
China was very interested in covering up the wet market aspect of it for a long time, and only allowed investigators to come a year later.
Nobody really demanded anything. They just asked nicely and China said no. And that was it. There was no real pressure against China to provide information.
Everyone else was happy that China didn't share any details. That is a bit suspicious.
No one was happy that China didn’t share any details. I’m not sure where you were getting the idea that they were. The WHO did try to put pressure on China, but China has more power over the WHO than vice versa, so they just couldn’t put all that much pressure on them.
Nobody put any REAL pressure towards China or any other suspicious party like EcoHealth Alliance.
WHO put extremely weak pressure on China, it's like they didn't even try.
I don't remeber seeing any politician being genuinely enraged for China hiding so much important information. Given that the negative effect of the pandemic was enormous, I would have expected to see lots of powerful people saying "you must show us everything you have on this topic, or else...". But nobody did that.
What comes after “or else…”? What pressure could the WHO put on China during the pandemic, when they needed Chinese cooperation especially?
Of course it's not only WHO but all the governments.
The pandemic had an enormous negative effect on the population of the world. People deserve to know what happened. The governments who represent the people should find out what happened but for some reason they don't even try.
If it's not enough to ask nicely, it's not hard to come up with harder ways to create pressure.
For example, setting some sanctions would definitely help. Similar to which are in effect against Russia now. They would be even more effective against China because it's so depending on imports from other countries. A few months of blocked fuel imports would give a nice incentive to start being more open.
“Hey, this bat we just collected on one of periodic field trips to yunnan had an interesting viral strain with some really interesting structure around how it’s furin site works and implications for protein folding”
“We should probably research it, is there any grant money we can use?”
“I think the Americans just announced some new initiatives?”
“Great, let’s apply. Write a proposal that says we will investigate Furin site placement and it’s impact on virology.”
….
“Did we get the funding?”
“No”
“But the grant reviewers didn’t know about the bat, right? Just that we proposed doing the furin site stuff?”
“Yea, we’ve changed the names of the ones we found and generally obfuscated that we have this specific one in house”
“Should we just do the research anyway?”
“… … … yea, probably, could be interesting”
“We haven’t really done this before, and maybe we don’t have the right safety procedures and protocols, specifically to make sure it doesn’t escape?”
“No. But this particular strain can’t even last that long in the lab on its own. So what’s the worst that could happen”
Five out of the six organizations on the DEFUSE proposal are American.
I have no strong opinion about the Lab Leak hypothesis. I have a very strong opinion that Bayesian reasoning does not, in practice, help with deciding such matters.
Early in the post, Scott offers the following. After reading the entire long post, I believe that it is, in fact, the correct overall summary:
> But the joke goes that you do Bayesian reasoning by doing normal reasoning while muttering “Bayes, Bayes, Bayes” under your breath. Nobody - not the statisticians, not Nate Silver, certainly not me - tries to do full Bayesian reasoning on fuzzy real-world problems. They’d be too hard to model. You’d make some philosophical mistake converting the situation into numbers, then end up much worse off than if you’d tried normal human intuition.
It's nice to be able to say so. Try saying that on LessWrong.
A very accurate observation of Scott there.
Well I’m still convinced it was a lab leak but at least now I know I’m probably wrong.
What a great reply!
Disclaimer: I have not watched the full videos. Apologies if this point is addressed at length.
To what extent are both parties relying on CCP-curated information for their data on the early cases? Was there a discussion of how (un) trustworthy that information might be?
I think at least some of the consideration is that whatever curation was going on in December 2019 is unlikely to have any evidence about the actual origin, since the CCP had no reason to suspect it was related to the virology institute.
If you assign 1/10,000 probability of that particular wet market being the location of the first major outbreak conditional on the lab leak hypothesis, you should also calculate the probability of that particular location under the zoonotic origin hypothesis. The population of Wuhan is <1% of the population of China (maybe the whole world is a better reference class, though rural areas and developed countries are going to have lower risk), for example, and that’s before you account for the fact that zoonotic origin could also cause major outbreaks at other locations within each city. All in all I’m not sure “covid began to break out at this particular wet market near WIV” is terribly strong evidence one way or the other.
Another piece of evidence I don’t see discussed here is the lack of openness from China, and relatedly the lack of evidence of a zoonotic source. If China had information against the lab leak hypothesis they would likely release it, so the relative lack of information we have is more likely given a lab leak imo.
For all the talk of Furin cleavage sites, there's one thing that confuses me. Could we have had a (far less deadly, probably not even noticed) coronavirus pandemic if there was no Furin cleavage site? My understanding of gain of function here is that there are multiple proteases that can do the cleaving even without Furin, but Furin does the job faster, and therefore sometimes outruns the immune system in a way COVID wouldn't do without that extra cleavage site. But that raises the question of why we got an outbreak of a virus with that Furin cleavage site, instead of a hypothetical precursor that lacks it. The existence of a Furin cleavage site is considered unusual, so for Zoonosis would you expect a precursor virus without it to be far more likely to cross over into humans first, and therefore prevent COVID from being a major issue in the first place?
Shi's team, in 2018, published a paper reporting serological evidence of bat-human coronavirus spillover around the caves in Yunnan province. Doesn't seem like this strain/these strains would've been infective enough to kick off a pandemic, but if this sort of transmission was happening regularly enough for long enough, then the people around these caves could've definitely served as a natural "gain-of-function" experiment.
Paper here: https://link.springer.com/article/10.1007/s12250-018-0012-7
Sometimes you can tell one of Scott's posts is going to be a banger just from the quality of the pun in the opening section header alone
Unfortunately it took me 16 hours to read this article 🥺
The notorious fact is that there were only 3 labs in the world at that time doing gain of function technology in the world...2 of those were im US and 3rd one in Wuhan..GOF was banned by Obama administration so the only destination to outsource technology was wuhan lab...🤗🤭 even famous defuse grant for the almost same virus creation by GOF was proposed and refused for wuhan lab by darpa im 2018...immpossible coincidencenes happen😅🤣
"Still, it’s awkward to use “conspiracy theory” as an insult when the conspiracies were real. Maybe a better question is whether lab leak is “pseudoscience”.
The argument against: lots of smart people and experts believed it was a lab leak. There were all those virologists giving 50-50 odds in their internal conversations. Even Peter says he started out leaning lab leak, back in 2021 when everyone was talking about it.
The argument in favor: since 2021, experts (and Peter) have shifted pretty far in favor of zoonosis. They’ve been convinced by new work - the identification of early cases, the wet market surveys, the genetic analysis. "
I have admittedly not looked into the details, and I cannot claim to truly know how to handle filtered evidence, but:
If the experts (mostly) all agreed to have a conspiracy pushing for A, and then all the evidence that turned up after that pointed to A, this really doesn't make me feel like I have to raise my probability of A all that much. For all *I* know for every published result supporting A there are 2 unpublished results supporting B, or claims that have not been investigated because experts (correctly) believe they would turn out to support B.
One issue I have with the opening presentation in the first video (I have watched 38 minutes so far) is that he lists, for example, about a dozen different theories about how the virus could have been made, and his slide says, "it's not possible for all these theories to be correct." So what? If there's one true explanation for something, the fact that 11 other people came up with a false explanation for it doesn't diminish the probability of the first explanation being true. Otherwise, I could make the probability of any true idea go down just by creating a bunch of false theories.
What you should do it reserve some probability mass for the theories you have never heard of, or that no one has even thought of. The probabilities of the set of theories that happen to exist at time T don't have to sum to 1, in general.
This is a good and important point, thanks.
Skipped the whole lot and am working purely from Scott's post-transcript summary.
>So the real p(wet market|lab leak) isn’t the 1/10,000 chance a pandemic arising in a random place hits the wet market, but the (higher?) probability that there’s something wrong with Peter’s argument.
This looks like an excuse to assign whatever numbers you like to any point. If you're readjusting part of your stats, surely you should be readjusting all the other parts with the same readjustment metric, and thus reach a conclusion indistinguishable from the original formula.
It does sound like Saar lost mostly because he got caught flat-footed, but considering he set up the system it still looks bad for his methodology. Surely he used it to calculate a proper debate setup beforehand, and it underperformed.
I learned more about this debate and the arguments in favor of each side in the hour and a half reading this post than I did in any previous discussion of these arguments. Thank you for the rich summary, and while I agree with your arguments about why this kind of thing probably can't happen very often, I really wish we'd see more of these kinds of rich debates
Same!
A systemized version of this plus community notes is my basic model for how all news should work. But with reputation scoring in there to smooth out some of these transaction costs for actors who have proven themselves credible.
“Dumb” question that I did not see answered anywhere, and not mentioned much at all even in these comments to date: does “lab leak” also incorporate the possibility that it was created in a lab and *deliberately* circulated in the Chinese public?
And if done deliberately is an option within “lab leak”, doesn’t that throw off the odds by several OOMs, given that for someone deliberately wanting the virus out there, but wanting to cover their tracks a bit, “releasing” it in a wet market that had all those animals would be a fairly high probability thing to do (in addition to being a very good way to ensure spread)?
Is deliberately put out into the world its own separate possibility, or is it incorporated into “lab leak”? And if the latter, what is faulty about my reasoning that it throws off the supposed unlikelihood of the outbreaks being centered in the wet market by many OOMs?
...Because that would be completely insane? Not the possibility of it happening, but... There is no potential rational justification for doing that, ever. It would just be plain and simple bioterrorism. Except it wouldn't even accomplish anything of value except kill a lot of people.
Like, I'm sure the probability isn't zero, but... people generally assume that the people involved in the situation aren't completely insane. And also it would be incorporated into the "lab leak" hypothesis, since the existence of the lab would still be responsible for the event.
Really? Seems to me all it would take would be a single Chinese bureaucrat - high, middle or even low-ranking (the last of which seems admittedly less likely) - who decided the benefits of the knowledge gleaned from human trials would outweigh the risks. And merely not understanding how high the risks were. The Nazis had “rational justification” for things we people with decent western liberal morals thought were far more ‘insane’.
And if this case is in fact covered within “lab leak” then it changes completely the logic that the wet market would be a 1 in 10,000 chance to something more on the order of 1 in 10…
...That would possibly be the stupidest, most psychotic way of doing human trials possible. And if the Chinese government was responsible, why the fuck would they spread it in their own country?? Even if they were completely stupid, spreading it in another country would at least have a non-zero chance of avoiding the virus spreading back to China. And all of this is irrelevant regardless, because if they just wanted to have human trials, I'm sure they have plenty of political dissidents they would work with.
Surprised to see no mention of Nick Patterson's latest: https://npatterson.substack.com/p/yet-more-on-covid-origins
Likewise
I loved this whole thing, thank you for writing it and I'm glad you enjoyed it!
I have no dog in this hunt and don't feel like any important policy-making we have to do hinges on this outcome. We know we need to keep labs safe and we know we need to deal with zoonotic risks. And I still found it riveting to read, as well as all the comments here. I think this lands in the genre of competency porn and I'm here for it.
A meta-observation: I started today 50-50 on this question. Right after reading it, I was briefly at 90-10 zoonotic. An hour or two later, I'd say it slipped to 75-25, and now I'm landing maybe at 60-40. That's an interesting experience just to watch over the course of a day. I don't know, but I think if I read it all over again, it might goose me back up to 80-20, temporarily.
When I ask myself why looking at the debate as presented swings me a lot but then that regresses over hours, I would summarize that as "mistrust". Mistrust of the institutions that produce information, mistrust of the arguments presented, mistrust of the biasing potential of the debate format. That's my emotional/psychological bias and so the further I get from direct contact with the elements of the debate (which I found persuasive), then that default bias starts to take over more.
While my intuition listens along to the content of the debate, it also has some deeply-held opinions about the trustworthiness (or lack thereof) of experts. I have my hobbyhorse example of this around guidance to parents for what position to put their infants down to sleep -- which has changed three times since my adulthood. It's not that anyone was misleading anyone, but that things that look absolutely true to science in one era turn out to be absolutely wrong in the next era and that's how science proceeds. We don't know what we don't know, and that makes all my reactions to these kinds of competing stories pretty uncertain. And it also makes me marvel at other people's certainty.
I don't know, but I think this might have some relevance for questions people ask about how to persuade people about the importance of large scale risks based on a mish-mash of reasoning and data. Where and how mistrust shows up is a huge issue.
"Where and how mistrust shows up is a huge issue."
This is indeed a core issue. And it can be studied systematically.
> Fourth, for the first time it made me see the coronavirus as one of God’s biggest and funniest jokes. Think about it. Either a zoonotic virus crossed over to humans fifteen miles from the biggest coronavirus laboratory in the Eastern Hemisphere. Or a lab leak virus first rose to public attention right near a raccoon-dog stall in a wet market. Either way is one of the century’s biggest coincidences, designed by some cosmic joker who wanted to keep the debate stayed acrimonious for years to come.
Scott messed up. The lab was built near wet markets because wet markets were a threat, and GOF research probably was conducted there because leaks could be blamed on wet markets.
The lab is fifteen miles of dense cityscape away from the wet market. There are thousands of buildings between the lab and the wet market. There's no way "sit around, wait for something to leak, surely it will surface at the wet market" would be a reasonable strategy.
Also, the lab was not built near wet markets because wet markets were a threat. This is completely confused.
The new notes USRTK obtained through FOIA in January 2024 aren't trivial. I had been pretty much on the fence about the lab leak hypothesis, but the new evidence is so overwhelmingly damning it literally made me nauseous reading it. The notes explicitly propose generating a consensus sequence with an inserted furin site. They propose using six BsmBI sites to assemble the infectious recombinant - confirming a 2023 hypotheses that the spacing of the six BsmBI sites in SARS-CoV-2 had a suspicious spacing. The notes proposed passaging the recombinant virus in human cell lines under BSL2 conditions.
It's a mistake to casually dismiss new smoking guns, all caps or otherwise.
"Why did the older strain start spreading later? Probably the virus crossed from bats into raccoon-dogs on some raccoon-dog farm out in the country. It spread in the raccoon-dogs for a while, racking up mutations, including the (less mutated) Lineage A strain and the (slightly more mutated) Lineage B strain. Then several raccoon-dogs were taken to Wuhan for sale, including one with Lineage A and another with Lineage B. The one with Lineage B passed its virus to humans earlier. Then 3-4 days later, the Lineage A one passed its virus to humans."
So apparently the zoonosis story requires *two separate* zoonotic transmissions, from the same location and within the same week, but from separate viral lineages? Am I understanding this right? And if so, then can someone please explain why both sides are just passing over this claim as a minor aside rather than a devastating hole in the theory?
I am not an expert in the field, at all, but this intuitively strikes me as *really* implausible.
The lab leak hypothesis seems like it either needs two leaks of its own, or a long enough period of spread to produce this mutation while somehow not leading to the explosion of hospitals until late January.
B is only two mutations away from A, and one of those mutations is silent. It could have happened in the same person.
Zhang et al 2024 has new early samples from Shanghai with a bunch of intermediates between A and B, consistent with a single spillover. I think their proposed phylogeny roots at an A0 strain which is one mutation closer than A to the bat viruses.
Out of curiosity, how widespread is the software bug Peter alluded to? If this is not fixed/ improved and the software in question is still widely used, and further if this bug might make such findings more likely, should we believe this latest finding?
I'm genuinely perfectly clueless and ignorant on the topic, so truly don't know.
I haven't seen anything about the bug he refers to, I would be interested to see a reference supporting his claim. I'm sure he's referring to something real but his interpretation seems overconfident.
Ok, I'm guessing he's talking about this line in Pekar et al 2022: "We identified numerous instances of C/C and T/T genomes sharing rare mutations with lineage A or lineage B viruses, often sequenced in the same laboratory, indicating that these intermediate genomes are likely artifacts of contamination or bioinformatics".
This is somewhat more cautious than Peter's "we can dismiss both" so maybe he was referring to something else?
The new paper (it's Lv et al 2024 contrary to my earlier ref, Zhang is the senior author) mentions this in a couple places but don't seem to think it affects their result.
That seems most plausible to me but it blows a huge hole in everybody’s timeline since it means cov2 was spreading around in humans before any of the cases flagged by the authorities. For a disease with a 10+ day incubation period and a relatively low rate of hospitalization though it could have easily been spreading for a month or so around young healthy people before anyone started looking for it.
See responses above at https://www.astralcodexten.com/p/practically-a-book-review-rootclaim/comment/52707601
Side note about the use of "conspiracy theory" in this context. It's got these two meanings, one being "nutty, unfounded belief" and another being "a theory about how people coordinated actions secretly behind the scenes." There's a Venn diagram where flat earthers believe an endless stream of government officials are conspiring to keep the truth from us.
It doesn't seem accurate to me to call the lab leak idea conspiracy theorizing in the "nutty, unfounded belief" sense. Smart people in very recent years, including in this debate, disagree. When you accuse the other side in that situation of "conspiracy theorizing", you're just engaging in ad hominem attacks which add no useful information.
Some lab leak people do believe that there was behind-the-scenes conspiring that happened, but so do some zoonotic supporters. There are more people on the lab leak side who do seem to be more of the nutty, unfounded type of believers. But it's a distraction to look at them, because the evidence in favor of lab leak is not like the evidence in favor of the earth being flat, etc.
Sometimes too the people we call conspiracy theorists are just people who have been slower to catch up to the preponderance of evidence, given that science is a process that unfolds over years. I don't think it's helpful to call people in this category conspiracy theorists.
Strong arguments were made on both sides of this. Peter's arguments look stronger. But as a person who doesn't have their hands on all the raw data, I wouldn't say that he "debunked" the other side. He just presented some other information that was more persuasive.
Anyway, I guess this is a wish for not calling people "conspiracy theorists" as it relates to things scientists themselves are still debating.
The whole Session 2 issue seems irrelevant to me. The simplest lab leak scenario is a virus being collected in the wild, brought to the lab, then escaping in guano on a tech's scrubs or something similar. No need for genetic engineering of the virus at all.
The argument seems to have been "natural virus in wet market" versus "engineered virus from lab" while neglecting other scenarios.
This kind of speaks to what's the relevance of this debate after all, other than all the political heat that's been generated around it.
Why do we really care? (I mean that as a sincere question rather than dismissively)
Zoonotic origin means we need to fund virologists more to be prepared for the next nasty virus to come out of the wild.
Lab leak means we need to put a choke chain on virologists to make them practice better lab hygiene and stop inventing nastier viruses.
So it's an significant impact on policy questions. (Having more virologists practicing to higher standards, alas, is probably not a viable option.)
We already know that both of these are big risks worthy of tons of attention. Lab leaks have already killed lots of people. Pandemics even more.
This is true.
In a sensible world, the uncertainty about whether the virus came through the door or the window would be seen as a good reason to close both the door and the window; in the real political world we live in it's being used as an excuse to leave both the window and the door open.
In an even more ideal world the rest of the world would demand trillions in reparations from the Chinese government.
> Why do we really care?
Aside from the obvious questions about whether we need to improve lab safety, and whether gain-of-function research caused a problem for us (yet)...
And aside from Saar's underlying motive, which appears to be an attempt to implement practical Bayesian reasoning...
Mostly because the discussion picked up political heat. Which warped the "official narrative", in ways that are still relevant. It's important to know which of our official scientific sources will follow along with tribal groupthink, and which will base their decisions on data. (Or rather, where on the spectrum each individual source falls.) And so this leads to a question of what the data actually are: what did we know, and when did we know it, and what kind of mistakes were made along the way, and why were those mistakes made, and what if anything can we do about it in the future.
Thank you, nicely said. I find my opinions about all this swinging around. Which I think mostly means I don't have very strong opinions about it.
Yes, the irony on my part is that I actually do want to "trust the science", but I also am slightly familiar with how the sausage gets made. And when we have scientists who are also public figures, making claims with far too much certainty, that makes me skeptical of everything they say: it's being a public figure and engaging in public relations*, not being a scientist and engaging in science. Give me something with error bars, and explicit limits on our certainty, and I'm much more likely to trust it. It's not a perfect heuristic, but I came to it the hard way. And it does give me a bit of moral high ground, in that if people would just stop trying to score political points and be honest, I'd be on their side.
* i.e., propaganda
"And when we have scientists who are also public figures, making claims with far too much certainty"
...It's a selection effect, at least in my country.
Journalists call around till they find some scientist that says what the journalists want to say themselves but do not have the academic cred to say. And/or till they find some scientist who is willing to say something dramatic or controversial about the topic. Those scientists receive new calls. The ones who offer error bars do not.
I do not blame journalists. Most live and die by clicks these days. Those who hate the system leave journalism, creating a new selection effect where the remaining are those who like to live by clicks.
In the old days, newspapers could print a few long-winded serious articles in each issue, since no-one knew for sure how many - or few - that read them. (The Germans for some strange cultural reason still have some newspapers like that. It also helps to have a potential pool of 82 million readers.) But after online publishing, editors and owners have accurate data on how often such pieces are read. Which has been a death knell to serious science journalism.
...admittedly, I am talking here of the social sciences (broadly defined). The natural sciences is somewhat better placed.
Instead, we have substack and a few other places where the global audience of people who are actually interested in nuanced stuff (with error bars) is large enough that it can fund at least some writers who have nurished a "the truth, whatever it is, shall set you free" attitude to their subject matter.
.
Yeah, it's weird to me that they didn't mention the third hypothesis, which I thought was considered as plausible as either of the other two initially. But I think that Session 2 deals with a lot of evidence that had been taken by some to be extremely strong, and it only supported the gain-of-function version, not the lab-collection-accident version.
"As mentioned earlier, the DEFUSE grant was rejected."
I do not know how many people here are familiar with how the federal grant system (whether NIH or DoD) for biomedicine works, but if one claims that the only way experiments get done is if 1) they are proposed in a grant and 2) that grant is funded, they are either unfamiliar with how the system works (like Peter) or else engaging in intentional misdirection (like the many virologists who have made this claim and certainly know better). As a general rule, the proposed experiments in a biomedical grant fall (often roughly equally) into 4 categories: 1) work you have already done but not published and inot included as preliminary data (you have to include just the right amount of data to justify the proposed experiments/money) ; 2) work you never intend to actually do but which looks good on the application; 3) work you do during the 1-2 years the grant is under review but no decision has been rendered; 4) work you actually will end up doing during the funding period.
That the DEFUSE grant itself was not funded is drastically less important (frankly not important at all) from the perspective of a Bayesian trying to figure out the origin than the fact they proposed making the precise modification that is so striking about Covid and has not been seen in any close relatives. David Baltimore gets this, Nick Patterson gets this, and I think anyone with knowledge of both how these grants work as well as the molecular biology gets this.
I wondered this as well, it seemed like non-evidence to me given how this works.
I agree, several of my friends are in grant-seeking areas and universally this has been the reported experience. Most times your current work is being funded by your last grant, your future work is being funded by your current grant, and you current grant was based on your past work.
Related, a saying by an old colleague: "Collecting the data takes 90 per cent of the available time. Reporting the findings takes the other 90 per cent".
>I do not know how many people here are familiar with how the federal grant system (whether NIH or DoD) for biomedicine works
I am _not_ familiar with it, and I appreciate the comment!
>1) work you have already done but not published and inot included as preliminary data (you have to include just the right amount of data to justify the proposed experiments/money) ;
Now that you mention it, that sounds _very_ familiar from other areas of research as well!
Saar raised the issue of ascertainment bias, but undersold it. There's no question imo: the case search *was* heavily focused on HSM, and the HSM sample search *was* heavily focused on the southwest corner. The Chinese authorities who did the search have been clear on this (Bahry, 2023, Table 1 and Fig. 1).
E.g. we now know that the raccoon dog stall was sampled 10 times, with 5/10 positive; while the pool of blood outside a beef stall in the east wing was only sampled once, with 1/1 positive (Liu et al., 2023; cf. Bahry, 2023, Fig. 1). That, not the racoon dog stall, was the highest positivity ratio. Peter's heatmap was based on a paper by some Western scientists who had no idea how many times each stall had been sampled, and had instead speculated that every sampled stall had been sampled equally (Worobey et al., 2022).* After Jesse Bloom also showed there's no positive association between raccoon dog and SARS-CoV-2 genetic material, nothing implicates them, not even in a God's-joke way (Bloom, 2023).
China CDC director George Gao has also emphasized, when asked by the host of a BBC podcast, that this overwhelming ascertainment bias means you can't assume genuine case clustering around HSM (https://www.bbc.co.uk/sounds/play/m001ng7c, 24:00–25:12).
References
Bahry, D. (2023). Rational discourse on virology and pandemics. mBio 14: e00313-23. https://doi.org/10.1128/mbio.00313-23
Bloom, JD. (2023). Association between SARS-CoV-2 and metagenomic content of samples from the Huanan Seafood Market. Virus Evolution 9: vead050. https://doi.org/10.1093/ve/vead050
Liu, WJ. et al. (2023). Surveillance of SARS-CoV-2 at the Huanan Seafood Market. Nature. https://doi.org/10.1038/s41586-023-06043-2
Worobey, M. et al. (2022). The Huanan Seafood Wholesale Market in Wuhan was the early epicenter of the COVID-19 pandemic. Science 377: 951–959. https://doi.org/10.1126/science.abp8715
*This zoonosis-theory paper is one of many by a prominent coauthor network centered on "Proximal origin" authors Andersen, Holmes, Rambaut, and Garry.
*edit: Peter's figure is the version from that co-author network's preprint (Crits-Christoph et al., 2023, https://zenodo.org/doi/10.5281/zenodo.7754298), not the paper (Worobey et al., 2022). (It also includes the partial metagenomic data that they scooped from Liu, Gao et al., getting in some trouble with GISAID for.) It's still obfuscatory, acting as though the SC2-negative samples weren't *also* concentrated in the sw corner.
Would love to hear Peter's response to this.
I do not believe that the pool of cow's blood in the east side of the market is the source of the covid pandemic based on 100% of its samples (1 out of 1) coming back positive.
Otherwise, I think I adequately answered all the questions about ascertainment bias at the debate.
Nobody's saying it came from the cow's blood (just like Bloom wasn't saying it came from fish). The point is that the Worobey et al. / Crits-Christoph et al. heatmap is an obfuscation, based on their making a baseless assumption that was already known to be baseless and is now known to be false. At present nothing genuinely points to raccoon dogs.
The only adequate answer to the clear, known, overwhelming ascertainment bias in the early case search is to say "Wow, you're right—the ascertainment bias *is* clear, known, and overwhelming, so we can't appeal to apparent clustering as evidence."
(For the HSM samples, now that we know how many samples were taken from each sampled stall, it probably *is* possible to properly redo the Worobey et al. / Crist-Christoph et al. heat map, modifying their code to use the positivity ratio. I haven't done it myself because I'm not good at coding. It could either be done with the measured positivity ratios, e.g. 5/10=50% and 1/1=100%; or using pseudocount smoothing like in my figure, e.g. [5+0.077]/[10+1]≈46% and [1+0.077]/[1+1]≈54%.
Note: Worobey et al. / Crits-Christoph et al. haven't properly redone the heat map either; instead, they've since chosen a new way to automatically downplay the less-sampled stalls away from the southwest corner, by mapping p-values instead of positivity [https://www.biorxiv.org/content/10.1101/2023.09.13.557637v1#comment-6279754825].)
For anyone who might want to try, here's their github for Worobey et al 2022 https://github.com/sars-cov-2-origins/huanan-market
Can you just say why George Gao said they focused too much on the market. If we're looking for strong evidence - the guy in charge of the case search said there was ascertainment bias both within environmental sampling in the market and in the city of Wuhan. Is he just very confused? What is going on here?
Question for Peter: don't you think you're being a little credulous with respect to the early case map, given an obvious and known effort from WHO/China to support anything but LL?
Separately--great job. To go into this with nothing but your own research and bet a good portion of your net worth against some rich guy, and win, is something from a movie. You should be proud for sure.
China very much opposed the wet market hypothesis early on as well. That's why they didn't allow any investigation of the wet market for a long time.
Peter claims the Worobey study's map is from "before ascertainment bias", but I just don't understand how. If the first known case was from the market, and with the knowledge of SARS zoonotic origin, you don't even need to attribute any bad faith to get ascertainment bias around the market. Once the LL hypothesis existed (and I don't really know when it was first proposed, or when the CCP thought it might be a possibility, you also have bad faith to contend with, even if their official story was propaganda about an American lab.
No they didn't. George Gao said it was from the wet market in Jan 2020
What should we make of this study, which found the presence of covid in Brazilian wastewater in late 2019?
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7938741/
There are lots of studies vaguely like this. I think we should be skeptical, because COVID has doubled every 3.5 days in every population it's been exposed to.
Suppose there was a person with COVID in Brazil on November 27th. The first confirmed case in the Americas was January 21st. By January 21st, the one COVID case in Brazil should have grown into ... 55 days divided by 3.5 = 16 doubling times, 2^16 = 65,000 cases. There's no way to miss 65,000 cases of COVID, especially when the next week that's quadrupled to 260,000.
So you would have to argue that COVID fizzled out sometime after November 27 - eg one person had it, and didn't pass it to anyone else. That's possible for one case - maybe that one virus just had bad luck, or the person quarantined. But it quickly because implausible once there's more than 2 or 3 cases.
So the question you have to ask is whether there was some way that there were only three cases of COVID ever in Brazil (until it officially arrived from China), and yet some scientist still got very lucky and found one of those three cases in the wastewater.
I would rather believe they made some kind of error, which also fits with the fact that nobody has any explanation for why a Chinese bat virus should be in Brazil before it's in China.
That's a good guess, and sound reasoning. If covid was in Brazil in November 2019, why isn't it also in the wastewater in other cities around the world? Why just there? Where are all the cases? All the deaths?
I had 20 more slides prepared, in case Saar brought up any of these outside of Wuhan issues. Since he never contested them, I ended up turning them into a blog post:
https://medium.com/p/e54c358736bb
The specific issue with the Brazilian study is probably lab contamination -- they tried sequencing some of the covid in the wastewater and it shared mutations with B.1 strains that weren't otherwise seen until later in 2020. The authors theorized that meant that B.1 was circulating worldwide before Wuhan (one lab in Italy also found something similar, probably for the same reason). But the B.1 mutations made the virus better adapted to humans, so it would need to then travel to Wuhan and revert back to something less well adapted and then mutate into B.1 again.
A far more likely theory is that the 1 sample from Brazil and 1 sample from Italy were processed in labs where someone was a little bit sick.
I think one of the Italy studies tested samples from a range of dates, with positivity showing up maybe in mid-December? It's harder to explain that as contamination.
I think they are better explained by false positives, right? It doesn't really add up given the lack of an outbreak in italy
False positives after some date but not before? I agree that it is hard to reconcile with the lack of an outbreak.
Here's the paper, it's actually harder to believe than I thought: https://pubmed.ncbi.nlm.nih.gov/36029839/
I don't follow this story, but someone said this paper was the strongest of the ones suggesting an early international spread.
> False positives after some date but not before?
Not trying to Texas Sharpshoot, but I can think of some plausible explanations:
1. The stored samples are physically sorted chronologically, because this is a research lab and research labs like to be organized and alphabetical order wouldn't make sense in this case. Material from an "actual, for-real positive" sample contaminates its neighbors.
2. The work of studying the samples is passed to multiple people. The work is divided into chronological batches. By a stroke of luck, a positive-but-asymptomatic employee happens to handle and contaminate the batch of samples immediately preceding the "actual, for-real positive" batch.
3. The samples are stored "bottom to top", chronological order, and processed "top to bottom". Residue from an actual, for-real positive sample is left behind in lab equipment.
4. (comedy option) All the samples are processed by one person in chronological order. None of the samples originally contained covid, but the one person processing them caught covid at some point and began contaminating the batches as they worked on them. All the positives are false.
Basically, jokes aside, it becomes a lot more likely (or at least less-unlikely) when you consider the on-the-ground human aspects of lab work and sample storage.
Informative blog post, thank you.
I mentioned this elsewhere, but how is an early December market outbreak consistent with a 3.5 day doubling time? Fig 1 of this paper reports ~200 excess deaths per 100K per week from pneumonia in Wuhan on 1/23/20, the day of the lockdowns.
https://www.bmj.com/content/372/bmj.n415
A 3.5 day doubling time for an outbreak starting on 12/1 combined with a .05% CFR predicts (2 ^ 54 days / 3.5) * .05 CFR * 2 three and a half day periods in a week = ~4400 deaths the week of the lockdowns. So unless Wuhan has a population of 2M people, this is completely inconsistent with the excess death reports.
I agree the outbreak probably started late November, which I think does work with the numbers you give.
See my comment at https://www.astralcodexten.com/p/practically-a-book-review-rootclaim/comment/52714955 , which said November 27, although I think your numbers suggest something more like November 21. See also the graph Peter took from Pekar of when he thinks the epidemic began: https://miro.medium.com/v2/resize:fit:720/format:webp/1*duXQpxzXkveoFHWO4sF2pQ.png
I played with this compartment model, and could not get the excess deaths to match that report without going back to the first week of November.
https://gabgoh.github.io/COVID/index.html
which is directionally consistent with the simpler doubling time argument.
Even so, we're being asked to believe that under a late November market emergence date, the first confirmed cases were third or maybe even second(!) generation cases for a respiratory disease in a crowded city that no one was looking for and no one knew anything about? I am skeptical .
Maybe this has been answered elsewhere, but isn't it awkward to be relying on Chinese death reports? I had understood that was one of the areas that was most suspect about their records - they consistently downplayed the amount of covid in their population to give the impression that the governmental response was good and/or that there was no problem that even needed a response. Basically deny that there's a problem until it can't be covered up, then cover it up as much as possible after that. I thought that was common even outside this particular debate, that nobody really believes that the Chinese govt handled covid so well that their population death rates were a tiny fraction of everywhere else.
I think (and may be wrong here) that there would be some lag time before the Chinese authorities noticed there was a problem to cover up. The first week(s) of increased death counts would look like statistical noise, or just the usual annual fluctuations between the relatively-less-morbid summer season and relatively-more-morbid winter season. So while the "deaths from Covid" statistics can be (and probably are) fudged downwards, we would still expect to see a "hiccup" in the "overall deaths" reports released prior to January that, with hindsight, isn't simply seasonal fluctuations.
The idea that all these data involved in this debate are suspect is a fair point. My personal view is that excess deaths is possibly the most robust way of looking at the early pandemic in Wuhan, but I might be mistaken about that.
I'm surprised that you have that kind of trust in a continuous epidemiological model with only a couple parameters, surely things are noisy at small population sizes? As an example from the low end, Pekar et al 2022's modeling suggested 8 spillovers of which 6 died out. There is clearly some point at which COVID does not double every 3.5 days.
There is also documented bias from the case reporting criteria changing over time. Wouldn't cases of a previously unknown disease be undercounted at first?
The experts seem to be using Monte Carlo for this rather than fitting an exponential.
I agree at small population sizes things are more iffy, which is why I'm saying the data don't rule out a very small COVID epidemic, but they do rule out a sizeable one. I guessed "not more than 2 or 3 cases", but you could convince me the real number was 10 cases or something. I wouldn't believe 100.
I would bet those 6/8 failed spillovers didn't reach the 3 case threshold, certainly not the 10 or 100.
I suspect you are underestimating the threshold.
Pekar's cutoff for a successful outbreak was ~500?
EDIT: Pekar says "failed introductions produced a mean of 2.06 infections and 0.10 hospitalizations", congratulations :-)
Sure. Complex models get you published in a fancy journal. They don't necessarily give you correct answers. I mean, 8 spillovers? In Wuhan, but no where else? Does any one seriously believe that?
Yeah, Pekar et al 2022 is likely to be retracted based on the work of twitter anon @nizzaneela.
But I don't think the basic spread model is controversial and apparently they generated lots of runs with lineages that died out completely.
All of the many errors uncovered in Pekar et al. were in the direction favoring zoonosis, which is noteworthy.
If you don't trust a doubling time model, see if you can plug numbers into the above compartment model predicting a such an enormous excess death figure by 1/23. I couldn't do it in a back half of november time frame without unrealistic assumptions about the CFR or R0.
A priori there is no continuous model that I would trust enough to extrapolate backwards in time, through all the discrete behavior around the spillover, and come up with a credible start time down to a week.
I'm not any kind of expert in the field, but isn't that a very suspect use of modelling?
I don't understand how the doubling time of 3.5 days can be anything other than noise when the number of cases is in the low single digits
Answered below:
https://www.astralcodexten.com/p/practically-a-book-review-rootclaim/comment/52716610
Well, all of this just confirmed my priors: most debate is utterly worthless. They can talk about likelihoods and logical fallacies all they want, but the problem is that no one has any hard evidence. Peter's obviously the smartest one in the room here, but just being intelligent is of no use if you don't have anything to work with. Sure, you can take a bunch of circumstantial evidence and then claim that they make your theory more likely by pulling numbers out of your ass, but none of that gets you any closer to verifying the actual truth. And at this point, actually verifying the cause of COVID is likely impossible-- any hard evidence would have been burned by the CCP, since even they don't have anything to gain by knowing the truth.
...So what's even the point of all of this? As others have pointed out, the actual dangers of wet markets or GoF research is completely independent of whether or not COVID in particular was caused by either of them. One event doesn't change the reality of the systems at play. We already agree on the information that matters: that the situation was potentially preventable by the CCP, and that they failed to do so. But even that wasn't enough for countries to stop doing buisness with them, so how would figuring out the origin of COVID change anything? This entire discussion is utterly pointless.
The (applied) point of the discussion is the hope that the CCP will react to this (for them) embarassing and apparently never-ending discussion by cleaning up both the wet markets and the research laboratories studying stuff like this.
...and from that angle, the ideal outcome of the discussion is 50-50 for the wet market versus the laboratory leak - hypothesis.
More generally, there are two motivations for caring about the probabilities for a lab leak versus zoonosis:
(1) You portray yourself as an observer of the world, and simply are intellectually curious about what goes on in it. You care about the probabilities only because of the off-chance that during your lifetime CCP will fall from power or something equally dramatic will happen in China, which makes the authorities open the archives so it is possible to check if what you thought was probably the cause, was actually the cause.
(2) You portray yourself as a participant in the world, and want to influence policy-makers and your fellow man to limit threatening risks. The probabilities you and/or others assign to possible causes serve as calls to action to do something to change/reduce the probability of similar pandemics in the future.
These motivations are very different. They are likely to influence your own stated probabilities, as well as how you interpret why others offer the probabilities they do.
This is a highly interesting topic. Unfortunately I don't have the energy tonight to get deep into all the great points brought up in the post. But I want to share a few scattered thoughts in babble mode:
(1) Taking a 5% chance of dying on a climb is pure insanity. Maaaybe taking a 5% *lifetime* risk is worth it if you really love climbing. Hardly an endorsement for risking $100k at 4%.
(2) It's a terrible idea to offer a standing bet on your true beliefs. You're literally doing adverse selection and end up with a sneaky rules lawyer as your debate "partner".
(3) It seems that Peter is just a good debater, of the "debate club" variety. Laser focused on winning the debate, rather than finding truth. I was repulsed by the dirty arguments made, for example the quibbling about whether to call the first outbreak a "superspreader event" or not. Regardless of the completely irrelevant doubling times Peter brings up multiple times, a place like the Wuhan market clearly is a mechanism for turning a handful of cases into 80k cases. I've only skimmed the *summary* of the debate, so please point me to where Peter tackles the real underlying issue head on.
(4) Re: orders of magnitude. There is no limit to how badly you can fuck up a bayesian estimate. For example, Peter's 1 in 10'000 probability of the first infected person working at the wet market. That's just a uniform prior! It's so, so bad. The easiest way to debunk this is to take these odds, and start listing conclusions that would need to follow if they were actually correct: "viruses are equally likely to spread to everyone regardless of their profession and their position in the social graph"... um no, everyone with an inch of common sense (or kids) knows that viruses spread where people gather: schools, workplaces, and yes, markets. The vendor at that market didn't need to be the first, merely among the first *few*, and then keep coming to work while contagious.
(4a) It strikes me this might be a valuable method to flesh out. Let's call it the *reverse inference test". I don't know why, I just feel like it latches onto my intuitions more cleanly than trying to refute a claim in the same direction it's made.
(5) Probability cascades are bullshit. To me, this is the classic example of "science as attire"; you pretend to be scientific by multiplying numbers together, but what you're really doing is spinning a narrative. It's extremely hard to come up with an exhaustive enumeration of posibilities. MECE, as it's sometimes called (mutually exclusive, collectively encompassing). Or for the mathy people, an equivalence relation on the possible states of reality. If you doubt this, go take an advanced algorithms class and report back. I'm not joking... repeated encounters with impartial yet implacable entities (compilers, test suites, theorem provers, etc) will drill this into you like nothing else. Peter's argument style is that of a college freshman, and any competent math mentor would beat it out of him before it rots his brain. To the extent Peter is aware of this, it just reinforces (3).
(5a) to clarify the above point, the typical mistakes made is multiplying together probabilities that aren't independent, and to implicitly assume that the only way to reach state X is the specific path towards X you've laid out. These can make a 50% event seem like a 1% event, or the other way around.
(6) Explicit bayesian calculations in the absence of a *formal model* are a horrible idea. It just takes confusion and multiplies it by orders of magnitude. What I've learned in my time as a scientist is that the relationship between the real world and the model you're working on, and how to transport assumptions and conclusions back and forth between them, is one of the hardest things to get right. And we don't really have a systematic way of teaching this skill.
(7) If you wanted to blackpill me on the applicability of bayesian calculations to everyday life, you've succeeded. In particular, I see no reason to update my beliefs regarding the lab leak hypothesis in response to this debate. I'm still 50%+ in favor.
Wow, that was a lot more than I thought. Thanks for reading, and I hope it was helpful, even in this low-effort format.
We disagree in the large, but +1 to this post because every object level point is exactly correct, and that's way more important to me than agreement.
May you be blessed with good luck.
Thanks! we should really continue our discussion about AI at some point!
Actually, I have a question, what specifically in advanced algorithms are you thinking about when you say that it'll help with models? I'd like to implement a couple of those things as an exercise you see.
(Personal guess it's some kind of probabilistic algorithm, or just seeing a bunch of greedy algorithms that looks optimal... but actually doesn't even converge. I would guess that I'm not thinking expansive enough but seeing the gap in my prediction and your answer would be illustrative of what I'm missing)
For one, combinatorial algorithms. Stuff like sat solving, computational geometry and so on. The state space boggles the mind. Graph algorithms, including stuff like reductions of various problems to max flow.
One of the most interesting classes I took in my master's was models of computation. We had to program algorithms on Turing machines, cellular automata, tile completion systems, and so on. The interesting part about it is that we could have our solutions automatically checked. Not sure if there's equivalent stuff out there on the internet.
There's so much more. Stuff like computer graphics or physics simulations, where the operations are very elementary, but you can't just debug by looking at a single code path. You kind of have to think on a systematic basis and judge the output as a whole.
That's just off the top of my head.
Very good points. (4) and (5) are particularly relevant.
I know I am being ridiculous, but has anyone made a 2-3000 word summary of this post? Would be very useful for semi-literates like me.
Scott did, it's the last part of this post.
That is the conclusion, I was hoping for a summary of the arguments.
My brain skipped a beat when I read the summary that Peter dismissed one of the later market spreads as "that didn't come from a human source, there were COVID particles on the fish". That feels eerily lab leak mechanics adjacent and I found myself slightly confused why he thought that was likely but COVID from the lab coming to the wet market and then spreading from there unlikely. Probably if I watched the videos, I would be less confused about this, and I'm not trying to make an argument, just registering my confusion.
Generally, I have no strong position on this question at all and a low prior on that the things I think about it are true. I thought this article was a fantastic deep-dive! I realise I could get even more information if I watched the videos, but I'm not really a video person. Thank you so much for this. :)
Something about that whole line of reasoning felt a bit off to me. If the explanation is due to non-COVID-zero imported goods, shouldn't we expect to see a similar occurrence of outbreaks with other goods importers. Why did it specifically happen in a wet market?
Frozen fish is the leading "anywhere but here" theory, whose popularity is almost entirely restricted to the government of China. Peter doesn't actually believe in it.
Regarding the 1 in 10000 chance of being discovered in a wet market if lab leak:
That is flawed. You might as well assign 1 in 365 that it was discovered December 30th. You can tell it is flawed because if it had been discovered in a school, Saar would assign 1/10000. Or in a restaurant, 1/10000. Or a bakery, 1/10000. He gets a free 1/10000 wherever because any public place can be differentiated by at least 1/10000. “Only 1/10000 Wuhans work at Wuhan IHOPs!”
That was discussed in the 3rd debate. You can't assign high numbers unless there's some importance. If someone gets murdered in one building in Wuhan, and there are 10 million buildings, that's not a 1 in 10 million coincidence.
On the other hand, there are only 4 markets in Wuhan selling wild animals (out of a giant city of 10 million plus people). Huanan market is the one selling the most animals (about 40% of all the wild animals coming into Wuhan), but it's otherwise not the biggest market or anywhere near the most crowded place in Wuhan.
So if you're trying to weigh the odds between "wild animals caused it" and "the lab caused it", then it is of significance that it first showed up a the place where the wild animals were sold, as opposed to 10,000 other places it could have first showed up in Wuhan.
You’re right, I was misinterpreting how they were using bayes theorem here
That's not the complete reference class, however. Had the virus first showed up in a zoo which treated wild animals, or which took wild animals for its exhibits, and the chain of custody involved intersected with the region in question, that would also be treated as equally valid evidence for zoonosis. If it first showed up in a restaurant which used products from the wet market, or somebody who made purchases there, that might also count as part of the reference class. Anybody involved in the logistics of moving animals and anybody who visited a farm (there's no reason to limit ourselves to raccoon dogs). The workplace or school of anybody who visited a farm. The workplace or school of anybody who traveled to the province in question. The raccoon dog farm already has two degrees of separation, so any plausible path with two degrees of separation is part of the reference class - and that's an enormous reference class.
Note that the reference class for zoonosis isn't limited to Wuhan, and, indeed, we wouldn't expect it to first show up there given a zoonotic origin.
Once you start critically examining the actual reference class to which the wet market belongs, without arbitrarily limiting it to particular observed qualities, it is pretty clear it is actually enormous.
Whereas the laboratory is almost a reference class in and of itself; there are a few partial matches, so it's not entirely its own reference class, but it is something like unique in a way that the wet market absolutely is not.
You nod at this briefly in the conclusion, but what I find fascinating about this debate is that the what truly happened doesn't seem to matter at all! If an oracle arrived and told us the source, I don't think it would (or should) change our planning at all. The investigation of COVID has revealed that there are huge risks from *both* lab leaks and zoonosis that we need to invest significant effort in reducing, and which one actually happened in this case doesn't change that.
It seems to me that it matters. When you are estimating whether the next pandemic will come from nature or a lab, the result here should affect your prior.
What matters more immediately is whether to even continue investigating the origin of COVID. The most visible advocates for a zoonotic spillover have been hopefully declaring the case closed for some time now. But if COVID really was related to the DEFUSE research, there is a huge amount of evidence within the reach of US investigators and it would be a shame not to look at it.
Maybe I'm missing something, but it seems like Peter's argument hinges on two separate cases of zoonosis within days of each other in the exact same location and the same mutation? Or maybe I misunderstand that point.
Hopefully Peter will show up and explain, but my impression of his case is something like:
- Raccoon-dog on a farm catches COVID from bats
- Lots of raccoon-dogs on that farm get infected
- Several raccoon-dogs from that farm are brought to Wuhan for sale
- This version of COVID is already good at infecting humans
- The virus spills from two of the raccoon dogs to humans, for the same reason that if a bunch of infected Chinese visited America, it might spill from two of the Chinese over to Americans - because the virus spreads quickly in a naive population.
So the story is that both strains arose on a farm, and then it crossed to zero humans at the farm, but swiftly infected a whole bunch at the wet market? Seems pretty tenuous.
It's entirely possible there were spillovers elsewhere that died out. Before the pandemic the CCP had encouraged rural people to get involved in farming animals to make money. It was part of Xi Jinping's drive to "end absolute poverty" (in reality the threshold of what "absolute poverty" meant was a shifting target, but China did give many people a way to make a living this way). So you had people in very remote regions raising animals, in some cases near bat caves, then shipping them into big cities to be sold.
Pandemics generally don't start in rural areas, because the average person has many fewer daily contacts, lowering R to the point where they're unsustainable. I think there was a study showing that a bunch of animal handlers had SARS antibodies (suggesting they'd gotten the disease) when all known SARS outbreaks were in cities.
There is a lot wrong with this causal chain.
1. Everyone *on both sides* including the Worobey group and the WIV, agrees that the ancestral virus is in Yunnan or SE Asia. Miller keeps denying this but it is so totally at odds with all of the evidence, that you have to ask questions of him here.
2. The claim therefore has to be that the raccoon dog was infected in Yunnan and taken to a farm in Hubei where it acquired the furin cleavage site in Oct or Nov 2019. There is no direct evidence of this, and raccoon dogs roam wild in Hubei, so it seems weird for them to ship them from Yunnan. There is no evidence that raccoon dogs can sustain such intense transmission to cause acquisition of the furin site as happens with avian influenza, and lab experiments suggest that they cannot do this. Freuling et al found an R0 of 1 in their experimental infection of raccoon dogs.
3. Raccoon dogs have never been found infected outside of a lab, unlike eg mink.
4. If lots of raccoon dogs on a farm get infected, why were there no outbreaks anywhere except Wuhan city where the farm was supplying? If you are envisioning thousands of infected raccoon dogs on a farm, this is a big BF in favor of the lab. It only really works if you are envisioning about 100-200 raccoon dogs on a farm, but then you are conditioning on a very specific farm size, and we have evidence that the typical farm is an order of magnitude larger.
5. The raccoon dogs were wild caught in Hubei. Wang et al 2022 tested 15 wild caught raccoon dogs at the suppliers of the market in Jan 2020 and they were found negative. Around 38 raccoon dogs were sold at four markets per month. So, this is half of the inventory. The raccoon dogs photographed at the market were agreed by the Worobey group to be wild caught in Hubei. Maybe you think half of the raccoon dogs were topped up from farms in Hubei. There is no evidence for this. If they were topped up from fur farms, they would have been in eastern China, even further from the ancestral bat viruses.
6. So, the claim is now that these ~10 raccoon dogs are responsible for the outbreak. This is an extremely small fraction of the total wildlife trade in China, and conditions the zoonosis hypothesis on an extremely specific proposition. To justify this, one would need extremely strong countervailing evidence of raccoon dogs being the intermediate host. But what we do have is extremely small traces of covid in their cages and a negative correlation of raccoon dog genetic material with covid. Whereas for the viruses that the animals in the market were actually infected by, like bamboo rats covs or canine covs, there is a strong correlation. Miller's claim that 'we have down to the stall' is just wrong.
The collective BFs in favor of this causal chain are extremely weak, and in fact rule out raccoon dogs.
Just noting this comment implicitly references Jesse Bloom's work. (twitter, github).
Yeah, it would be:
some infected animals are shipped to the market.
some of those animals get people sick.
2 of those infections go on to infect significant numbers of other people.
The odds of there being 2 spillovers as opposed to 1 are about the same, once you have the right components in place. Eric asked a very similar question in the first week debate, if you want to take some time to watch.
Could you lay out from start to finish how the ancestral virus got into the human population. e.g. is it something like:
Ancestral bat virus in Yunnan infects raccoon dog
Some of these taken to farm in Hubei
Intense transmission leads to acquisition of FCS
Within a month of that, 20 raccoon dogs take to Wuhan in November
8 spillovers occur
2 catch on
Is that right?
Many Thanks!
>The six estimates span twenty-three orders of magnitude. Even if we remove Peter (who’s kind of trolling), the remaining estimates span a range of ~7 OOMs. And even if we remove Saar (limiting the analysis to neutral non-participants), we’re still left with a factor-of-50 difference.
ouch, ouch, _ouch_, _OUCH_.
And I thought that the single order of magnitude disagreement between the superforcasters and the domain experts on AI risk was bad... ( as I'd commented in https://www.astralcodexten.com/p/mantic-monday-31124/comment/51596120 )
In a way, the scatter in estimated here is even more disturbing, because AI risk unavoidably involves estimates of future events, many with only weak analogies to known past events. _This_ estimate is all about events which _do_ have strong analogs to known past events.
I don't know what to suggest. One large class of worries is indeed treating some probabilities as uncorrelated when one could indeed expect correlations. E.g. as you said, researchers modifying viruses in ways intended to be less researcher-like and more nature-like in order to anticipate natural changes.
One other thing that generally bothers me is the blanks in a lot of the probability calculations where one side estimated an effect and the other side just didn't discuss it. Maybe it would have been helpful to have a final pass where all sides gave estimates for _all_ of the factors?
Does peter have a blog for us to follow his future work?
https://medium.com/@tgof137
There's an option to subscribe via e-mail, if you don't like that site.
I could try making a mirror on substack, have just never gotten around to it.
So, I should go back to pitching my children's animated TV series "Wuhanimals" about a raccoon dog, a bat, and a pangolin who are friends?
South park had it right, all along:
https://www.youtube.com/watch?v=4-aeJEOLhCA
I truly hate to be a meta-commentator refusing to engage with the meat and bones of the debate, but I'm still puzzled by why this matters enough for the debate to be worth at least 100000 dollars and a few dozen hours of a few intelligent people's time (probably a lot more time and money, actually, given the extent to which all of this was discussed in other sources, of varying levels of sophistication, that I don't follow).
Would it help anyone in particular, even people who lost loved ones or whose businesses did not survive the economic consequences of the pandemic, to know whether the virus was zoonotical or originated from a WIV leak?
I doubt there would be meaningful political or economic consequences either way. I'm moderately sure virological research is still being funded about the same way it did before the pandemic, I'm a bit more sure wet markets are still as unsanitary as they were before the pandemic, I'm quite sure trade/political relations/economic dependence of the world on Chinese economy and industry is about the same as it was before the pandemic, and I'm very sure human beings still die the same way they did before (and during) the pandemic.
How much time time and energy has been spent debating whether the Holocaust happened as claimed, or what were the root causes of the American Civil War, or what caused the Great Depression, or how the pyramids at Giza were constructed? People like to discuss and understand important events in world history, even if there are no direct political or economic consequences hinging on the outcomes of these discussions.
Regardless of whether or not this pandemic originated in a lab, it's still the case that GoF research is potentially quite dangerous, its claimed benefits quite dubious, and we certainly should be having a serious debate about whether or not it should be more strictly regulated than it is. If COVID hadn't happened, we could make that a dispassionate theoretical debate. But here and now, we're stuck with a debate involving people who remember and care about COVID. And the way real people actually work, if one of the things they "remember" about COVID is that a bunch of conspiracy theorists claimed that it came from GoF research, but they were later shown to be lying liars who all voted for Donald Trump, then we are not going to get a fair discussion about how GoF research ought to be regulated going forward.
The same may be true of wet markets, but the case is I think weaker there because the animal reservoirs will still exist and will still interact with local human populations, and it's not clear how much benefit there would be to shutting down the wet markets if it just delays the remote village -> cosmopolitan city step a bit.
But so long as both hypotheses are credible, we need to keep making it clear that both hypotheses are credible and arguing against anyone who claims to have "proven" that it was 99.99...% certain to have been one or the other. And if both are credible but one is genuinely more probable than the other, then we want to know that so we better know how to allocate our political capital in a world where we may not be able to shut down both the wet markets and the GoF labs,
...since the main applied point of the debate is to shame CCP into making sure neither wet markets nor laboratory slip-ups cause the next Big Pandemic, a conclusion that the wet market and the lab leak hypothesis are equally likely explanations of the Covid pandemic would be the ideal (applied) outcome of the debate.
...that is, if your main motivation for caring about these probabilities in the first place, is as a call to action to limit major risks of similar pandemics in the future.
Great write up, and incredibly intellectually stimulating. One thing that struck me is how much harder it is to understand the past than to predict the future! If Rootclaim is so good, it should prove itself on the prediction markets, instead of relying on judges. This also solves the problem of financial imbalance, challenging people to high stakes debates is like bullying someone at the poker table when you have a pile of chips.
And then when those people win, you tell them it was rigged and they have to bet you again otherwise it doesn't count.
Super obnoxious.
"One thing that struck me is how much harder it is to understand the past than to predict the future"
Are you serious?
Importantly, Miller incorrectly claimed the N501Y mutation would result from passage in hACE2 mice (mixed them up with BALB/c mice). The major papers Miller relied on have been seriously challenged since the debate. See Stoyan and Chiu (2024), Weissman (2024), Bloom (2023) and Lv et al (2024). Overall the circumstantial evidence makes lab v plausible:
1. Chinese researchers Botao & Lei Xiao observed lab origin was likely given the nearest known relatives to SARS-CoV-2 were far from Wuhan. Wuhan Institute of Virology (WIV) sampled SARS-related bat coronaviruses where the nearest relatives are found in Yunnan, Laos and Vietnam ~1500km away. They refuse to share their records.
2. Patrick Berche, DG at Institut Pasteur in Lille 2014-18, notes you would expect secondary outbreaks if it arose via the live animal trade. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10234839/
3.Molecular data: Only sarbecovirus with a furin cleavage site. Well adapted to human ACE2 cells. Low genetic diversity indicating a lack of prior circulation (Berche 2023). The CGG-CGG arginine codon usage is particularly unusual but used in synthetic biology.
Restriction site SARS-CoV-2 BsaI/BsmBI restriction map falls neatly within the ideal range for a reverse genetics system and used previously at WIV and UNC. Ngram analysis of the codon usage per Professor Louis Nemzer https://twitter.com/BiophysicsFL/status/1667232580255490053?t=IJgitS5cw364ioclzVWxaA&s=19
The SARS2 backbone is very low in CG and CpG. While the 12-nt insert that gives it the FCS is extremely high in both. Almost as if it was some kind of chimera of a consensus sequence and a codon-optimized polybasic cleavage site? https://twitter.com/BiophysicsFL/status/1752800486837678377?t=EpIRgyybJVaPgeMP5xdstA&s=19
https://www.biorxiv.org/content/10.1101/2022.10.18.512756v1
https://link.springer.com/article/10.1007/s10311-021-01211-0?fbclid=IwAR1HMUMtLIAFOFppVasQDeoIAYrVhP8j4YoPO4wnaTOUiKLsllZl_oKryOw
4. DEFUSE full proposal: virus 20% different from SARS1, consensus seq assembled with 6 segments, without disrupting coding seq, BsmBI order, FCS. SARS2: 20% different than SARS1, 6 evenly spaced fragments w BsmBI and BsaI restriction sites, FCS.
Jesse Bloom, Jack Nunberg, Robert Townley, Alexandre Hassanin have observed this workflow could have lead to SARS-CoV-2. Work often begins before funding sought or goes ahead anyway.
5. Market cases were all lineage B. Lv et al (2024) indicates there was a single point of emergence and A came before B. So market cases not the primary cases. See also Bloom (2021), Kumar et al (2022). Peter Ben Embarek said there were likely already thousands of cases in Wuhan in December 2019.https://t.co/50kFV9zSb6
https://www.ncbi.nlm.nih.gov/pmc/articles/pmid/34398234/
https://academic.oup.com/bioinformatics/article/38/10/2719/6553661
6. Evidence for lineage A in the market is based on a low quality sample according to Liu et. al. (2023).
7. Bloom (2023) shows market samples do not support market origin. There is also no evidence of transmission in the claimed susceptible animals elsewhere. https://academic.oup.com/ve/advance-article/doi/10.1093/ve/vead089/7504441
8. Lineage A and B only two mutations apart. François Ballox, Bloom and Virginie Courtier-Orgogozo note this is unlikely to reflect two separate animal spillovers as opposed to incomplete case ascertainment of human to human transmission (Bloom 2021).
9. Sampling bias. George Gao, Chinese CDC head at the time, acknowledged to the BBC stating they may have focused too much on and around the market and missed cases on the other side of the city. David Bahry outlines the documented bias. Michael Weissman has shown this mathematically.
https://journals.asm.org/doi/10.1128/mbio.00313-23
https://academic.oup.com/jrsssa/advance-article-abstract/doi/10.1093/jrsssa/qnae021/7632556
10. Spatial statistics experts show the Worobey claim the market was the early epicentre was flawed.
https://academic.oup.com/jrsssa/advance-article-abstract/doi/10.1093/jrsssa/qnad139/7557954
11. Wuhan used as a control for a 2015 serological study on SARS-related bat coronaviruses due to its urban location.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6178078/
12. Superspreader events also seen at wet markets in Beijing and Singapore (Xinfadi and Jurong).
13. WIV refuse to share their records with NIH who terminated subaward in 2022. Wider suspension over biosafety concerns. https://www.bloomberg.com/news/articles/2023-07-18/us-suspends-wuhan-institute-funds-over-covid-stonewalling
14. PLA involvement at WIV and MERS research prior to SARS-COV-2. MERS features several similarities with SARS-CoV-2.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7022351/
15. SARS1 leaked several times and SARS-COV-2 has leaked from a BSL-3 lab in Taiwan.
16. Unpublished infectious clone identified from Wuhan contradicting arguments such reverse genetics systems would be published.
https://www.biorxiv.org/content/10.1101/2023.02.12.528210v1.full
Miller incorrectly claimed the N501Y mutation would result from passage in hACE2 mice (he mixed them up with BALB/c mice). Since the debate the key papers that Miller relied on have been badly undermined:
1. Worobey et al featured on Retraction Watch after spatial statistics experts eviscerated their claim the Huanan Seafood Market was early epicenter. https://academic.oup.com/jrsssa/advance-article-abstract/doi/10.1093/jrsssa/qnad139/7557954
2. Lv et al (2024) found new intermediate genomes so the multiple spillover theory is unlikely (it was anyway given lineage A and B are only two mutations apart). Single point of emergence is more likely with lineage A coming first. So market cases are not the primary cases (all lineage B). Their findings are consistent with Caraballo-Ortiz (2022), Bloom (2021).
t.co/50kFV9zSb6
3. Jesse Bloom showed again the market samples don't support market origin. t.co/rorquFs1wm
4. Michael Weissman shows mathematically ascertainment bias in the early case data the judges relied on (George Gao, Chinese CDC head at the time, acknowledged this to the BBC last year - they focused too much on and around the market and may have missed cases on the other side of the city).
https://academic.oup.com/jrsssa/advance-article-abstract/doi/10.1093/jrsssa/qnae021/7632556
5. Account that identified errors in Pekar et. al. leading to an erratum last year has found another significant error. Single spillover looks more likely. t.co/GAPihZu51P
> I have a weird urge to visit Wuhan as a tourist, see the Wuhan Institute of Virology, stroll through the Huanan Central Seafood Market (unfortunately closed), maybe eat a raccoon-dog.
Scott N Alexander N plans N to N eat N Raphtalia N CONFIRMED
Thank you for an incredible write-up and analysis.
I am actually shocked at how poorly Saar fares here. The vast majority of his evidence is basically "if things were different in a way we can't know, then our conclusions would also be different, therefore lab leak." E.g. "if COVID spread rate was lower in the beginning", "if an infected WIV worker spread to the wet market then isolated", and so on. That's it, really? I was expecting so much more actual evidence. And then what data points they do provide (e.g. "Mr Chen", 90 cases, intermediates, etc) have apparently valid reasons why they're not evidence. I started out 60/40 zoonotic, but based on this I feel I have no choice but to update to 99/1.
I feel the real problem with this setup/framework is that a specific position is taken, with overly-precise numbers, and the goal of the debate is to simply justify that final percentage. You see this with Saar's "but the judges just did the math wrong" sour grapes complaint.
In the IT and infosec world, we have the FAIR method--a similar method for building quantified models of risks. The real value of these isn't that you convinced yourself you have only a 1-in-10000 chance of being compromised by a state actor in the next 10 years... it's the shared understanding of the risks your business thinks are important, how you think those risks are correlated, how you think they're evolving over time, etc. That enables your organization to act from shared goals, and not get bogged down in interminable debates day-to-day. The final numbers are fine, especially if you use confidence intervals, but you know it's only a snapshot in time using your best data and estimates as of the last time you underwent the exercise; by the time you update in a year/quarter, you will have more data and different estimates. Sometimes, risks that weren't even on your model last cycle, are now the top priority because of what's happening in the world (or because you gained new information that you missed previously). It's a fact of life.
With that in mind, I think Saar fares better if his arguments are framed as reducing absolute confidence in zootonic origin, and not the reason why LL is the only possible explanation (which it obviously isn't, even from Saar's own arguments). So yeah, I was only joking about updating to 99% in favor of Z--I think there's enough doubt thrown in from Saar's evidence alone that it could not possibly be more than 80%. But Peter/Z still wins this version of the debate, and Saar's over-confidence in "no way it's anything other than LL, here's the math to prove it, and if you disagree then you're doing the math wrong" does border on pseudoscience here.
To that end, I think Scott's final takeaways are the most important. If we have any shared goals anymore as a society, then we should devote resources to better preparing for pandemics from both origins, and in also improving government/organizational transparency so it's easier next time to understand what went wrong. As it is, either a natural disaster can be ignored because "teh government conspiracy," or else somebody got away with murder; either way, the debates that focus solely on convincing everybody that one's favored position is the only explanation means we're not more prepared for future disasters than we were with this one, we're just more divided and easier to manipulate.
"The final numbers are fine, especially if you use confidence intervals, but you know it's only a snapshot in time using your best data and estimates as of the last time you underwent the exercise; by the time you update in a year/quarter, you will have more data and different estimates. Sometimes, risks that weren't even on your model last cycle, are now the top priority because of what's happening in the world (or because you gained new information that you missed previously). It's a fact of life."
Thanks for this.
I found this debate to be very valuable. I started with 70:30 odds on lab leak's side (I didn't do any extensive research on COVID origins) but after this debate that shifted to 5:95 (i.e., favoring zoonosis). I wish there are more such debates on contentious topics.
The Rootclaim model seems pretty sensitive to modest differences in probabilities; I'm guessing a factor of having too many variables so small differences across all of them add up to a big effect? I tried to average the six guesses giving Saar the least weight and add some of my own intuition and got (I think) a modest pro-lab leak outcome. https://docs.google.com/spreadsheets/d/14OkrvJ21qCqQcXINQnMO4-a1Xi-npidF/edit#gid=2057332411
> About half of cancer patients lose their hair, and Putin hasn’t, so we’ll divide by two.
This leapt out as methodologically unsound. How do you know whether Putin's lost his hair? Here are some possible scenarios:
1. Putin has never had cancer or anything cancer-like. His hair is fine.
2. Putin has had chemotherapy. He wears, or has worn, prosthetic hair to hide this fact.
Note that these two scenarios are perceptually identical. You have no way to tell one from the other.
If the question you're evaluating is "Is Putin covering up a case of cancer?", his visible hair provides no information at all, because all scenarios predict visible hair. In the "no" case, the hair is irrelevant, and in the "yes" case, the hair is deceptive.
It's kind of like making an adjustment to your evaluation of "Is this a coral snake, or is it a king snake?" for the observation that the snake in question has red and yellow stripes. That observation was already part of the premise of the question; you're double counting (among other weird mistakes) if you adjust for it again.
> Note that these two scenarios are perceptually identical. You have no way to tell one from the other.
🤔
?
Of course we can tell a wig from real hair.
Not your finest argument.
Haven't you ever watched a movie? You know how the same actor will show up in different movies with different hairstyles? (Or... the same movie, with different hairstyles?)
Generally those are wigs. It isn't possible to tell the difference between a wig and real hair; often they are real hair, just not real hair that the person to whom they are attached grew.
If you had the opportunity to physically manipulate Putin's hair, you might be able to demonstrate that something was off. But you don't.
Good illustration.
It reminds me of a Taleb quip directed against only-thinks-within-the-box-rationalists, in one of his books. Someone rolls a dice nine times. It comes up with sice (six) each time. He asks a statistics professor how likely it will come up with six also in the tenth roll. The professor answers "1/6 chance". He then asks Fat Tony, a street-smart mafioso. Fat Tony answers: 100 percent chance. Why? "Obviously, the dice is loaded."
> Another of Saar’s concerns with the verdict was that Peter was an extraordinary debater, to the point where it could have overwhelmed the signal from the evidence.
I've mentioned this before in the ACX comments, but this is such a great lead-in that I'll say again that this is why I didn't make a habit of reading Overcoming Bias. Eliezer Yudkowsky used that blog to make a lot of excellent points in an entertaining and well-written fashion.
And I got the sense, reading his work, that sounding persuasive was a bigger part of his strategy than being right, meaning that reading him was dangerous. So I enjoyed a number of pieces, but I didn't follow him or seek out his work.
My suspicion here is that using estimated probabilities instead of probability distributions leads to the same issue as for the Fermi paradox, in that your result is extremely sensitive to your assumptions and you end up with extreme outcomes on either side like we have here: 10^-25 and 534 are both clearly wrong.
This is a very cogent observation. Like Eric Drexler et al. in 'Dissolving the Fermi Paradox' they realized using point estimates instead of entire distributions lead to weird & paradoxical results. Good job!
If you have expanded on this somewhere, I would be interested in the reference.
My overwhelming feeling while reading this was a sense that a lot of the important evidence is itself subject to long debates and explanations that are being glossed over. Is a certain length of protein sequence likely or unlikely to be done by a mutation in the wild? What is the right way to construct priors about wet market and virus lab locations? Everyone was just stating certain things confidently, while providing detailed documentation for other things.
In the end, “evidence for my claims” seems to have a fractal quality sometimes. The closer you look, the more evidence you need to back up your use of evidence. It’s turtles all the way down.
While I'm terribly sympathetic to most of Wilf's points, one thing I absolutely agree with him on is the superiority of written to real-time/spoken debate, because the things it rewards correlate much more strongly with being right.
It seems to me that Saar's evidence-weighing analysis is basically "outside view" with extra steps.
As I see it, Saar's analysis hinges on the point that the zoonosis hypothesis contains a coincidence that is very unlikely to be explained by some out-of-model error (which I agree with), whereas the lab leak hypothesis only contains a coincidence that could very plausibly be an out-of-model error (I agree with this as well). This sets up a barrier that is almost impossible for any amount of mere digging into the details to overcome, as that details-digging is all itself quite vulnerable to out-of-model errors. And so the end result is an outside-view judgement that no amount of inside-view analysis can outweigh.
This surely seems like an error to me, though those with different positions on the outside-versus-inside-view debate might disagree.
Could you specify/elaborate on what you mean by "inside view" versus "outside view"? (or give a reference?)
I genuinely do not understand why there is not more talk about the third option: that one of the many bat cave spelunkers from Wuhan caught COVID-19 and brought it back to Wuhan.
The WIV and the market were fairly far apart from each other, as noted in the debate. The Wuhan CDC is *right next to the market*. And while it doesn't seem that the Wuhan CDC was conducting gain of function research, their employees were fanning out all over China going into bat caves.
From this Washington Post article - https://www.washingtonpost.com/world/asia_pacific/coronavirus-bats-china-wuhan/2021/06/02/772ef984-beb2-11eb-922a-c40c9774bc48_story.html
"In the video, the researchers scale the cavern wall, their headlamps ghostly blue.
“If our skin is exposed, it can easily come in contact with bat excrement and contaminated matter, which means this is quite risky,” says Tian Junhua ["associate chief technician in the Wuhan CDC’s pest-control department, but he has a reputation as a swaggering adventurer in his work with bats and insects"], one of the bat hunters.
...
The video was released by national science authorities and Chinese state broadcaster CCTV on Dec. 10, 2019, and circulated on social media. It’s a high-quality production, designed to promote China’s world-leading viral research.
...
Tian and his team from the Wuhan Center for Disease Control and Prevention are filmed catching horseshoe and pipistrelle bats and collecting samples of guano, in search of new bat-borne diseases and the basis of new vaccines. Tian talks about the need for caution. “It is while discovering new viruses that we are most at risk of infection,” he says, though he is shown handling sample vials without wearing full protective gear.
...
In 2017, Tian told the state-run Wuhan Evening News he once forgot personal protective equipment and was splattered with bat urine, leading him to quarantine at home for two weeks. On multiple occasions, bat blood squirted onto his skin while he was trying to grasp the animals with a clamp, he told the paper."
So the questions would be - exactly when did the Wuhan CDC move to be close by the market? The move finished on December 2, 2019 - how long was the move and how was it conducted? Does the science require an intermediate animal or could it have jumped directly from bats to humans with the cave exploration?
USAID seems to have been sufficiently aligned with your concerns by late 2023 to end the DEEP VZN program:
https://www.science.org/content/article/u-s-cancels-program-aimed-identifying-potential-pandemic-viruses
Thanks, very interesting! I'd like your comment if I could.
I am now scared that the next pandemic will be started by debaters on either side doing gain of function research to shore up the evidence for auxiliary points in their debate.
There's so much stuff on this page that Chrome is having difficulty rendering it even on a high-end desktop. Consider closing comments and creating a separate page for that.
This is an issue on most of Scott's posts; Substack just isn't designed to handle lots of comments. Given that this has been an issue for years, it seems unlikely they care to fix it. SSC was much better in this respect.
Substack just isn't designed to handle, period. It's unbelievable that a text-based site could possibly run so poorly, and yet here we are, on computers so powerful they can emulate computers *emulating* computers from twenty five years ago, chugging along worse than Slashdot twenty five years ago, Livejournal twenty years ago, Blogspot fifteen years ago, or Reddit ten years ago.
Sites with lots of text and comments are a *solved issue*. I shouldn't have to literally download Firefox just to browse the comments for more than ten minutes without crashing (on Firefox it "only" crashes every twenty minutes or so).
(I don't remember slashdot ever struggling as much as substack is here...)
I remember when "nested" comments first came out (different from "threaded" view, which was more like old-school UBB) it could get a little bit laggy... on a Pentium II. A computer far less powerful than the cheapest burner phone out there today. And it still wouldn't crash, per se.
Substack is just embarrassing.
At the outset of the pandemic, I was at about 15% lab leak, 85% natural origin, and thought it unimportant because I didn't think we'd ever have enough evidence to do better than guess. By the start of this year, after way more than 15 hours of studying way more evidence than I imagined we would ever have, I was at 80% lab leak, 20% natural origin, After this, maybe I should update to 70% lab leak, 30% natural origin. But I'm going to indulge in a bit of emotion and make that 25% lab leak, 100% annoyed.
Annoyed in that it seems like the rationalist and rationalist-adjacent community, parts of which at least I value and respect, has decided that on this one issue we're going to flip a $100K coin to decide what to believe, and I suspect anyone who dissents will at best be quietly dismissed as an ignoramus by much of the community. I will grant that by paying $100K, we got a coin that is *slightly* biased towards the truth,
Seriously, one-on-one personal realtime debate is a *lousy* way to find the truth. There's a reason we don't have debates as a central feature of e.g. scientific conferences. Throwing in high-value bets doesn't improve things. The process selects for the most skilled and charismatic debater, and in some cases for the favor of biased judges, with the truth mostly just serving as a tiebreaker.
In this case, there was no tie to break. Miller, by all accounts including this one, is an exceptionally capable debater. Wilf, while clearly skilled as an entrepreneur and possibly as an analyst, does not seem to be very good at in-person debate. From this very thorough account, Wilf completely failed to raise some of the very strong points in favor of the lab leak hypothesis, failed to call out Miller on some of his indefensibly weak points, and agreed to a structure where he'd open by spending the first third of the debate on the very weak assertion, "maybe it didn't come from the Wet Market" even though that is in no way necessary to the lab leak case. Or maybe he did raise the strong points and call out Miller's weak points, but in a way that neither Scott nor the judges noticed.
As for Scott shifting is odds from 50-50 to 90-10 on the basis of this debate, I think that requires p<0.20 for "Wilf isn't very good at this and left the winning arguments at home", which I don't think is defensible. You'd need a statistically significant sample of debates to be confident you have seen both sides well-argued.
I will continue to value and respect the rationalist and rationalist-adjacent community, particularly Scott's corner. But it has a clear blind spot in its enthusiasm for clever and novel truth-seeking mechanisms (bets, prediction markets, debates, etc) on the basis of their cleverness and novelty without adequate regard for their in many cases obvious flaws.
I'm not taking the results of this debate too seriously. I'm taking the fact that the debate exposed me to the arguments on both sides, that I spent a long time grilling the debaters on their arguments, that I read their followups and blog posts, and then that I read many other blog posts and a couple of papers on this topic seriously. Plus the surveys of experts and superforecasters.
The post has long cut and pasted sections of unchallenged borderline dishonest arguments from Peter. The quotes are so long that sometimes I had to double check whether it was you or Peter talking. Many of the arguments come from a paper which is flunking a careful replication attempt, not mentioned. Weissman's probabilities are not included because the items didn't match up with the rows of a spreadsheet built from Peter's analysis or something. There is not much evidence of research beyond the debate here.
I'm glad to hear that and, knowing you, not terribly surprised. But this,
"For what it’s worth, I was close to 50-50 before the debate, and now I’m 90-10 in favor of zoonosis"
does kind of read like it was just the debate. And, as I said, I don't think it would be reasonable to make that kind of shift on the basis of just a debate, because it requires unwarranted confidence in the process and participants. You say that the debate exposed you to the arguments on both sides, but in reality it only exposed you to a subset of those arguments. Ideally, that would be a comprehensive subset including all of the strong and relevant arguments on both sides, but I am skeptical that is the case here. And I'm certain that you have at least some knowledge of missing arguments from your other readings, but most people (including most rationalists) don't.
Unfortunately, as I see this debate spread out into the broader internet, it is largely being stripped of nuance and turned into "Two really smart people held a very thorough debate on COVID origins, and now all the smart people agree that it's almost certainly natural zoonotic origin, nothing more to see here, move along".
> The paper that claimed that defined how well COVID was adapted
Typo
> Something like this is true for the other Chinese wet market based outbreaks we know about it.
Extraneous "it" at the end.
> the furin cleavage site really does stand on a genetic map
I think this is supposed to be "stand out"?
> and which which Viewers Like You did not appreciate
Duplicate "which".
> something that everyone already everyone else had already considered
Repeated phrase
If it turns out it was natural origin — as I am more likely to now believe — it’s sad it will be that much harder to end gain of function research. Even if the lab did not leak the pandemic, the risk that the lab would leak a pandemic eventually was still too damn high. Hopefully this event will nonetheless lead people to take the risks more seriously.
I’ve spent some time in virologist subreddits since covid, and I find their attitude towards these risks to be terrifying. They remain very blasé about it. It appears to be deep in their culture, similar to how civil engineers are culturally committed to designing for cars, and architects are culturally against ornamentation.
The science has to decide whether it’s a lab created virus or not. However created in a lab is not the same as lab leak. In fact your prior for this being a leak should be close to, or at, 0% since it’s the only time a major pandemic has happened in that fashion.
When you look at the distance between the lab and the market and the fact that the origin is clearly the market if lab created pathogen is proven true then the most likely explanation is an external attack.
>In fact your prior...
The 1977 fku pandemic is feeling very neglected here.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4542197/
> So for example, suppose there are fifty things about a virus. You should expect at least one of those to have a one-in-fifty coincidence by pure chance.
Minor math error here I think: You should expect *exactly* one, not at least one. It's equally likely to be more than one or less than one.
Thank you to Saar for running this in good faith, and thank you to Scott for a great write up. One of the key things that stood out to me while reading the analysis was that there are a few relatively simple factual items that were disagreed on that might be relatively easy to chase down if we weren't worried about governments trying to obfuscate things. There's also a few simple pieces of information that could shift my thinking a bit too, like where do the lab workers live? I wouldn't expect a lab leak outbreak to be centered around the lab, but around one of its employees homes.
I noticed in myself a high risk of falling victim to a classic 'david and goliath' story too, which is potentially a very bad bias. Almost everyone reading that first section is going to be rooting for Peter, and enjoy it everytime he seems to win a point.
I also found myself with the David vs Goliath bias, despite having no opinions on COVID origins and not particularly caring about it.
Despite this, Saar’s arguments (as described here) often ended up fairly weak due to him leaning on what seems like a lot of “bad claims” about things like the pre market cases that Peter easily dismissed just by following the logical consequences that should have occured if they were true. I think this ends up with him showing weaker than his argument really was.
Also agree, huge props to Saar for doing the thing in good faith, and hopefully this taught him a lesson that when someone bets a large % of their NW versus your proportionally small bet, they will prepare 10-100x as hard as you will (because they have so much more to lose), and the fact that they took the bet in the first place implies they are likely already exceptionally good at whatever you are betting on.
I must admit that I don't care nearly as much about whether COVID was zoonotic or not, as I do about the government paying social media sites to enforce a particular orthodoxy about the question.
Some of this sounds like an “Encyclopedia Brown” level of argument. E.g. the idea that because the University of North Carolina said they’d do furin site insertion in a grant application from years before the pandemic, therefore the WIV cannot have done any insertions of their own.
Research institution capabilities, and researcher ambitions, and the real world generally, are a lot more complicated than that reasoning implies (I say this despite putting little weight personally on COVID being engineered).
Likewise with the idea that the WIV would have fully released the details of all viruses they have sequenced.
I haven’t watched the debate though, so maybe it was better than Scott makes it sound.
It strikes me that whether you should debate theories you think are crazy depends on how popular they are. Consider:
1/10000 believe in theory A. You debate someone arguing A, and you win 99% of the audience (considering what humans are like, this is super impressive). But now, 1/100 of the audience believes in the crazy. And *no-one* can convince 99.99%.
66% believe in theory B. Now when you debate someone arguing theory B, even if you only convince *half* the audience, things have improved!
This, of course, assumes that the people believing what you think is crazy can be convinced in the first place. This might well be true about lab leaks or *possibly* even UFOs; it will *not* be true about creationism (evidence isn't even a factor in why someone is a creationist).
The JFK assassination is another of those cosmic jokes. Sensible people should not believe in conspiracy theories about it. But dear Lord, was Lee Harvey Oswald a weirdo, the rare person who had been involved with *all* the kinds of people who _might_ be suspected of conspiracies; and murdered before he had a chance to testify; and with an investigation that couldn't have seemed more suspicious and cover-up:y if it tried.
If a lone gunner killed a reformist Russian president, if the gunner was himself killed by a man with mafia links and if that man died of cancer a few months later we would all be fairly conspiratorial.
Really fantastic read. I already leaned towards zoonosis but this removed most of the remaining doubt. It so valuable to see one side put up their strongest arguments and then see if the other side can knock them down, which in this case they did pretty dramatically.
Absolutely fantastic.
“… And so on and so forth, until we end up with the final calculation: 86% chance Putin doesn’t have cancer, too bad.”
What Rootclaim is doing is interesting. But should not such hard-core Bayesianism also include sensitivity analyses? That is, investigate how robust the conclusion is (here: “86 percent chance”), at least for minor-to-moderate changes in priors, age-related cancer statistics, and the like?
Using sensitivity tests in cost-benefit analyses is usually recommended. It appears that Bayesianism could benefit from such tests, for the same reason.
It would also quell the skepticism people like me automatically feel when encountering predictions spelled out in fine-grained percentages. (Plus, why stop at percentages? Why not add a few decimal places as well: 86,xy percent chance Putin doesn’t have cancer– that would make the prediction even more “accurate”….)
Do not get me wrong – I respect what Rootclaim is doing. But I would be (genuinely) interested in why Rootclaim does not add/construct some kind of confidence interval around their percentage-estimates. (And why Bayesians in general do not do this.)
SARS2 leaked from a Wuhan lab but it's not Chinese junk
https://jimhaslam.substack.com/p/lets-be-blunt-sars2-leaked-from-a
This post caused me to strongly update against Bayesian methods. Maybe it's the impossibility of seeing accurate priors, or the tendency to engage in cherry-picked absurd probability estimates. Or maybe it's the qualitative difference between statistical probabilities prospectively estimated vs retrospectively measured. I no longer have confidence this method produces more signal than noise.
For example, what prior do you set for an outbreak in WSM? You might think that's low, but given multiplicity you're not really looking at a prior probability anymore, are you? Both a spillover event and a lab leak might have come from any number of places. We'd tell some plausible-sounding story about why that place was unique, then set absurd priors as a fence around that story. But at heart it's still a post-hoc story, and all the prior-setting in the world will only serve to reinforce your story with more plausible-sounding stories. This seems to be what Saar is fighting against in making adjustments for out of model variance, but that seems like the wrong approach.
The better approach is probably abandoning Bayes - at least explicitly/strictly. I'm not yet convinced about the whole Bayesian enterprise, but it feels like it's on much more shaky ground for me today than it ever has since I first learned about it.
"This post caused me to strongly update against Bayesian methods."
This is also my major takeaway from the post & from the comment section.
I sometimes teach students who enter occupations where they have real power over other humans. Such as child protection officers, who in my country have a lot of influence when deciding in a child should be taken from its parents and moved to foster care.
I have toyed with the idea to include some "real" Bayesianism in their curriculum, with regard to calculate the risk of various forms of child abuse in various types of families. (As of now, I only make references to what I half-jokingly label "everyday bayesianism", which is basically what Scott talks about early in the post - also captured by the expression "inference to the best explanation".)
This post, and this extensive comments section, has made me update (as they say here) strongly against this idea. Few things are more dangerous than professionals with real power believing that "science" can give them analytical tools that makes if quite certain when sufficiently serious child abuse goes on, that drastic measures are called for (and can be defended in the courts).
Putting numbers on inherently uncertain things creates professional peace of mind. It is a temptation in all professions given power from the state to dramatically interfere in the lives and health of other human beings. So no wonder that "Bayesianism" is an increasingly popular catch-word in the helping & caring professions, as elsewhere. But judging from this blog post & comments, "real" bayesianism is a danger, not a help, in reducing type I errors (not interfering when one should) and type II errors (interfering when one should not).
Well said.
After thinking it over more thoroughly, I'm not sure the problem is that Bayesian thinking is wrong, per se. The problem is that some of us have been convinced to apply it more broadly than it should be. (Here I'm thinking of Scott's "P(A|B) = [P(A)*P(B|A)]/P(B), all the rest is commentary.")
Simple heuristic: before you try to apply Bayesian reasoning, ask yourself, can you accurately fill out the equation? "Not quite, but I think I can estimate-" No. The answer is no. Move along, or you'll be getting yourself into trouble, trying to apply (even loosely) principles that will only lead you to error.
That's a useful rule of thumb, I'll remember it.
I personally agree it was not a lab leak but a pretty important was lost in the debate (or at least poorly factored in). Namely China was making hiding evidence. While this may impact priors…the bigger impact is that, if it was a lab leak we only know what information was released (which almost certainly would be anything that boosted their preferred narrative) and do nit have all the evidence that was presumably withheld (which would be all the evidence they could suppress that went against the preferred narrative).
It’s kind of like sample bias in that the evidence you use to make the assessment is systematically biased.
I unsure how you assess these probabilities because they are fundamentally unknowable.
I did a poor job articulating this and want to highlight that I do not presume to know what the CCPs preferred narrative even is and it could have changed. They could and have started by withholding all evidence at the local level (this does appear to have happened) but then at higher levels decided on a new narrative. We just don’t know and cannot reliably intuit this with any degree of precision
The point is that we can assume the evidence we are using is biased in a systematic way but have no reasonable way to assess this.
"maybe you should lower your disbelief in each hypothesis to something more reasonable, like 99%. But now the chance that the sun rises tomorrow is 0.99^100, aka 36%. Seems bad."
All these hypotheses are very correlated; extreme probabilities of base events are not required, just extreme probabilities of conditionals.
'If all 99 other hypotheses are false, and I'm not making some silly maths mistake or whatever, then hypothesis 100 is 99.999999% false' is not necessarily overconfident even for a human.
Right?
I have done my own extensive research and believe the evidence is compelling that there was a lab leak. IMO, Peter has a real axe to grind and his strong personal conviction against the lab leak hypothesis has made him blind to common sense.
The odds that the virus evolved free of human-engineering are probably much less than in 1/10,000. Peter has no idea how to figure those odds and frankly no one else does either. The actions of the Chinese government in stonewalling are far more convincing to me that there was a lab leak. If the Chinese government didn't have something to hide why be so opaque???
They stone walled at the start but that was hiding the emergence of the virus in the wet market.
A few thoughts:
For an adversarial setup (such as debates or US court cases) to converge on something close to the truth, the case for each side needs to be represented at a much more similar level than was done here. Even if an innocent man is accused of a crime (so he *knows* that the truth will set him free), it is still generally recommended that they hire a competent lawyer to represent their side, as opposed to defending themselves. This is probably not as important in countries with a less adversarial court system.
Calling Peter a better debater doesn't quite capture what happened here. I don't think there was a great difference in rhetorical skill, but there was an extreme difference in both the level of preparedness and the commitment to the cause. This case involves 50+ pieces of circumstantial evidence of various degrees of strength, some favoring lab leak and some favoring zoonosis. All the ones favoring lab leak were heavily contested, but at least to me it did not seem like the same was true for many of the ones favoring zoonosis. This was probably a mixture of a few things. Saar didn't seem to not know enough of the precise details to effectively push back against some of the bolder zoonosis arguments, such as how confident we should be that we actually have a good picture of the first cluster. In other words, he conceded far too much, allowing the zoonosis side to look much stronger than I suspect many people still believe it is. Secondly, I don't think Saar fully understood what his "role" in this debate was. Instead of being a soldier for his side, he mostly stuck with his 2-3 year old argument for what he thought the "truth" was. Peter, on the other hand, seemed eager to actually win the debate, so he made the maximally strong argument for his cause, not granting anything, as opposed to trying to converge on the "truth". So for every piece of evidence, he was able to put on the "maximally pro zoonosis" spin, perhaps resulting in his absurdly low final lab leak probability.
It is perhaps interesting to think about how this compares to adversarial collaborations, which are popular in our community. Using the language of Julia Galef, one can perhaps think of an adversarial collaboration as a "dance" between two scouts. Similarly, a debate can perhaps be thought of as a battle between two soldiers. The incentive is to win, not to find the truth. The rootclaim debate looked like a battle between a soldier and a scout, with the predictable result.
(Disclosure that I support in BiosafetyNow, but these are my views and don't necesarily reflect those of BN)
The Twitter origins debate, which Scott understandably shies away from (ref. his most recent tweets, and ofc I certainly don't blame him) represented the richest, most boisterous most tumultuous, most rancorous, most acrimonious form of the origins debate I've ever encountered online. The current ACX substack post & its comments don't reflect the Twitter debate's diversity of opinions, evidence, and individual credibilities IMO.
The key items I feel are insufficiently highlighted (even if they are mentoned at least once) from my scan of the comments section:
• the value of 'adverse inference' facing:
° PRC's gov
° the most outspoken virologists (+ 1 evolutionary biologist, + 1 disease ecologist, and yes not *all* virologists) especially when relevant materials turned up from their side only following subpoena
• No progenitor or 'sister' virus with an MRCA dated to shortly before outbreak has been found in animal populations despite extensive sampling.
• I don't feel any post-outbreak (and some near-pre-outbreak) data out of the PRC should be accepted on faith; rather it & all conclusions drawn therefrom should be placed in its own little suspect epistemic box. By contrast, any data in non-PRC hands pre-outbreak is 'gold'. The remaining quadrant, data & analysis produced in the ordinary course ex-PRC but post-outbreak, is tainted by the need to avoid 'rocking the geopolitical boat'. Often we can't reference such data that for example underlies post-outbreak intel findings, like that of the ODNI (atop DoE, FBI, CIA DIA...) so I can only give it the 'silver' label treatment.
And transcending all these quadrants, verified data that turned up ex-PRC from leaks, FOIAs, sequences reconstructed from SDAs, etc. is 'platinum'.
* S̶D̶A̶s SRA's - Sequence Read Archive
1. Peer reviewed papers since the debate that show ascertainment bias in early case data and intermediate genomes that undermine the multiple spillover theory indicating lineage A arose first so the market cases are not the primary cases (Weissman 2024, Lv et al (2024), Stoyan and Chiu (2024).
2. Miller's claims about the furin cleavage site are dubious:
Miller incorrectly claimed the N501Y mutation would result from passage in hACE2 mice (he mixed them up with BALB/c mice). WIV was performing in vivo experiments in transgenic (human ACE2 expressing) mice and civets in 2018 and 2019 in SARS-like CoVs.
I think both sides may have also overlooked the MERS furin cleavage site shares several structural and functional similarities. The sequence looks quite similar. In 2019 WIV researchers were involved in MERS research. Yusen Zhou who died in mysterious circumstances in May 2020 and was involved in an early Covid-19 vaccine was also involved in this research that involved manipulating the FCS. t.co/7zcSUPR60T
Broad Institute biologist Alina Chan also observes the S1/S2 FCS PRRA insertion in SARS-CoV-2 generates a Class IIS restriction enzyme site (BsaXI). This was used by WIV and Ralph Baric at UNC previously.
Dr Andreas Martin Lisewski discusses similarities with a MERS infectious clone described in 2017 here. The argument no one would engineer an FCS like this overlooks the MERS example.
t.co/fAVUlJu0TK
You've posted approximately this same comment three times on here now, plus five or six times on Twitter. If you post it further, I will ban you.
I think it's wrong to characterize this as a tragedy for Saar and Rootclaim. Sure the contest didn't validate their methods -- but I think their purpose was bigger than just winning internet arguments. Their goal is basically furthering scientific knowledge. Every model and method is flawed, but some are still useful. I see this as a heroic story, and Saar's stubbornness doesn't negate his accomplishment. By gracefully hosting and funding this exercise, he's made us all better off and advanced our knowledge of important issues. And that is the most important thing.
So the Chinese want us to think that they caused a global pandemic?
Although not infallible, humans are actually pretty good at ascertaining the intentions of enemies and potential enemies, and that is why it is a universal human inclination. Being oblivious to the likely motivations and intentions of other people is a big evolutionary disadvantage.
> Newtonian mechanics wasn’t pseudoscience when Newton discovered it, but if someone argued for it today (against relativity), that would be pseudoscientific.
I disagree strongly with the ontological assumptions here. Newtonian mechanics and relativity are both great theories. Newtonian mechanics is superior for terrestrial, human-scale phenomena because it provides a simpler representation of the physical laws. Relativity is superior at speeds approaching c because the complexity is justified by the accuracy gained. To claim that Einstein > Newton for modeling the trajectory of a baseball would be the same error as saying that Newton > Einstein for modeling a black hole: both arguments would be domain-of-applicability errors.
I self-identify as a flat earther, because for 99% of daily life, the flat earth theory (viz., conceptualizing the Earth as a flat plane) is simpler and more useful than conceptualizing it as a globe. The error made by "actual" flat Earthers is not that their theory is bad - they have an excellent and very useful theory - it's that they misunderstand their theory's domain of applicability. Flat Earth theory gets you to the grocery store; global earth theory gets you to the moon.
To state my core point a bit more explicitly: the "best" theory (for any particular purpose/problem) is that which most simply encodes all information that's relevant (to the particular thing being studied). In some cases this is Newtonian mechanics; in others, it's relativity.
Quoting a comment I made on a lesswrong:
I think most points here are good points to make, but I also think it's useful as a general caution against this type of exercise being used as an argument at all! So I'd obviously caution against anyone taking your response itself as a reasonable attempt at an estimate of the "correct" Bayes factors, because this is all very bad epistemic practice! Public explanations and arguments are social claims, and usually contain heavily filtered evidence (even if unconsciously). Don't do this in public.
That is, this type of informal Bayesian estimate is useful as part of a ritual for changing your own mind, when done carefully. That requires a significant degree of self-composure, a willingness to change one's mind, and a high degree of justified confidence n your own mastery of unbiased reasoning.
Here, though, it is presented as an argument, which is not how any of this should work. And in this case, it was written by someone who already had a strong view of what the outcome should be, repeated publicly frequently, which makes it doubly hard to accept the implicit necessary claim that it was performed starting from an unbiased point at face value! At the very least, we need strong evidence that it was not an exercise in motivated reasoning, that the bottom line wasn't written before the evaluation started - which statement is completely missing, though to be fair, it would be unbelievable if it had been stated.
https://www.lesswrong.com/posts/NrKQGyggC7jcersuJ/on-coincidences-and-bayesian-reasoning-as-applied-to-the?commentId=HnxnDuqwucFAsG6bP
Counterposition here- given the verified attempts of multiple authorities to shape the "narrative" around covid, plus the powerful incentives for them to do so, does this mean that the sum of evidence available to assess the situation is probably unreliable and unfit for the purpose of this debate? The added layer of information control typical of the Chinese government makes the situation even more opaque.
Maybe this isn't a debate which can be settled based on the available information. Maybe that is by design.
The real question for me is this- why is any of this important in the long run? The obvious answer is that it has some bearing on how the world might respond to future pandemics. If it is yet another zoonoses then all the efforts to date to prevent them didn't prevent this one and probably wont prevent the next one, so maybe we should attempt to apply democratic pressure to make governments do more. I'm not sure they can do much more either way.
If laboratory research is behind covid, then this is also not completely unprecedented (though the level of genetic engineering might be somewhat new). Again, the technology isnt something that can be easily regulated by governments. Nucleotide sequence synthesis is relatively centrally controlled by a few large companies with the technical capacity, but that is only one of many ways to generate genetically modified pathogens. Again, we could attempt to apply democratic pressure to democratic governments, but again that would likely have limited impact.
The only thing that really shines through with the whole covid affair was the government and health authority overreach, panic, incompetence and opportunism. Once again none of this is new. The corporate capture of the medical system has been progressing for decades. At best the whole covid affair is just another reminder that authorities do not often have the interests of the individual as their main priority, and even when they do their capacity to deliver functional outcomes is usually highly limited. All trends point to this situation worsening in the future.
Novel pathogens will likely always remain potential black swans, which means that most of the time authorities will over react to their emergence relative to the actual risk posed. Over time this will probably build up broad social resistance to government responses until a genuinely threatening pandemic emerges. The relative difference in perceived threat between pathogens that are novel and fast moving versus those that are familiar and slow (such as tuberculosis) is also worth contemplating. Our information ecosystem thrives on novelty. Nobody is going to buy a newspaper that reports the same steady rate of slow deaths from tuberculosis or malaria. No government minister is going to win the fanatical devotion of the masses by heroically fighting such unexciting killers.
>"Peter seems to have a photographic memory for every detail of every study he’s ever read. He has some kind of 3D model in his brain of Wuhan, the wet market, and how all of its ventilation ducts and drains interacted with each other."
Mostly I think it's just keeping up with the Standard Arguments™️, from papers and preprints by zoo crew (Holmes, Andersen, Garry, Rambaut, Worobey, Pekar, Rasmussen, Débarre, Crits-Christoph, Neil etc.—the "Proximal origin" authors and their collaborators). They're the group of Western virologists who ~yearly come out with a new paper finally proving market zoonosis for real this time (Andersen et al. 2020, Holmes et al. 2021, Worobey et al. 2022, Pekar et al. 2022, Crits-Christoph et al. 2023a, Crits-Christoph et al. 2023b, etc.) and get a ~yearly media circus about it. Maybe he also read some pop-sci books like Chan & Ridley's "Viral" for pro-lab (2021) and Quammen's "Breathless" for pro-zoo (2023). Probably also the Chinese HSM-sampling paper (Liu et al., 2023) and WHO-China joint study (2021), following zoo crew in how to interpret them.
E.g. Peter's heat map slide is just Fig. 1 of (Crits-Christoph et al., 2023a), which is an elaboration of Fig. 4 of (Worobey et al., 2022); and his "drains" argument is from Fig. 2C of (Crits-Christoph et al., 2023b).*
References
Andersen, KG. et al. (2020). The proximal origin of SARS-CoV-2. Nature Medicine 26: 450–452. https://doi.org/10.1038/s41591-020-0820-9
Bahry, D. (2023). Rational discourse on virology and pandemics. mBio 14: e0031323. https://doi.org/10.1128/mbio.00313-23
Bloom, JD. (2023). Association between SARS-CoV-2 and metagenomic content of samples from the Huanan Seafood Market. Virus Evolution 9: vead050. https://doi.org/10.1093/ve/vead050
Chan, A. and Ridley, M. (2021). Viral: The Search for the Origin of COVID-19. HarperCollins.
Crits-Christoph, A. et al. (2023a). Genetic evidence of susceptible wildlife in SARS-CoV-2 positive samples at the Huanan Wholesale Seafood Market, Wuhan: Analysis and interpretation of data released by the Chinese Center for Disease Control [preprint]. Zenodo. https://doi.org/10.5281/zenodo.7754299
Crits-Christoph, A. et al. (2023b). Genetic tracing of market wildlife and viruses at the epicenter of the COVID-19 pandemic [preprint]. bioRxiv. https://doi.org/10.1101/2023.09.13.557637
Holmes, E. et al. (2021). The origins of SARS-CoV-2: A critical review. Cell 184: 4848-4856. https://doi.org/10.1016/j.cell.2021.08.017
Liu, WJ. et al. (2023). Surveillance of SARS-CoV-2 at the Huanan Seafood Market. Nature [online ahead of print]. https://doi.org/10.1038/s41586-023-06043-2
Pekar, J. et al. (2022). The molecular epidemiology of multiple zoonotic origins of SARS-CoV-2. Science 377: 960-966. https://doi.org/10.1126/science.abp8337
Quammen, D. (2023). Breathless: The Scientific Race to Defeat a Deadly Virus. Simon & Schuster.
WHO (2021). WHO-convened global study of origins of SARS-CoV-2: China Part. https://www.who.int/publications/i/item/who-convened-global-study-of-origins-of-sars-cov-2-china-part
Worobey, M. et al. (2022). The Huanan Seafood Wholesale Market in Wuhan was the early epicenter of the COVID-19 pandemic. Science 377: 951-959. https://doi.org/10.1126/science.abp8715
*The heat map is misleading: it was based on implausibly assuming all sampled stalls were sampled equally; in fact the wildlife-corner stalls were sampled far more heavily (Liu et al., 2023). For a critique of dismissing ascertainment bias, see my (Bahry, 2023). For a critique of Crits-Christoph et al. 2023a, see Bloom (2023). For a critique of their later, also-misleading heatmap (Crits-Christoph et al. 2023b Fig. 2a), see (https://www.biorxiv.org/content/10.1101/2023.09.13.557637v1#comment-6279754825).
I find the "drains" argument mildly interesting (Crits-Christoph et al., 2023b Fig. 2c): nonzero evidence, but just one real datapoint. (The two downstream of the racoon dog stall are *in general* downstream of a big chunk of the western wing and the entire western wing respectively, so those mean little.])
I guess I don't quite understand your point. Scott points out that Peter has an eidetic memory for the arguments in this debate, and you respond: "Well, he just memorized the studies and arguments, especially those that support his view."
Aren't you saying the same thing as Scott? Surely, we wouldn't expect Peter to actually author the papers himself. So what else is there to do than defending his position using other people's work?
And if Peter's understanding of the research is actually much shallower than it looks to a layperson (which is I think what you're saying), that seems fairly embarassing for Rootclaim, no?
My point is that Peter was well-prepared—but it was the high end of the normal range of preparedness for someone who'd been following the topic, not savant-level (compare to Scott's "I can’t imagine how many mutations it would take to make me even a fraction as competent as Peter was").
E.g. to make the drains argument, he didn't need a working 3D model of every brick and rivet in the market; he just needed to follow zoo crew on Twitter, see one of them share Crits-Christoph et al. 2023b, read it make the drains argument, and see its Fig. 2C.
Note that this comment isn't intended to disparage Peter, just to bring things down to Earth—and to give some pointers on where the arguments came from so they can be seen in their original publications, so readers can evaluate them further, etc.
(It is meant to disparage zoo crew though lol)
I agree that Rootclaim should have been more prepared
Scott- I can roughly translate my factors into the grouping you use in your table.
Net priors using your grouping)
me: 12 you: 16
combo of all HSM/lineage factors
me: 1 you 0.002
FCS-ish
me 25 you 25
reasons WIV wouldn't do it
me: in priors you: 0.17
cover up success
you 0.5 me: did it?
other factor restriction enzyme pattern
me 70 you NA
The RE pattern is new, and I think you'll find my argument at least reasonable.
The big disagreement is about Worobey/Pekar. If I were to argue for ZW, I'd stay the hell away from Pekar, whose errors are now verging on something worse than "wrong".
Worobey does correctly point out that there was a spreading event at HSM, but so many specifics (lineage, internal DNA-RNA correlations, Wuhan share of relevant species (if any) don't fit that ZW has to rely on the non-market versions, which actually were generally considered before hand to have more of the priors than a market version- including by DEFUSERs.
Why don't I then get stronger LL odds? There's a lot of uncertainty in the priors. I do "robust Bayes" to take that into account, integrating over a broad fat-tailed distribution of priors. One tail keeps ZW in the game. Changing likelihood factors makes much less difference in the final odds than you might expect, because the enormous likelihood ratio has already been diluted down by the uncertain priors.
Lots of people in the comments seem to have additional evidence one way or another that they want to add in. Maybe we need a coronavirus version of TalkOrigins.
I just cannot get myself to the point where I take anything being stated about this as reliable. For 2-3 years, everything that was being said about COVID by "experts" was either on what Zvi calls Simulacra level 2 (the public health establishment; saying X, despite ~X, to make you believe X and act in a way they want you to) or on what he calls Simulacra level 3 (Twitter users saying something not because THEY believe it's true, or even that they expect YOU to believe it, but just to signal that they're good non-racist Science-lovers). So even while I find Peter's points as you've summarized them to be pretty reasonable, he hurts himself IMHO by making the characterization of lab leak as a "conspiracy theory", because accusing an opponent of that also happens to be great tribal signaling. And the data on which all of this is premised comes from sources like the WHO which is clearly willing to lie for a variety of reasons.
So despite finding Peter's arguments here pretty reasonable, I can't get over the hump. I also, were I as dishonest as the PH establishment, would be inclined to just say it was a lab leak to induce public fear of gain-of-function research to help eliminate it. I can live with that having been wrong for this one virus, but still wanting to ban such research because next time it may NOT be wrong, but since most people don't reason that way then zoonosis being true is quite a blow to efforts to end this dangerous stuff. So let's give public health a dose of its own medicine and keep a nice healthy percentage on lab leak regardless.
Sneaking in late just to say this is my second ever favorite post ever, can I change my survey answer to NOT favoring lab leak? : )
Even after this excellent post, I find Michael Weissman's v5.7 analysis from March (https://michaelweissman.substack.com/p/an-inconvenient-probability-v57) to be the most compelling information available online to date, and it contains significant evidence updates from his v5.6 post that you linked in the write-up from January. I've kept an open mind through the years about Covid origins and found the Rootclaim debate to be especially interesting and cause for shifting my probabilities, but I have not encountered anything that has so significantly updated my probabilities than Michael's latest v5.7 analysis does. I'm curious if people have read that analysis in full and what everyone's thoughts are.
You seem to initially be interested specifically in Michael's Bayesian probabilities to compare his results to your table (which I think he posted recently in the comments above). I understand that you likely have dedicated dozens of hours to this post already and reading through a massive updated analysis is likely not especially appealing, but just looking at the probabilities alone does a disservice to the immense amount of research, evidence, and analysis that is included in Michael's write-up. For what it's worth, I've read an incredible amount of high-effort research and debate on this topic throughout years of following this, and Michael's research/analysis is the first one that I find convincing. In addition to more convincing Bayesian logic, his research includes crucial compelling evidence/information that came out after the debate (or was available but dismissed or potentially misrepresented at the time of the debate). I am trying to be as fair to both sides as possible and I honestly can't see how someone can read Michael's v5.7 post in full and come away with a 90% probability of Zoonosis. In my mental model, the two most likely reasons are that either people haven't read Michael's 5.7 analysis due to low exposure of a small blog, or I am fundamentally misunderstanding something.
Either way, I want to know what people think. If Michael's post isn't as convincing as it seems to me, I'd like to know why so I can update. If it is convincing, I'd like others to know so they can update. @Scott, I historically have placed a lot of value on your opinions and I'm genuinely interested in what your thoughts are here.
Thankyou for writing this up.
Personally, I don't have a lot of emotional/ideological investment in the issue so wouldn't sit through the debate but it great to get a detailed enough write up of the main arguments and their problems. I guess I should be over it, but I remain amazed by the level and vehemence of the responses here and elsewhere. Lab leaks remain possible whether or not it happened here so there's an ongoing risk anyway. There are plenty of risks. In the long term, I'm actually more worried about the intentional release of an engineered virus than an accidental leak.
The Rootclaim response in case it hasn't already been shared. Although the main issue to me is several papers have superseded the key papers relied on for the debate anyway. https://blog.rootclaim.com/covid-origins-debate-response-to-scott-alexander/
Peter had more at stake than Saar.
100k is about 0.05% of Saar's net worth. I assume it is a much much larger share of Peter's net worth. More to lose. More to gain.
The incentive for debate performance is massively skewed toward Peter in this case. Worth noting when analyzing the truth-seeking utility of this format.
Sorry if this is noted elsewhere, it is a big comment section.
Big comment section? Nah, it's barely 100,000 words. You should be able to knock it off in a day if you don't break for lunch.
This was such an interesting read. There are a lot of smart people here in the comments continuing the discussion, but I just wanted to chime in and say THANK YOU for writing something so compelling to read for somebody (me) who has no prior knowledge or real opinion on the subject(s). Legend ⭐
I believe it is absolutely impossible that an animal or animal specimen transported to Wuhan lab for study caused the pandemic.
I only believe that an animal or animal specimen transported to the Wuhan wet market to be EATEN, could have possibly caused the pandemic. Obviously.
I believe the Chinese govt is LYING when they say no pangolins, bats or racoon dogs were sold at the market at that time. I believe the Chinese govt is LYING that the only COVID found in the market were on fish, who can't get COVID .
I absolutely believe the Chinese government when they say with exact certitude when and where the earliest cases were, so I can see my dots on the map. The cases neatly surround the fish on the map where the COVID was found.
I believe the Chinese govt when they say they took down their virus database in Sept because of fear of hackers. Cause you know, hacking had just been invented. I also believe the Chinese govt when they say they increased safety protocols at Wuhan just because. I understand the Chinese govt blocking attempts to get to the bottom of this and silencing doctors and scientist who tired to warn the world.
I believe the former head of the CDC, the FBI and the Department of Energy are all wrong on their beliefs of the Covid orgin.