"Throughout"? Nay, that was weeks beforehand. It takes time to book an appointment with a psychologist, get the gender dysphoria diagnosis, go to the pharmacy, etc.
Additionally, the contents of the post detail collected impressions over quite a long period of time, from late 2021 to present. It would have been impractical/expensive to confound this with something short lasting like ketamine!
17. Seems hard to draw much of any conclusions about when the trend started given the massive gap in data before 1910 and between 1930 and 1980. What the hell happened to poetry at those times?
25. You will however go to hell for being a Cubs fan.
Yeah, the data in #17 seem incredibly fishy. Looks like just fitting to artefacts.
Looks like a random spread before 1930 (though with suspicious peaks at 0/25/50/75/100), then virtually nothing until 1980, then a vast amount of un-rhyming poetry collected since ~2010 or so pulling the average down. And why was there loads of poetry in one year (maybe 2015)? It's all so bizarre.
Like, look at the actual data from 1930-1980. In what world is that red trend line a fit to those dots?
The dots are individual poems. Poems with strict rhyme schemes score round numbers. ABAA is 75%. 25% is weirdly low, though.
It would be good to look at other averaging methods, like dichotomizing < 25% or maybe <= 25%, which is probably what people mean, not a switch from ABAB to ABCB.
Copyright issues for the 1930-1980 gap? And maybe people are just less likely to put those pre-1910 poets like Robert Bridges online than they are Pound and Eliot.
To Scott's point, non-rhyming doesn't equal modernism. Regular meter might be an interesting thing to measure like this, although it might be tough. Poems in iambic pentameter are going to have plenty of lines with variations, so it might be harder to see a pattern over time.
Gap is likely explained in this way: Before a certain year, poems are in the public domain and easily findable via book archives. After a certain year, poems are commonly published online. Between that is all the stuff that was published in print, but is still under copyright, and so is less available for automated searches.
17 seems like a good opportunity to pound my hobby horse that song lyrics are by far the most popular and pervasive form of poetry from the 20th Century on and it's silly to talk about poetry in general and popular poetry especially without including them.
#23. This is also true of skateboarding and has been for over two decades. Jordan’s shoe deal reworked this category and sneaker heads amplify. Consider only two people have dedicated shoe facilities at Nike. Jordan and Janoski.
It’s an interesting meeting point for urban culture and sport.
#10: Frankly this isn't very surprising to me. I'd imagine this graph correlates very strongly with the popularity of figures like Andrew Tate and etc with this age group. The amount of internet celebrities popular with teenage boys that express this kind of sexist attitude really started to take off ~2020. Would be interesting if there is any data on the popularity of those figures over time.
Supporting a society in which women get paid less than men for the same work is one thing, but actively believing that it’s actually better to pay women less for doing the same thing seems more egregiously sexist than I can believe 20% of people are.
I feel like that question is really just measuring how important boys think it is to virtue-signal that they aren’t gay liberal cucks, and apparently it has become noticeably more important to them since 2018.
Even if there is such a correlation, it wouldn't answer whether those figures are causing these beliefs, or becoming more popular because they're validating the beliefs people are picking up for other reasons.
I wonder if based on the questions some of it is "fighting" with the survey team. I'm reminded of the classic Onion article "1 in 5 Americans believe Obama is a cactus": with some of these questions you can recognize "ah this is one of those *woke* questions, I'm gonna give them the *based* answer just to piss them off!" I remember reading a while back on how ~10% of high school boys report being certified submarine captains.
A way to confirm or refute this theory would be with a somewhat more oblique question or experiment that tries to get at someone's "real" views on gender. Maybe one of those money-negotiating experiments (like the Ultimatum Game) where you experimentally manipulate the name + photo of the person you are negotiating with, to see if these "I hate women" survey boys actually try to give less money to girls (even at the cost of money for themselves.
Kind of a tangent, but when they actually interviewed a panel of official submarine captains and asked them what they most fear down there, the number one answer (by far) was, of course, “dogs”.
It's not surprising, just like it's not surprising that figures like Tate are popular when we just spent 20 years telling boys (especially white ones) that they are evil, useless, and should be discriminated against. Of course they are going to lash out at the people vilifying them.
47. But did you know that this effect comes almost entirely from daughters?
I really really struggle to believe this. Unfortunately the post is paywalled, but I'm pretty sure I did read it, or a similar post; and the study it's based on is not very big- around 400 women total (200 bearing daughters).
They look at only the gender of the first child, but ignore the evidence (which they do cite, but don't account for) that first born girls are more likely to be followed by a sibling than first born boys. That is, the effect could be entirely that the mothers of first born daughters have larger families, and hence reduced employment. They kind of sweep it under the rug by saying "if parents anticipate building a larger family, they may be more likely to specialize, with one more focused on market work and the other on home production."
I don't have time to read the paper, but do they control for how long the mothers left the workforce? How about for multiple kids? People always say daughters are easier to raise when they are young, so I could see mothers of daughters being more likely to spend more time with them before returning to work. I could also see families that start with a daughter being more likely to have extra kids because the father wants a son.
#5 Would you let a dead spider, controlled by AI robotics, perform surgery on your eyeball? If the alternative was going blind? No. I would rather go blind.
#9 - Come back 20 years from now and the following sentence will have aged like milk: “Looking at . . . the current state of Meta’s VR universe, it is absolutely impossible to imagine a world where this kind of thing is ever going to succeed.”
Once you spend enough time trudging through the internet's gamer communities, you realize that there are millions of people out there who would drop everything real in their lives if they could "live" in virtual worlds, embodying a perfect version of themselves (often an all-powerful character) and creating romantic links with other people's avatars. Meta's strategy with VR has been to buy everything and everyone who does anything remotely interesting in that space - until the technology catches up (GPUs that can render realistic worlds, systems that suppress motion sickness and improve real world mobility -> virtual mobility) that technology looks like a massive failure on all fronts because there's almost no revenue to be found, and people who already have a dislike for video games find the whole concept extremely virgin-coded (they aren't wrong).
>Once you spend enough time trudging through the internet's gamer communities, you realize that there are millions of people out there who would drop everything real in their lives if they could "live" in virtual worlds, embodying a perfect version of themselves (often an all-powerful character) and creating romantic links with other people's avatars.
ETA: Second Life is still there, alive and well, in case that wasn't clear, and currently celebrating its 22nd birthday. The film I linked to is a celebration of what is, not a lament for what was.
A guy decided to check out an old virtual reality game from the mid-2000s that was still running. At first it seemed like no one was there, but then he ran into some users. It went places I didn’t expect and in the end I got a little verklempt.
Second life is still alive and well, but has mostly been totally supplanted in the VR space by VRChat! VRC is basically the modern day second life, with a few million people owning it and tons and tons and tons of active users. It is what i imagine Meta wanted it's metaverse to be, but i don't think Meta wants all the gay furries.
Of course the concept of living in cyberspace has been around since at least William Gibson. But most people don't want to just sit and interior decorate, and definitely not with PS2 level assets. They want something to do, a world to explore and a game to play. Then they also want to be doing the popular thing that other people are doing. What does it mean to be "an all powerful character" in Horizon Worlds?
In their defense most of the money is probably going to hardware development of the headsets which have real engineering challenges. But if you wanted to make a virtual world that people live in, really you don't need VR, you need to seamlessly merge the gaming space with a discord-like multimedia sharing app, with social media, and then with outside corps and organizations. People already basically do live in that cyberspace it just isn't integrated as it likely could be.
The thing is VRChat is still just chugging along - I just wander around weird worlds random people have created about once a week. It's an amazing experience that's completely at odds with Meta's culture and business model, and I say this as a former Meta engineer.
The metaverse kind of exists - but only for a small population of weirdos. Something like 60% of them are furries, and there's not much money in it aside from independent digital creatives selling custom avatars.
I think VR will persist as a niche weird thing for the foreseeable future. If Meta abandons it I expect a few other companies to fill the gap, albeit with less excessive subsidies. Valve is a private company beholden only to the whims of a handful of nerds with access to a free money printer and a demonstrated interest in making stuff that they want to make in VR.
If meta made VR chat with the GDP of the country I grew up in behind it instead of worse second life, I would really feel the urge to get inside the GiveUp-Machine.
If Meta did that it wouldn't have the qualities that make VRChat what it is. VRChat works because it's completely dominated by a small set of weirdos and has little mainstream appeal. This protects it from the profit incentives that would lead to metric-maxing optimization, and then you get what every mainstream social media converges towards.
You want enough capitalism to push people to make better stuff but not so much that "exploit every weakness of human psychology at industrial scale" becomes a strategy worth pursuing.
The idea behind gaming is to escape the real world (in a certain sense). I don't pay close attention to Meta tbh, but from what I remember, they seem all too eager to bring the real world with all its irritations to the virtual one.
Doesn't it have huge implications for non-virgin coded non-video game applications? Enterprise stuff, realistic essential worker trainings and so forth.
"you realize that there are millions of people out there who would drop everything real in their lives if they could "live" in virtual worlds, embodying a perfect version of themselves "
Until that becomes an actual option, we really don't know. Millions of hippies swore they were willing to drop anything to live in a commune *if only*...
"Meta's strategy with VR has been to buy everything and everyone who does anything remotely interesting in that space - until the technology catches up"
Except, there is no guarantee the technology ever catches up. Also having a VR set, even much better than the current ones, is still not quite "Ready Player One". How many technologies have been 2 years away for 30 years? And how many technologies that were believed to be the next Industrial Revolution just became a niche hobby little influence on most people's lives? 3D printing is an obvious examples
#26: perhaps link to the story on substack: https://alexsalvinews.substack.com/p/the-election-of-pope-leo-xiv , to avoid the dozens of clicks required to read it on X? I'm sure there's a way to "unroll" the thread on X, but a webpage where that's already done for you seems good.
In general, it’s really annoying to see a link to a Tweet, figure there’s probably more to it than just one Tweet, and then realize that there’s probably a thread, and you’re just not allowed to read a thread if you aren’t logged in to a Twitter account.
"Then universities borrow money now against that future windfall. What goes wrong?"
The same thing that goes wrong if I go borrow a million quid based on "next week for sure I will win the lottery so I definitely will be able to pay you back, honest!"
"I thought you weren’t allowed to do any of this, and certainly not talk about doing any of it, on pain of excommunication?"
We're Catholics. You seriously expect us to stick to the rules? 😁
The difference is that you probably aren't going to win the lottery, but it's overwhelmingly likely that Democrats will have unified control of the U.S. federal government at some point in the next several decades. Not saying there aren't reasons why this might not work, but that isn't one.
Of course they will, the right hates everything about this. The whole argument (which, again, is not entirely free of holes) is that the Democrats will eventually have the opportunity to pass legislation on a party-line basis over the right's objections.
It won't just be the right, I wouldn't like to be the Democrat official charged with explaining why a moneybags Ivy League college is getting taxpayer money to repay their borrowings, while [insert issue here: housing crisis, potholes in the roads, systemic racism, healthcare costs, etc.] are crying out for funding.
Oh, the Democrats are likely to get back into power. But when they do, will the promised money be there? Or will it be spent on other, more "worthy"/"needy" causes? Suppose Harvard borrows against the promised payout, do you really think there won't be outcry from the more lefty side of the coalition about rich endowment white privilege colleges getting a government bailout instead of state schools., or black colleges, or any of the rest of the list of "more urgent social needs"?
See Biden's student loan forgiveness, yes it did happen, but not as broadly as people thought it would:
"Despite failing to deliver his promise for broad student loan forgiveness, President Joe Biden has now overseen the cancellation of student loans for more than 5 million Americans — more than any other president in U.S. history.
In a last-minute action on Monday, the Education Department canceled loans for 150,000 borrowers through programs that existed before Biden took office. His administration expanded those programs and used them to their fullest extent, pressing on with cancellation even after the Supreme Court rejected Biden’s plan for a new forgiveness policy."
1. Loan forgiveness was severely curbed by the courts (fortunately, imho). Unlike the current admin, the previous one was in the habit to respect sentences
2. We are not talking about huge sums, at least when compared to the national budget. For the left to be able to block a 0.x% addition to the budget, that the center-left campaigned on and publicly promised, they'd need a level of control not even remotely close to the one they have
3. The left (as in, the Squad) has been pretty vocal about the selective grant cancellations and has repeatedly framed it as an attack on research rather than on Harvard. Nobody guarantees consistency obviously, but to completely renege their current position for no appreciable reason would be deeply misguided and there is no reason to think they would.
> Democrats will have unified control of the U.S. federal government at some point in the next several decades
I disagree. While I do agree that *some* left-leaning party will have unified control at some point in the future, the present-day Democrats have decayed past any point of recovery. Their putatively successful political descendants will have to deliberately distance themselves from present-day Democrats to such an extent that making high-confidence predictions about their policies is difficult (if not outright impossible).
Are you saying that there'll be an ideological realignment like in the 20th century, or are you literally saying that the United States Democratic Party will never again have unified control of the federal government? If the latter, well, I think that's less than 20% likely and would bet accordingly.
I wouldn't say "ideological", I'd say "political" or perhaps "messaging" realignment (as I doubt that politicians have anything resembling a consistent ideology). And there will likely be an organization *named* "US Democratic Party" that may, in the next 10..30 years, gain full control of the US Government. But that organization's internal structure, policies, messaging, strategies, financing priorities, etc., will be so different from the current Democratic Party that making confident predictions about their behaviour is premature. Otherwise, they'd have no hope of taking any kind of control over anything.
The problem is planning for a sure thing, and planning for a time period. If the Democrats regain control of the federal government in 2028, that looks very different for a university carrying large loads of debt than it does if Democrats regain control in 2036. Also, maybe the Democrats don't back pay any grants, or at a lesser percentage than full, or even decide that it worked out pretty well and don't even add new grants. Then the universities are in a really bad position. This would be a very bad bet to make.
Any bet made at the whims and capability of elected officials to do something that they may or may not even want to do is a bad bet. There is no guarantee that Democrats as a whole would make such a change, let alone individual Democrat politicians. There is even less guarantee that they would have the ability to do it, such as failing to get all three branches, having too small of a majority to pass disputed legislation/filibuster, etc.
Maybe, but there is nothing stopping them from selling state-contingent claims (x$ will be paid to the bearer *if* Dems win AND make good on their word). If they trade at 10c on the dollar, too bad, it's not Harvard's problem as long as they have buyers.
Also, once these state-contingent claims are on the market, they create a natural, powerful constituency pressuring Dems to in fact make good on their word.
This depends on a much longer chain of hypotheticals than that.
For this scheme to work, it's necessary that 1. Ds regain unified control sometime in the next decade or two (pretty likely), then 2. those Ds retain enough respect for commitments made today that they make a serious effort to pay out the (probably very large) back funding, and 3. that effort successfully beats out all other competing demands for funds, in an increasingly strained budget environment.
Then there's the whole question of timing. This bet looks very different depending on whether you anticipate the payout in 4 years, vs 8, 12 or 16. The further into the future it is that Ds regain control, the less likely that points 2 and 3 above will succeed. Then, of course, there's the time cost of money to consider.
Meanwhile, this policy requires a very public commitment and funding negotiations; thus, it can't be done quietly. Trump has all kinds of ways to interfere with this process, from going after the lenders in various ways (all banks are extremely vulnerable to USG pressure) to using other levers against the universities.
Think about this as if you are the prospective private lender. Would you be willing to pay out $billions to universities, with repayment contingent on someone else fighting and winning a thankless political struggle in an unclear number of years? I am pretty sure that would not pass any conventional underwriting process.
Whereas my comparison would be paying in US Dollars during the War for Independence, or in Confederate dollars during the Civil War. It's yet another dangerous escalation in the political landscape.
44: Ultimately, I think the most concise way to think about "The purpose of a system is what it does", is to see that most people are split into two diametrically opposite interpretations of it:
1) What a system does must be the purpose that was intended by its designers
2) The "purpose" of a system is a concept that can, and should, be described independently of any intent or design
And the way to resolve this is to realise that POSIWID itself is a very confusing sentence out of context, and Stafford Beer (presumably) never intended it to confuse people, so:
Under interpretation 1), POSIWID is a self-contradiction;
Under interpretation 2), POSIWID proves itself - its a meme that perpetuates discussion of itself by the conflict between both interpretations, and it's useful to imagine that as its "purpose" in the context of asking how it sticks around.
If a system is sufficiently "bad" it will be torn down. I think that's the best way to hear POSIWID. Democracy, worldwide, is used as a tool to let the CIA/United states government influence elections and policy in other countries.
It seems like what people ordinarily point to when using the word "purpose" is the ways in which a thing promotes some given terminal goal- so not just the intent of a designer, but also not really independent of intent.
Most of the people doing the pointing, however, don't actually have a solid grasp of terminal goals as a property of agents and distinct from instrumental goals, so the way people end up conceptualizing purpose tends to be more like "the ways in which a thing promotes [unnamed, half-formed concept]". People aren't sure whether that half-formed concept is an objective property of reality or something subjective, so we get lots of very confused memes like "the meaning of life" that assume purpose can exist independently of any agent.
Reading through the linked blog post, it sounds like those early cyberneticists noticed that anything a thing does could constitute its purpose to some hypothetical agent with some particular terminal goal. But they also didn't really have a solid concept of terminal goals and were operating under this confused cultural assumption that purposes are properties of things that can exist independently of agents, so the observation just looked to them like "anything a thing does could constitute its purpose"- phrased even more confusingly as POSIWID.
#15: yeah, someone whose funding gets cut and loses their academic position is most likely to look for a job in industry, not to start a startup. Launching a startup (at least in biotech) usually requires preliminary data which you can't get if you don't have a lab.
When we recently hired for a PhD-level position we got over 300 applicants the first day alone.
#16: an interesting idea, but what if the Republicans win in 2028?
I don't know, Stephen Fry also seems like the kind of person to be into this. The surprise is that he chose to spend his weirdness points this way given that he has fewer to spend as a much more mainstream figure.
#8 - I can get 1/100th of the karma of saving a life by simply praising good deeds? This does not seem remotely play-tested. Surely praising good deeds a hundred times is easier than saving someone's life. If paying someone to praise good deeds on my behalf counts, I'm sure I can hire someone to do that at a much more effective rate than it would cost to spend the money saving lives.
And "adopting an orphan", something that brings with it daily obligations for years, is just 50 karma! I'm never going to adopt any orphans with this scoring scheme. (Though maybe it made more sense in Ming China, where children may have been economic investments rather than money sinks. But still.)
On the other hand, saving lives is *also* surprisingly cheap. GiveWell generally gives estimates of $3500-$5000 in the modern world, and perhaps it was even cheaper in Ming China due to more people near you living close to the subsistence threshold.
On the other, other hand, you don't merely have to pay someone to praise good deeds; praising is the easy part. You have to pay them to identify good deeds! Which is such a thorny proposition that modern philosophers still have not solved it.
I think it's about $2 to keep a child from going blind. That may not be saving their life (or it might be, consider orphans and other destitute), but it's certainly keeping them as "valued members of society."
I always like to run that as a metric for "How Much does this charity deserve to be funded?"
Look at the link below... Looks like we still need more money/effort on this front, hundreds of thousands of children are still going blind (and that's with a 50% mortality rate after blindness).
Do you have a citation? Usually when cost numbers are that low, someone is conflating 'cost per treatment' with 'cost per preventing negative outcome', which are only the same if 100% of treated people would have the negative outcome without the treatment and 0% would have it with the treatment.
We have upwards of 200,000 children a year going blind (with a mortality rate of 50% in the first year alone after blindness). This is after distributing 400 million doses per year.
"Globally, VAD is estimated to affect about one-third of children under the age of five, causing an estimated 670,000 deaths in children under five annually.[8][9] It is most prevalent in sub-Saharan Africa (48 percent) and South Asia (44 percent).[8]"
Yes, this is cost per treatment (vitamin A shots). To "fully prevent" Vitamin a Deficiency you need a shot a year...
OK, let's assume that it takes 4 shots then, to fully prevent VAD (since this is largely a <5-year-old issue.)
Your paper doesn't give blindness stats (which I know was the initial claim), but it does give all-cause mortality improvement. It drops all-cause mortality from 2.6% to 2.3%. That's 3 lives per thousand treated children, or per 4000 doses, or about $1,333 per life saved. Very, very efficient!
But... while this doesn't give blindness data, note that 50% of kids die within a year of developing blindness. That caps the # of blindnesses prevented at twice the number of deaths prevented (which would be true if all the mortality improvement was due to preventing blindness; unlikely.) That means that you have to pay at least $666 per case of blindness prevented (and probably quite a bit more.)
Still a bargain; but not wildly implausible like $2 apiece would be.
GiveWell's rough estimate is more like $6,000 apiece.
Each treatment costs $2.70 ($1.35 per shot, but you need two shots). Divide that by the probability that the child would have gone blind without the treatment but won't with it, which they calculate by multiplying the following numbers together:
- 1.3% prevalence of night blindness in Guinea, Burkina Faso, and Mali
- 10% of cases of night blindness progress to full blindness (there apparently are no studies of this so this number is a guess but they think it's a generous one)
- 68% chance the treatment works
- 50% chance of surviving early childhood after going blind (as your comment accounts for)
They then go on to calculate DALYs (divide by 0.187 disability weight for total blindness, and by 25 more years of average life expectancy for a blind child who doesn't die right away) and conclude that blindness prevention through vitamin A supplementation costs about $1,200 per DALY. This is worse than giving cash, which they estimate improves lives at a rate of $1,000 per DALY. So ultimately not the most good you can do.
They recommend Helen Keller International because of the mortality benefits, for which the numbers are better ($3,500 per life saved, not sure the exact conversion factor to DALYs but it's obviously way better than cash).
I think it's clearly written for a different society, where words had more weight, anonimity was not an option, and if you told a butcher to their face they had to change job you could have been in for an unpleasant retort.
Presumably, as both the material impact and the cost of praising good deeds and harassing butchers decreased, the karma points would decrease in proportion
#1 - Scott, you say that post helped you crystallize some thoughts on what good writing and good taste are in general. Sounds like the perfect topic for you to write about - will we get a glimpse?
Despite having a successful therapy, VERV stock really didn't do well over the last 4 years. They recently got acquired by Eli Lily for $1.3B but this is down from about $2B in 2022 (and recently it was as low as $300M).
I personally lost about $15k on this. Just goes to show that biotech is hard to make money in.
That is odd. It seems like the stock should be higher than that. Any idea why it's so low? Even on the day the results were announced (April 14, 2025), the stock only went up ~$0.90 to ~$4, off from a peak of >$70 four years ago.
> looks like the rhyme decline started well before then.
Paradise Lost, an epic poem published in 1667 by John Milton doesn't rhyme. I gather many people were critical of it at the time for that reason. My only complaint about Milton's poetry generally is the plethora of ancient Greek characters, divine or otherwise, who pop up in it. Presumably readers in the 1600s would have been familiar with these references, but most people today find them obscure and baffling. Piers Plowman, written towards the end of the 1300s, doesn't rhyme either.
Doesn't Milton treat rhyming as newfangled? My understanding is that English poetry didn't rhyme early on (Beowulf used alliteration, Shakespeare used patterns of stressed syllables), and that rhyming was introduced via French poetry.
AFAIK, classical Greek and Roman poetry doesn't rhyme either -- it's based on metric structure, like Old English and Sanskrit poetry (which suggests it's an Indo-European thing). Poetry in Latin starts rhyming consistently in the Middle Ages.
Shakespeare doesn't predate the initial burst of English rhyming poetry, he came in towards the end of it - his sonnets with their ABAB CDCD scheme were a compromise away from the ABBA ABBA Petrarchan sonnets that had a vogue in the 1500s (which are far easier to write in the very rhyme-friendly Italian lexicon than the English one).
Yes, rhyming was from French poetry. As with all barbaric, agglomerating English -- rhyming is much, much more difficult in English than languages that do it natively.
#48: Bret Deveraux notes that the "strategic" effects of the air campaign against Iraq, i.e. strikes on leadership, communications, electric power etc., were minimal.
> The DoD’s own report, issued in 1993 (Cohen and Keaney, Gulf War Air Power Survey (1993)) noted first that only some 15% of strikes were against ‘strategic’ targets, while strikes against Iraqi ground forces consumed 56% of strikes; as Coalition air forces exhausted their list of strategic targets, they switched over to strikes against ground forces.
> Despite effectively running out the entire list of strategic targets, Cohen and Keaney nevertheless conclude that strategic effects were broadly not achieved. Despite striking the Iraqi communications network with more than 580 strikes, “the Iraqi government had been able to continue launching Scuds” and “sufficient ‘connectivity’ persisted for Baghdad to order a withdrawal from the theater [Kuwait] that included some redeployments aimed at screening the retreat.” Consequently, “these attacks clearly fell short of fulling the ambitious hope” to “put enough pressure on the regime to bring about its overthrow and completely sever communications between Baghdad and their military forces.”
So no, strategic air power was neither a necessary nor a sufficient reason why the Iraqi army crumbled. They crumbled because the Coalition forces represented the undisputed peak of modern warfare on a perfect battlefield (flat desert) whereas the Iraqi forces had neither the equipment, nor the training, nor the doctrine to resist effectively. There is no reason to assume that an invasion of Taiwan would be the first proof of the effectiveness of strategic airpower after a century of failed tests.
I would generally caution against using Deveraux as an authoritative source outside his area of expertise, i.e. Roman era military history. The post about strategic bombing draws on WWII for a lot of the material, where bombing was highly inaccurate and the belligerents had strategic goals not clearly linked to military objectives like "bomb the civilians to make the enemy regime collapse" (which definitely didn't work). Anyway, what #48 is proposing isn't strategic bombing in the sense Deveraux is talking about, where air power alone is the deciding factor. He's talking about a combined arms approach where the enemy forces are crippled by precision strikes taking out key assets - so that a ground force can then overwhelm the disorganized enemy. The ground force is key here. But it doesn't matter for Taiwan because the ground force has to be lifted across an ocean for a contested amphibious landing, which is an entirely different type of operation.
I'm not sure what the rest of the ACOUP post has to do with my objection. I presented a specific quote from the post, not the entire post, because that quote dealt with the question of whether you can use air power to disrupt an army to the point that it simply breaks down at the slightest pressure. That was the theory developed for the 1990 Gulf War, it was tested in that war, and unless you're saying Deveraux misquoted or misunderstood the DoD report, the answer seems to be "no, that theory didn't work". The #48 tweet repeated that disproven theory, elevated it to fact, and builds its argument on that faulty foundation. That is why I object to #48.
Yes, that's the point. That is very different from the imagined strategic use of airpower, i.e. to basically win the war entirely from the air rather than by providing boring old ground support.
#9 - Reporting on how much money Meta has lost on "the metaverse" often conflates their crappy-looking VR game Horizon Worlds with their entire Reality Labs division. It's the division that's lost $70 billion, but it also makes VR headsets and does all sorts of pie-in-the-sky VR/AR research.
Horizon worlds has by all accounts been horribly mismanaged, with developers that refuse to actually put on a headset and test the game, but not quite to the level of spending 50 Star Citizens worth of money.
The brute fact of software engineering being extremely inconvenient to do while wearing a VR headset may be the single biggest obstacle the technology faces.
# 17, the poetry data from the economist could equally support the hypothesis that rhyming poetry has survived better. They clearly have a lot more data for recent poetry, and the dataset is short on unrhyming poetry from the past. They also clearly have a lot of rhyming poetry from today, but it’s getting overwhelmed by the unrhyming stuff.
#1 - Some points seem similar to the old discussions about AI Dungeon prose quality and the new sampling tricks to force LLMs to stop repeating the same cliches. I wonder if there are more in-depth writings about both history and state of the art of AI writing, including the new experiments and tricks.
I'm curious about the karma thing. Why is it virtuous to persuade someone not to abandon their village during a famine?
Seems to me that if food is scarce and someone leaves, there's one less mouth to feed and more food to go around for the rest of the village. What is the moral calculus that deems this not a good thing?
Just a guess but it might be something to do with family obligations especially to ancestors. I believe making offerings to ancestors graves was considered important at the time and this might not be possible if you leave.
#19 - Mifepristone is the "abortion pill", not the "morning-after pill". The morning-after pill (Plan B) is a large dose of (iirc) estradiol - basically a large dose of regular birth control.
Morning after pill is levonorogestrel, a progesterone receptor agonist. Birth control is reliant on progesterone to limit ovulation, although most formulations contain some estrogen.
> I’d always heard modernism was a reaction to the trauma of the World Wars, but it looks like the rhyme decline started well before then.
It's also well to keep track of the language; I think it's not a universal phenomenon, but a regional one. At least, intuitively the more modern Russian-language poetry is much more rhymey than its English counterpart. (If you suffer from a need of rhyming poetry, go ahead and learn Russian and I'll share some of my favorites.) I've been re-reading Sorokin's novel "the norm" (1980-ies) recently, which is all postmodernist / countercultural / in-your-face-reader, and the verses there rhyme — that's the one tradition Sorokin doesn't break. (Although of course Russian vers-libre also exists.)
(Also, how would war trauma induce rhymeless verse? Did the wars include rap battles?)
I think Brodsky, who was both a Russian and an English (well, American) poet, had quite a few things to say about that - in particular, not just that Russian was linguistically more suited to thymes, but also that Russian poetry was much younger and still hadn't grown out of its formal stage.
I'm not even sure what it means for poetry to have a formal stage. A stage where significance is placed upon the form of a poem? As opposed to what, not noticing form, not having form? The words in which a poem is written are still part of it, it doesn't matter what conventions it does or doesn't follow — you still can't get away from the fact that the words are light or heavy, or curt, or clumsy, or whatever. If you happen to speak Russian (which I though you might?), this is one of my favorite Russian free verses (by Ivan Akhmetyev):
русская поэзия катилась
на ритмических салазках
пока не выехала на советский асфальт
веррлиббрр
No rhyme, no regular rhythm, and it wouldn't work without the form. When Kharms and Vvedensky were dismantling all previous literary conventions of their literature, were they writing rhyming verses because they hadn't outgrown a stage? What does it even mean? If they had a working tool that contributed to their unique styles, why were they supposed to outgrow it?
***
Actually, now I'm a little worried and confused: when a connoisseur of English poetry reads, I don't know, Walt Whitman, do they just not notice the texture of the material that their hand touches because it's just form?
I don't think it's a technical term, but I meant with it poems with a specific rhyme scheme and rhythm, yes. I assume most of the rhymed poems in 1900 in that Economist graph also had a fixed rhythm.
Honestly, I'm not a connoisseur of poetry, English or otherwise, and am not able to answer any of your questions (or even understand some of them). Except that I think it's generally accepted that in most of art, there are periods when certain things are done, and then periods when they are cast out. Scott has written a lot on architecture, for example, and you really couldn't write a novel with Victorian-style chapter summaries nowadays without it coming off as a parody. Of course most manifestations aren't as clear-cut, but I think it's pretty clear that the corresponding graph for Russian rhymed poetry would be shifted right by at least 50 years, probably more.
I see. My point is that the history of Russian poetry includes some events of throwing out all previous conventions and "things were done this way" (the Oberiu group is a salient example), but the fact of rhyming remains preserved because it still works as a tool in terms of the artistic impression it makes on the reader, even when other conventions are broken, so it can't be thought of in terms of "someone just reusing the same tools because they didn't think to try anything else". Breaking with tradition is something of a common theme in the history of Russian literature.
I don't disagree with that - I just vaguely remembered Brodsky saying something on the lines that Russian poetry hasn't been using rhymes as long as English poetry, and so they still worked just fine, which I think is what you say as well.
Sorry if I did a bad job putting that into words...
In any case, modernism in all the arts clearly starts before the world wars. In music, there’s clear breakdown of tonality and meter in Debussy at the turn of the century, and even some Wagner and Liszt in the 1880s, and you have very obviously modernist music like Pierrot Lunaire and The Rite of Spring just before the First World War. In visual art, you’ve again got the post-impressionists at the turn of the century, the impressionists a couple decades earlier (who are clearly modernist, even if they no longer trigger anti-modernist viewers) and before the First World War you’ve got works like Les Demoiselles d’Avignon and Kandinsky’s Improvisations. Literature may have been slower to turn to clear modernism, but people like Dostoevsky, Kafka, and Jules Verne all have elements that seem modernist well before the world wars.
#31 Ross Douthat interviews J D Vance. - Well, I guess I guess I can quit wondering if I’m being locally transgressive by linking to a piece in that particular source.
It seems they didn't "fire someone" exactly, they terminated their funding for that person's organization (the Mill Institute). Still pretty clear that UATX's dedication to free speech means "free speech for people who agree with me." i.e. they are just picking the opposite pole in the culture war from where most of academia sits, not actually trying to create an open ideological space.
Strange that folks can believe that father's could be biased toward sons but not that mothers might be biased toward daughters. That would also explain the difference: mothers preferring to spend time at home with daughters more than with sons.
54 - The FDA regulation in question just lets you avoid FDA regulation. You would still be violating the US patents, and a court can (and would) order the importation to cease.
>It would be convenient if there were seem deep synergy between the Jewish religion and psychedelia. But there isn’t. So they can either drop the issue, or else confabulate something.
I am kinda surprised. To me Judaism would be the second least psychedelic religion after Islam. First because it revolves heavily around "lawyering". Second because it focuses on text, not visual imagination.
Of the religions popular around here, I would nominate Orthodoxy as the most psychedelic - the concept of theoria.
A large mainline branch of Orthodox Judaism, Hasidism, is heavily grounded in mystical works attempting to describe the relationship between Man and G-d and the nature of reality, a lot of it using geometric symbology. Fertile psychedelic ground, and there is in fact a growing movement on the fringes to incorporate psychedelics into religious practice.
VR: I thought nobody is using this anymore, and then I found out it still has a fringe following among people who like to see themselves in non-human bodies. Which means... Zuck spent $50Bn to build... Furry Heaven.
Beyond the humorous aspects, can someone tell me how are people like this rich? Muh super efficient meritocratic markets?
> Beyond the humorous aspects, can someone tell me how are people like this rich? Muh super efficient meritocratic markets?
He's spending Meta's money... and I think this is a pretty reasonable choice for humanity, if not necessarily for Meta. They've pushed the frontier on VR/AR tech quite a bit, both in fidelity and cost-efficiency. I don't know how that pays off for humanity in the future, but it seems like a pretty good bet that it will, eventually.
I know multiple people that use a Quest instead of a monitor when working because they like being able to have a variable number of "monitors". Also, a lot of kids have a quest and play the games, so it's something that I think will gain user share as people that grew up with it age.
#43 I have always thought militant anti-colonialism was odd and not in tune with the public in most places. The implication of anti-colonialism is that pre-colonial feudal kings who might still be foreign were better than European liberals who built roads, trains and had some form of rule of law.
Note that the question seems to be asking about the modern influence of the former colonizer, not about whether the colonization itself was good or bad. You can believe that the 1930s Germany that invaded your country was literally Hitler while still believing that modern Germany has a positive influence on your country.
The two countries with the highest positive view of their former colonizer, Cabo Verde and São Tomé and Príncipe, have a very large migrant community living in the territory of that former colonizer, Portugal.
A quick back of the envelope calculation tells me that the equivalent of 20% of São Tomé and Príncipe's population (40k compared to 200k) lives in Portugal, and the figure is 17% for Cabo Verde.
I suspect that for countries with a large community living in the former colonizer, and where most people would know at least someone living there, the answer might have more to do with the general image they have of the country (Portugal) than with the actual political and economic influence of the former colonizer.
Cabo Verde was uninhabited before the Portuguese showed up. No Africans at all are indigenous to it. The island was colonized, but no group of people were "colonized" in the political sense of the word. There's only African-descended people there at all for the same reason they're in Alabama. That's a pretty unique case, it appears to have been swept up by the marxist pan-African fantasies of some guys in the 60s but it was never really of a kind with the other nations in that poll.
The world had a log-normal distribution of income in 1820, but had formed a pretty dramatic bimodal distribution by 1975 between colonizer and colonized nations, and then since 1975 has returned to a log-normal distribution: https://ourworldindata.org/the-history-of-global-economic-inequality. This strongly suggests that colonialism did hold back the Global South on average. The absence of colonialism doesn’t necessarily mean those countries would have had feudal kings forever; they could have modernized better through trade with the developed countries, rather than being colonized, like Meiji Japan did.
I disagree, Finland, S Korea, Taiwan, Norway, Poland, Canada etc were colonies, the richest states in Europe like Switzerland, Sweden, Luxembourg had no colonies.
There is no correlation between colonialness and current GDP. We can easily test the hypothesis with neighbouring countries with different histories.
#2 I love the fact that 664 Texans apparently weren't satisfied with the choice on the ballot presented to them (Allan Shivers the Democrat nominee vs. Allan Shivers the Republican nominee) and went for a write-in candidate, Allan Shivers.
#18: I don't really understand what you got out of the Dylan Black articles. Your argument about Tegmark's Mathematical Universe was the correct response to the fine-tuning argument. What Dylan Black says seems between straight-up wrong and missing the point.
He's using a very specific formalization of probability theory from Andrey Kolmogorov from 1933. I think there has in fact been some progress in the field of probability theory since then. For example, it's totally possible to have a uniform prior over the natural numbers (or even the real line): it's called an improper prior [0] and can be rigorously defined. The Beta(0,0) distribution is another example of an improper prior which is often used.
Or take this example in his article:
"Problem Statement: You're presented with an urn. I tell you nothing about this urn. You reach in, and draw one red ball. What is the probability that it contains 50% red balls? If you feel uncomfortable, good. You should. This isn't a hard problem or an imprecise problem. It's not a problem at all. It's a grammatically correct but meaningless arrangement of mathematical symbols, it is ill-posed."
I mean sure, the problem is underspecified if you want to calculate an objective number, but one of the central points of books like E.T. Jaynes' *Probability Theory* (2003) is that it all depends on what prior information you have. A human does have prior information about what kind of things are stored in urns in mathematics departments. And so, gun-to-their-head, they could give some probability for the above question. A superintelligence could use some approximation of Solomonoff induction to give an even better probability in answer to this question.
I didn’t read the Tegmark or Black, but the post you’re replying to doesn’t mention Kolmogorov complexity in any way. Kolmogorov’s foundational work on probability comes years before Turing’s idea of computation was described, let alone Kolmogorov and others using that to develop theories of complexity.
#56: that paper is ancient. But so's this other one I found on Gwern's server somewhere, one of my favorites next to that German one about fucking vacuum cleaners: https://gwern.net/doc/technology/1958-bain.pdf
This initially seems like an industrial design study, but soon gets into unintended psychology: some children, when trapped in a fridge, do nothing. Plus the summary in Interlingua is a cute historical touch.
They can just trade state-contingent claims. The neat thing is that since the state they are contingent on is a political decision the bearers of the claims can influence, the more claims are emitted, the more likely the state is to be fulfilled.
Even if some non-Trump Republic will succeed him, it's hard to imagine he'd resist the pressure forever
I write, and almost exclusively read, poetry that rhymes. I also spend a decent chunk of time looking for modern poets who do the same, people like Chelsea Rathburn, A. E. Stallings, and Faith Thompson. I know there's a famous tweet (once linked to on this newsletter) that says something like, "People want poetry that rhymes. But we won't give it to them." So why aren't the three people I just named more famous? They're brilliant. Maybe it's the lack of interest in poetry in general. It's certainly not a new state of affairs.
36. This are valid points but his waving away the concerns about diversion doesn't quite convince me
"Well, the Israeli government says it's because of large-scale aid diversion by Hamas. But they have never once provided evidence of that, publicly or privately"
(notice "large-scale" btw)
There is some evidence but I don't want to get into that discussion. From a purely Bayesian perspective, what are the odds that Hamas didn't exercise some kind of control over the old process? Does he seriously think that they didn't get food for their fighters?
I couldn’t believe when I got to that part. He doesn’t think the government of a polity which is currently being starved-out in a war for survival would commandeer resource shipments? If Hamas *wasn’t” using food aid to feed its troops, shouldn’t every Hamas soldier have starved by now? It’s absurd.
If hypothetically Hamas wasn't stealing food, couldn't Hamas members just show up at the aid distribution site dressed as regular people and get food like any other person?
1. One has to show that there is still a significant number of Hamas "soldiers" around after 2 years of war
2. Even then, they are not really soldiers, they are irregulars. They can very well just tell their people to get food on their own, much like any other guerriglia force did.
3. At this point (actually, it's been more than one year), Hamas cannot keep criminals in prison, cannot prevent them from raiding armories, and formely incarcerated gangsters such as Yasser Abu Shabab are openly claiming territory for their gangs and looting aid on their own. It is not hard to believe that Hamas might be unable to seize aid simply because other local actors have disloged them
I'll add that GHF workers have been killed or tortured by Hamas (https://www.fdd.org/analysis/2025/06/29/hamas-places-bounty-on-aid-workers-ghf/), which (even aside from any Israeli evidence) implies pretty strongly that Hamas has reasons to not want aid distributed through them. I can't think of an explanation for this that isn't that it's because they control the alternate means of aid distribution.
It's also worth noting here that UN (and affiliated aid orgs) have been suspiciously quiet about this (see e.g. https://www.jns.org/ghf-lack-of-un-condemnation-of-hamas-killing-of-aid-workers-shows-bodys-underbelly/ ). There's a range of how we can interpret this - from somewhat benign ("they don't want to piss off Hamas in order to keep their neutrality) to amorally greedy (they don't want anyone else cashing in on their aid bucks project) to outright hostile (they hate Israel and don't want anything that can help Israel reduce Hamas's power).
But whichever interpretation you go with, the underlying fact that UN aid orgs are incredibly politically hostile to GHF stands, and UN aid insider criticism of them should be seen in that light.
I could think of any number of explanations - they see GHF as enemy combatants (as they are run and mostly manned by the US government which is also supplying Israel with weapons) while the UN not, or they (like many humanitarian orgs) see the GHF as an attempt at ethnic cleansing, or they fear the GHF will shift the populace away from Hamas strongholds and deprive them of human shields...
It just passed the Senate 51–50, with Vance breaking the tie. I expect it'll start getting coverage now that it has to pass the House again, since many prominent people highlighted this is reason they'd vote it down.
#3 - I think this is a good test between people who truly understand physics, biology, chemistry, etc, and people who simply memorize facts.
If a chess master is burning as much chemical heat energy as a professional athlete, then where does all the energy go? Dry skin in a room temperature environment is unable to radiate heat at a 6000 calorie/day burn rate. Chess players would have to sweat like basketball players if the claim were true.
I don't think this is necessarily true, because there's a lot of variability in how concentrated heat-generating activity can be to achieve a rate of 6000 calories/day. Basal metabolic rate is roughly 100 calories per hour. Working out in a cool room, I sweat heavily at caloric expenditures of 800 calories per hour, but not at 400 per hour. A chess tournament can last much longer than a basketball game, and with the activity distributed enough, the participants would not necessarily become particularly sweaty. Also, my understanding is that some chess players do become quite sweaty at tournaments, if not as sweaty as basketball players. Personally, I tend to sweat when I concentrate intensely, so I'd be surprised if that weren't the case for at least some chess players.
The issue I think is harder to square is that chess players don't seem to *breathe* as heavily as is associated with even modest cardio exercise. They might sweat heavily due to stress, but if they're engaging in increased respiration, they'd have to take in more oxygen.
"The issue I think is harder to square is that chess players don't seem to *breathe* as heavily as is associated with even modest cardio exercise. They might sweat heavily due to stress, but if they're engaging in increased respiration, they'd have to take in more oxygen."
Did you actually check out the link? The whole thing started with someone observing rapid breathing from chess players in tournaments (for short periods of time), and someone else extrapolating (poorly) on how many calories they could burn with that much extra oxygen.
To be slightly more precise they report the chest movement rate (from impedance pneumography) and interpret it as breathing rate- since your chest can move for reasons other than breathing and they do not specify a method for filtering the data it is not clear that the maximum signal they reported should even be considered a breathing rate in the first place
Exactly. Weirdly on-point anecdote but my ex had taken Sapolsky's class at Stanford and casually repeated this claim to me many years ago. I have a physics degree and my immediate reaction was "no f-ing way, show me the evidence" and made exactly that waste-heat argument.
Ha! Thank you for this comment. I didn't think to explicitly search through Sapolsky's lectures, but sure enough, it's in lecture 1 of the recorded version of his Introduction to Human Behavioral Biology (albeit just "thousands", not 6000). Note that this video has over 18 million views.
> You have two humans, two individuals, who are going through a ritual. They are sitting at a table. They're absolutely silent. They're making no eye contact, and they do nothing more physically taxing than every now and then, one of them picks up their hand and moves a little piece of wood on the table. And if these happen to be the right two individuals in the middle of a chess grandmaster tournament, these people are maintaining blood pressure for six hours running that you only see in a marathon runner. **These people are going through thousands of calories a day doing nothing more than thinking. And this is outrageous.** Because you look at one of these chess grandmasters who's just taken down an opponent, took down their queen or whatever, and they will have the exact same physiology as some male baboon on the Savannah who's just ripped the stomach open of his worst rival. And we're doing it there just with thought.
#13 Note that the huge income tax "cuts" in the BBB is mainly just maintaining the current tax rates! No one actually gets lower taxes than they currently pay.
This counts as a tax cut because the original bill only lowered taxes for the next 10 years, when rates would go back up.
This is done to make the official forecasts look good, but the intention was to keep the new rates in the future, which is now.
I'm not saying this is good (or bad) policy, but it's important to understand that taxes largely stay the same, even though it is in a technical sense a tax cut.
Thanks for posting this. I came to learn only this week that the 'tax cuts' which people have spilled a lot of ink over are 'the same tax policy we've had for the majority of the last decade'. Maybe I'm dumb but this feels like an underdiscussed point.
It's still important though because, when the original tax cuts were passed, the entire deficit/impact forecast and subsequent conversation at _that_ point was under the assumption that would expire the way that they were meant to. This is the bait and switch that is continually getting played (by both sides). Get what you want now, minimize the projected impact by giving it an expiration date in order to reduce public backlash and criticism, and then, when it comes time to actually let it expire, convince everyone that "oh this is just the status quo, it's not that big a deal to let it continue and cancel the expiration". So now, we are at the point when the lie of the expiration is finally being made explicit. So, relative to what was promised when they got passed, this _is_ additional tax cuts.
If you want to argue from the perspective that this is nothing new and just status quo, then I hope that the next time someone proposes something with such an expiration, you demand a forecast assuming it won't expire, and take your view based on that forecast.
I am, if anything, inclined to penalize this even more than usual specifically to try and discourage this bait and switch tactic.
I don't disagree that the TCJA's original proponents made their arguments in bad faith, what has been weird for me is the framing. I feel like both sides are selling this as a major tax cut, with the republicans hoping that people will like that and the democrats hoping that people won't like that because it includes/favors rich people. But, unless you are on medicare or frequent a medical provider with a high percentage of medicare beneficiaries, nothing about your life is likely to change if this bill passes. Like, I don't understand how the bill's authors expect it to win them in any votes in the midterm, given that it makes no-one's life better and some people's worse than the status quo. Democrats can and are pointing to people losing out from benefits being cut, but they will not be able to point to a rich CEO making more money from the bill, because the CEOs will be paying the same as they were in 2019.
If the republicans were not so committed to donor enrichment they could have just let the 199A deduction expire, not touched benefits, and said "SEE DEMOCRATS WANT TO RAISE YOUR TAXES" and the democrats would have had to get an argument with the ref, both in the style you've gestured at. But no one has done either of those things, so the rhetoric hasn't matched the situation and I don't see how it will pay off for them a year from now.
10: The explanation for this that I've always heard is that sometime during Trump's first term, social media algorithms changed to start pushing the manosphere much harder to teenage boys, realizing it was a good way to get them really hooked into the social media systems. I haven't done any research into whether or not this is true, that's just what I've heard.
33: "I guess this is good (the verdict wasn’t dependent on a few ignorant or dishonest people), but maybe also bad (shouldn’t being familiar with the best evidence for one side or the other make you believe that side more?)" Isn't this what you should expect from a healthy scientific field? The experts in a subject matter who don't know the details of a specific topic still know whose opinion they should trust, and thus end up converging on the truth? This sort of shortcut is exactly how science is supposed to work, right?
#20 Wonderful timing. The Roman Catholic lectionary cycles through much of the Old Testament every few years, and today's Old Testament reading features the destruction of Sodom and Gomorrah!
#1 - The way that the author identifies the relatively small set of emotive words the AIs consistently alight on (ghost, whisper, echo, buzz, hum, pulse, etc.) reminded me of a common observation about another poorly-trained artificially intelligent storytelling entity, the modern Walt Disney Corporation. It's an in-joke in theme park fandom that Disney only knows about a dozen words (dream, fantasy, wish, enchantment, adventure, magic, etc.) and combines them in various ways to serve every possible purpose.
Other corporate entities surely have their own favored vocabularies (synergy, sustainability, difference, etc.) just as AIs bent toward nonfiction do, but something about narrowing the domain to the production of fiction makes them even easier to surface.
It makes me think of the "start to Clench" races for Stephen Donaldson books, where everyone opens a different book to a random page and sees how long it takes until the word "clench" is used.
(In trying to find that article again, I've discovered Google lists no blogs or articles called 'Rickety Fiction', which seems like a fun enough name that SOMEONE would have used it.)
Re #3: Man, that's really disappointing about Sapolsky. I really respect him and though he was a responsible scientist, but this is pretty irresponsible. Now I will have to take every claim from him under suspicion. I wonder how many of the stories in A Primate's Memoir were embellished or even made up.
I just want to add that one thing I really like about Scott is that he is not a bullshiter. I really dislike bullshiters. I think Scott is about as honest as people come, and I really respect that.
Thanks. I think relatively few people are bullshitters, and it's more of a question of how careful your back-of-the-envelope calculations are. It looks like Sapolsky's issue was that someone said chessmasters breathed three times as fast (at one point), and he seems to have assumed breathing was a perfect correlate for metabolic rate so you could multiply daily metabolic expenditure by 3x. I can imagine making this mistake if I filed this fact away in my head as "something something 3x something metabolism chess" and then remembered it while writing an essay about something else and rolled a critical failure while trying to recall it.
I have probably done things approximately this stupid; when I don't, I'm protected less by careful source tracking, and more by a good enough mental model of how metabolism works that if I felt tempted to say something like this, I would think "no, that's crazy, I know that an active brain doesn't consume that many more calories than an inactive brain" and then check my source.
I don't think Sapolsky deserves too much disapproval for having gotten the calculation wrong. I think he deserves a lot of disapproval for, once it became clear that it was wrong, going around falsely telling people that it was a journalistic misrepresentation of what he said, instead of admitting that he made a mistake. I think that you would not have done that.
I think bullshitting is a spectrum, and I think almost everyone does it to a certain extent. I also think that there is a continuum between genuine mistakes and intentional bullshitting. The question then is, how likely are you to make a mistake about a fact or an inference from given facts, based on your commitment to truth and your love of bullshitting? In this case, and of course this is speculation, but as an intellectual, would it not be seductively attractive that thinking burns as many calories as athletic competition? Someone committed to truth would be wise to stop and take a step back, when such a seductive fact presents itself. But if you are prone to exaggeration, you will dismiss such thoughts.
Perhaps I am being too harsh, and I am sorry about that. But even in the most generous interpretation, this reveals a carelessness that I find off putting in a scientist. I still like Sapolsky and will continue to enjoy his books. I will just approach his writing with more skepticism, which I suppose is not a bad thing.
17) "I’d always heard modernism was a reaction to the trauma of the World Wars," No, it was a reaction to the glutinous over-reach of late 19C art. The more every new building is ornately Gothic, the more you're likely to dream of clean Bauhaus lines merely as a refresher. The problem is that then the reaction went even more overboard than what it was reacting to.
Useful to remember that Virginia Woolf wrote, "On or about December 1910, human nature changed."
56) I also hated dogs as a child, still do. And not for any psychological displacement reasons. Dogs attack strangers, they bark, they bite, they knock you over. Nothing used on the child in this story would have worked on me at all. Isn't there a reason for signs reading "Beware of dog"? You don't see any signs reading "Beware of cat" or "Beware of goldfish."
I'm not Big Swan, but I assume they meant to imply that the (radical?) left's support of Palestine is mostly about "decolonization", i.e, that they're fully on board with terrorism against Israel, support 10/7, etc. And that this has nothing to do with the recent occupation of Gaza by Israel.
ETA: to be clear, I'm merely relaying claims I've heard from rightists about the supposed true motives of leftists.
I do think it's a vibe based question (IMO most teenagers are going to answer surveys based on vibes and will probably have a higher lizardman constant), but there are definitely enough people to hit 20% on that question. The replies on that link are completely full of people arguing the premise - IE "if women did the same work they should get paid the same, but women don't do work as well as men" and things like that. The age they're surveying is particularly susceptible as well, I personally remember being pretty shitty at that age and growing out of it in college with exposure to more people and becoming friends with more women.
13. Trump's BBB "lowers taxes (especially on the rich)"
1. Most of the so called lowering consists of making permanent provisions that were enacted in 2017 and were scheduled to expire later this year. I have been a tax professional for 50 years, during that time there have been major revisions to the tax law in almost every administration. Any thing that is in effect for 8 years should be viewed as permanent. Taxes are imposed on a dynamic economy. We do not know what the effect of most changes are until years after they are put in place. A real reform would be leaving things the same for long stretches.
2. It is a tax on income. People who do not have much income do not pay much taxes. The standard deduction for a married couple filing jointly is not $29,200. Any decrease is mostly going to be a decrease for those who are better off.
3. The US has had an income tax system for more than a century now. it has not affected the distribution of wealth or income in any meaningful way. It probably cannot. The distribution is determined by factors that are not related to taxation. The wealthiest people in the country, Musk, Bezos, Gates, Ellison, Page, Brin, all made their money by exploiting a new technology. In the gilded age it was railroads, now its the internet. The real impact of taxes is imposed by the forces of supply and demand. Any tax that on a growing segment will simply be passed along as a consumer cost. Taxes should be designed to collect revenue cheaply and and easily. Getting the most feathers from the goose wit the least amount of hissing.
It's unfortunately paywalled, but here's my summary of his argument: It's a cohort effect. Most millennials came of age in the late 2000s or early 2010s, which was when the coolness of progressivism peaked. The conservative establishment had three huge high-profile political failures in short order (the Global War on Terror, the Great Recession, and same-sex marriage), which created a sense that the right was just bad and wrong about everything. The election of Barack Obama was the most exciting thing in ages and put tons of cultural momentum on the progressive side. And the cultural hegemony of scoldy Christian moralists was in just the right phase of its decline: recent enough that millennials remembered it as the uncool establishment to rebel against, but too advanced to push back effectively against that rebellion.
Now, as Yglesias's title says, those progressive millennials are old and uncool. The new cohort that's now coming of age is doing so in an environment where the left hasn't been doing so hot, and grew up being scolded by *woke* authority figures and so thinks of wokeness as the moralistic establishment that's cool to rebel against.
Also Democrats used to benefit from the decreasing share of the white vote, but that's stopped working as the right has gotten better at appealing to racial minorities. I'm not sure to what extent it makes sense to think of that as the same trend, though.
Your summary of the social media consensus is wrong (to be clear, that's because the reporting and tweeting on this by the authors was misleading). I'm not sure what the plot is from (not the preprint), but the study did a Delphi process and it looks like this is the stage 1, not the final vague statements that they had consensus on (see third link).
A representation of the final consensus, although the first author of the preprint (Capraro) says the best representation is the table in the preprint.
#10: Although the woke movement peaked in 2022, specific lines of woke argument peaked long before then. One of these is the alleged 23% gender-based pay gap. Although it was debunked by mainstream outlets in 2012, the myth persisted among feminists for *years*. I remember seeing new videos and articles decrying the pay gap in 2017, and perhaps as late as 2019.
For some data points about the history of this debate, see section 6 on this page:
My first hypothesis was that some boys took the "reversed stupidity" response to the pay gap claim. Feminists claimed "Women should not earn 23% less than men", and although many boys inverted the claim correctly ("Women *do not* earn 23% less than men"), some of them inverted it incorrectly ("Women *should* earn 23% less than men). This explanation seems plausible in principle, but I think the timing is wrong.
My next hypothesis is that it's related to the MeToo movement, which began in 2017. The movement may have contributed to the proliferation of workshops and other programs about sexual consent. (Presumably, these would have began 1-2 years later.) Many of the boys would have felt like they were being blamed, profiled, or punished for bad things that they didn't do.
In addition to stoking a general feeling of resentment in the boys, I suspect that some of the boys also responded with solidarity toward adults who actually did mistreat women. Both of these psychological factors could contribute toward the development of sexist attitudes.
My bet is on MeToo. It peaked in 2018 and I believe exhausted many men's tolerance for modern feminism. For me personally I know that's when I switched from considering it a little nutty to outright adversarial and in bad faith.
13. What I find interesting about this is that it was largely the religious right that drove this over concerns about pornography and child protection. It's also a good reminder that one reason prediction is hard is that success or failure can come from basically exogenous factors to the narrow area you're examining.
16. Decker is not as smart as he thinks he is. No bank would lend on the promise of a political party, in part because the Democrats can neither credibly promise to win the next election nor that they will pass that bill. You could calculate the risk discount rate but it would be huge. You'd need to discount that they'd win the presidency and both houses which, even if they're individually favored in all three, would net out to a lower than half chance. Then you'd need to discount they actually fulfilled the promise.
Giving some fairly favorable assumptions:
P(Presidency win) = 60%
P(House win) = 55%
P(Senate win) = 55%
P(Passing the bill even with control) = 70%
You get about 87% discount rate. That is insanely high. That's 4x the discount rate that VC gets for investing in companies that don't exist yet. That's higher than the discount rate given to Zimbabwe during its crisis.
That math isn't right because it assumes that the three electoral outcomes are uncorrelated, and because it represents the probability of *any given* election resulting in unified Democratic control, without accounting for their getting another chance every two years.
The math is simplified for sure. If you want to assume they're positively correlated and produce a P(Unified Gov) that does increase your chances but it still ends up at like 40%. Which produces a discount rate of like 73%. Which is still gigantic. And that's in a good year, if they have a less than 50% chance of winning it drops.
They get a new chance every four years, not two, because they need the presidency. But you're right that's the biggest simplification: you'd need to calculate the time value. You get an expected wait of 3.7 cycles through that, rounding to 16 years. Multiply them together and you get 89% even with the higher correlated number.
I'm not sure I'm following your discount-rate math. If (we assume for the sake of argument) the banks will have to wait 16 years in expectation to get paid back, then can't you just apply the normal interest rate to figure out how much they have to charge? The concern would be the possibility that they never get paid back at all, and I'm not sure how your back-of-the-envelope calculation purports to calculate that.
No, because the bank is not trying to give the universities the lowest possible interest rate. It needs to take into account default risk, political risk, time horizon uncertainty, etc. This means the cost of capital converges to the risk adjusted rate, not the prime interest rate.
What I think you're kind of feeling is the universities have other, cheaper sources of capital already. Which is true.
Somewhat that, but also I'm literally just trying to understand your math; I don't work with financial math that frequently. When you say "89% discount rate", does that mean that a lender would have to charge 669% APR to break even on expectation? If not, what does it mean? And can you show how that number is derived from the probability of unified government per election?
36. While I'm sympathetic to the idea that experts tend to know things, I'm also sympathetic to the idea that experts have a vested interest in finding whatever replaced them failed. There's no mention of the reason the GHF concentrated in a few checkpoints: because they were under military guard after Hamas supporters attacked those more distributed checkpoints.
He also does not mention the reason the UNRWA and other aid organizations were able to disperse aid checkpoints is because they paid off Hamas. Now, maybe he just doesn't think this is worth mentioning, but I think he knows the optics of "we paid Hamas so they would allow us to feed starving Gazans" would harm already waning support for UNRWA. And so on and so forth. This is a narrow selection of facts that veer into speculation, memetically fit but not a mature discussion of the situation.
48. The Gulf War was fought over land though which meant that the US could sweep in hammer after the degredation including actual, hard fighting. China's hammer would need to go over a rough sea. There's also the problem that the Gulf War, which is widely regarded as rapid, took six months. That's plenty of time for even a relatively unprepared America, Japan, etc to get itself together and intervene. And once they decide to intervene China has no option but a slugging match with no clear diplomatic exist. Of course, "a slugging match with no clear diplomatic exit" is also how I would have described Ukraine.
Another important fact is that neither Ukraine nor Iraq could retaliate against the home territory of the attacking country. Taiwan absolutely can.
Sure, maybe China can hope for the victory to be so quick, and the Taiwanese military so rational and dispassionate, that they just negotiate a dignified surrender rather than ravaging Xiamen in defiance. But it's an additional risk they're taking compared to Russia and the US.
55. China has had various institutes for understanding and controlling the effects of AI. It was seen as a natural extension of their control of other forms of technology and information production. I would bet the point of making an AISI that's modeled on the western ones is not to signal a new commitment to AI safety but to provide a Chinese voice in such conversations abroad. They do this a lot, having an internal version that controls domestic affairs and an external version that's meant to push the Chinese point of view in international organizations or create commitments China itself is not bound by.
A key tell: can you get commitments from actual Chinese regulators and are they sending over top party officials? Do they actually make laws and regulations that are enforced internally? Are they allowed to advocate for it openly inside China in Mandarin? Or are they just sending over people who are mostly specialized in dealing with foreigners with little domestic power? So far the CnAISDA style outreach has been the latter.
That said, I stand by my comment China invests more into AI alignment research than anyone else in the world. They just want to align it with the CCP.
Compare the image manipulation retraction of the Sodom & Gomorrah paper to the image manipulation retraction of the Alzheimer's amyloid beta paper. The archaeology paper was published in 2021 and immediately had concerns about repetitive elements. The Alzheimer's paper also had concerns in 2021, but was published in 2006. And my understanding is that no one looked at it merely because it was an influential paper, widely known to be unreplicable garbage, maybe widely believed to be fraud, but only because they were doing general sweeps, having discovered > 1% image fraud rate.
Why didn't the Alzheimer's paper get any attention to its images? Because no one thought it would be easy to prove fraud? Or because no one thought proving fraud was worth the personal career risk of being a whistle-blower? If image manipulation fraud is so common, why aren't there any whistle-blowers testifying to having witnessed it? Do the people doing systematic studies of fraud have a secret origin story of witnessing fraud but don't want to throw their mentors under the bus?
22. So it’s sorta like Cameo, but for dorks like us?
23. I’m surprised it’s only 32 players. I thought the general practice was to take any player of prominence and slap his name on a random athletic shoe design.
#1 - Is the quick and easy answer for why AI writes like that simply because the training data for fiction writing is overwhelmingly things like tumblr posts, mediocre mass-market stuff, and, perhaps most significantly, fan fiction? When I was reading the AI output it reminded me most of all of cringe-inducing tumblr posts and that time in high school (I've fruitlessly tried to erase from memory) when I went to see what all the fanfic fuss was about.
Aren't there lots of traditionally published books and journalism and stuff in the training corpus? Also, I don't see why fanfic in particular would be more than a small percentage of it, most user-generated text on the public internet isn't fanfic.
I'm not familiar enough with what's in the AI corpus, nor do I understand how much training on non-fiction (most of the internet, I assume) informs an AI's fiction stylings. But a very quick search indicates that fanfic, specifically, numbers in the lower 8-digits, and the number of traditionally-published fiction books is in the high 5-digits yearly. How many of those are available for AI training? I have no clue.
Self-publishing (mostly e-books), which probably has some overlap with fanfics, as I assume people often publish to both fanfic sites and create e-books of their works, numbers roughly 3-4x traditional publishing. I'd hazard to guess most of that, in terms of quality, is closer to the fanfic average than the traditional average. But, of course, the traditional publishing average is probably closer to the fanfic average than it is to "great" writing.
But it seems obvious to me that it will be extremely hard or impossible to get AI to produce great fiction, given that it's something of an aggregation machine. Nearly all humans can't write great fiction either!
That said, the particular stylistic quirks shown in the article resemble a particular style of internet writing, at least to my eyes. I'm not confident in this argument at all, as you can probably tell.
I'm not terribly familiar with Cremieux's take on the lead/crime hypothesis -- I think he has mostly argued that cohort studies are heavily confounded by race and class, so you can't just compare violence rates in high blood lead neighborhoods against low blood lead neighborhoods.
Meta-analysis of such studies and their confounders is difficult, but there are many other reasons why the lead/crime hypothesis fails to be some kind of unified theory for the rise and fall of crime across the western world in the second half of the 20th century:
In short, the lead/crime hypothesis fails for many other countries: Japan, Hong Kong, Mexico, Brazil, etc. The lead/crime hypothesis also fails in the US, if you divide the data up by age groups -- the rise in violence in the 1960's was a period effect, not a cohort effect, as you'd expect if this were an issue of subsequent generations getting lead poisoned in youth.
The rough synchronization of crime trends across western countries is fascinating, but may ultimately have simpler explanations. Demographics is one of them -- these countries all had a post war baby boom and then a decline as birth control became more widely available in the 60's/70's. A proper analysis would, at the very least, age adjust the data before just noticing the bump in the murders per capita graph.
There does remain a plausible theory that blood lead levels are one small factor in the bigger picture of crime rates, and that leaded gasoline did raise crime somewhat (10-20%?), all else equal. A sufficiently careful analysis of cohort studies and their confounders might be able to find whether this effect exists, and how large it is. But that's a hard thing to tease apart from the data and separate from other variables that did have an effect on the 80's crime wave and subsequent decline. Other strong factors may include: demographic changes, crack cocaine, mass incarceration, changes in policing, and medical changes that reduced the lethality of gun shots:
"Related: Cube_Flipper on his (?) experience taking estrogen. “What did change was my sense of space. This one’s quite subtle – it was the kind of thing that was more noticeable when I experimented with deliberately spiking my hormones. I’ll do my best to explain. It’s as if I took the entire volumetric representation of the space around me and increased the degree to which every point within that could influence the location of every other point, recursively. This allows everything to elastically settle into a more harmonious equilibrium. This effect is basically identical to what a small dose of psychedelics can do, specifically a tryptamine like psilocybin or DMT.”"
Might be an absolutely wild thing to say but I've been thinking this for a while, a big difference between women and men is that women are basically constantly microdosing on LSD/psychedlics. Thats there experience of the world
17. My pet theory is that the "decline of rhyming poetry" is largely an illusion caused by the rise of sound recording and broadcast media. There is still a ton of rhyming poetry being written, and much of it is extremely popular and commercially successful, but we call it "song lyrics" rather than "poetry" because it comes to us as audio/video recordings of the poem being performed set to music. Even for genres where the lyrics are often spoken rhythmically than sung (certain subgenres of folk, blues, and rap), we still think of it as music rather than poetry.
Before phonographs and radio became widespread, lyrics were commonly composed and published without any particular music being intended, and people could read them as poems or set them to whatever tune they liked that fit the meter of the poem. And poetic forms were standardized enough that you had a large menu of tunes to choose from for any given poem. Especially for "common meter" or "ballad meter", which is both very specific and very widespread, and is why you can sing the lyrics of the Gilligan's Island theme song to the tune of Amazing Grace and vice versa. One major and very popular form of poetry/music in the 19th century was "broadsheet ballads", poems written in ballad meter and published in broadsheet format (basically a single sheet of a full-sized newspaper).
"Chess Grandmasters Do Not Burn 6000 Calories Per Day"
There is a reddit post with the exact same title from a deleted account, is it your post? Also, more generally how good is substack for search engine optimization. I dont think its very good, I cant find your post on google, limiting the value of correcting this misninfo. I've had this problem with other substacks as well
I think it depends on how you define poetry. If you limit yourself to the academic journals, maybe yes. But, I think that is snobbery. I think the great mass of poetry these days comes from in the form of lyrics and rap, both of which are deemed to be musical genres.
Didn't the august body that awards the Nobel Prize for literature concede my point by giving the laurel to Bob Dylan? He characterized himself as a poet when young:
Though you might hear laughing, spinning,
swinging madly across the sun.
It's not aimed at anyone.
it's just escaping on the run.
And before the sky there are no fences facing
And if you hear vague traces of skipping reels of rhyme
To your tambourine in time
It's just a ragged clown behind
I wouldn't pay it any mind
It's just a shadow you're seeing,
that he's chasing
Rap of course rhymes. I detest it because of its coarseness and vulgarity, but it does often rhyme.
The thing about musical lyrics is that they have an unfair advantage. There's something about hearing words set to a tune you like that changes how you process them. For instance in song, phrases or sentences are often repeated, sometimes over and over with no other phrases intervening, and not only is that not tedious, it is often very powerful. For instance I love hearing the Talking Heads' sing "take me to the river, drop me in the water" over and over. Or there's the end of a hallelujah chorus -- "forever and ever" & "hallelujuh" over and over. But it's not just repetition we experience differently in song. Cliches are much more tolerable -- can sound heartfelt and obviously true -- .and so are phrases that don't make sense. I've often looked up the lyrics of songs I love and felt taken aback by how lame they are with the tune stripped away.
Allan Shivers did not "beat himself" in the election. His vote share from each party line was pooled together to make his total vote share. It wasn't like there was a competition over whether he gets elected as a Democrat or as a Republican. It's also not unprecedented for candidates to be nominated by both major parties. e.g. Earl Warren won both major party's nominations for CA Governor in 1946. Robert Heinlein wrote a book about his experiences in politics, in which he complained about how he had very narrowly lost an election in which everyone told him they wouldn't bother helping him because he had no chance; I fact-checked this and discovered that he had in fact narrowly lost the Democratic primary to the Republican incumbant. In New York, candidates still can get listed on multiple parties' ballot lines, and it is common for the Democratic and Working Families parties to nominate the same candidates. It's no longer common for both major parties to endorse the same candidate, but Simcha Felder did get both major parties' endorsements for his NY State Senate run in 2024.
Until the Top 2 Primary ballot measure took effect (in 2012, IIRC), California still had a weak back-door ballot fusion system. You could only be a listed candidate on the primary ballot for one party, but you could register as a write-in candidate for other party primaries for the same office. I don't remember if you got separate ballot lines if you won more than one nomination, like in NYC, or if you had to choose one (so the benefit would just be less competition on the general election ballot), or if you got one line with both parties listed on it.
I know about this because in 2010, I was peripherally involved in a Congressional campaign in a safely Democratic district where there was a relatively strong Libertarian candidate and no qualified candidates for the Republican nomination. The Libertarian, Edward Gonzales, tried to use the process to also get the Republican nomination. It didn't work: the local Republican party organization recruited and endorsed a different write-in candidate who won the primary by a large margin, and the Democratic incumbent (Zoe Lofgren) was reelected by an enormous margin (69% to 24% to 7.8%) in the general election.
#10 (fewer boys believe in gender equality): the explanation seems clear to me, 2018 was the year #MeToo peaked. Weinstein got arrested, the Golden Globes was a public MeToo rally with celebrities wearing black, the phrase "believe women" gained currency, and the Kavanaugh hearings happened. It pushed a certain subset of men to conclude that the 'equality' movement was just rhetorical cover for a zero-sum political gender war. So "gender equality" got (appropriately, in my view) re-coded as "anti-man" by many men and so they stopped supporting it.
46. Very interesting. This offers a lesson about prior restraint reverberating through the ages. José Clemente Orozco famously self-sabotaged in the same fashion when painting his tremendous mural "Prometheus" (aka "the hunk without the junk"), located in Pomona College's Frary dining hall.
When I was a student, I heard that Prometheus' penis had been removed or obscured by a third party. It was understood to have been due to puritanical inclinations, but one can't dismiss considerations of appetites might have played a role.
At any rate, I haven't thought of Prometheus in a while, so I looked it up and was surprised to learn that Orozco had actually omitted the titanic genitals himself!
However, seemingly filled with regret, "He attempted to add one when he visited Pomona several months after initially completing the mural, but it did not adhere properly to the wall."
#43: So portuguese have been pretty cool about their colonialism? Good catholics who know about humility. At least, by a glance, germans seem to score better than english and french. I wonder what african view on china may look like a century from now.
#35 - I'm impressed by the problem solving strategy the LLM is using. It seems like it's either a novel strategy that others aren't using, or it's using one of the most advanced existing strategies. Are there other domains where LLMs are coming up with novel ways to go about solving a problem? Right now it seems like LLMs serve as quickly-responding encyclopedias of known problem-solving strategies. Are there domains where LLMs are producing novel problem-solving strategies without expert prompt engineering? If/when this starts happening, I wonder which domains it will happen in first and how it will spread across domains/topics.
48: "T. Greer on China’s strategy for invading Taiwan."
The Chinese strategy described sounds just like the Russian strategy for invading Ukraine. The Ukrainians are a lot more like Taiwan than the Houthis, who are tribesmen wandering around a desert.
But the all time champions of fighting off invading armies of larger countries are the Israelis. And that is who the Taiwanese should model themselves after, if they are serious about staving off the CCR regime. Of course, it would need to triple its spending, start producing its own armaments, and institute mandatory universal military training and reserve service.
But, if they do something like that, their prospects aren't that bad. The Island is 380 km long and only 130 km wide, but, the eastern two thirds of the country is quite mountainous, The mountains can be a shelter from attack and a defensive redoubt. Given time and money the Taiwanese can defend themselves. they have the money. The questions are whether they have the time and the will.
I object that Bentham got "Eulered" there. If you're going to try to make an argument that is so heavily mathematical in nature, it is entirely correct to point out any issues with the math. Math provides a high level of certainty (or at least, the impression such) *due to the unrelenting demand of rigorous argument.* The issues that Dylan Black raises are pretty core to the use of probability on infinite sets. If you don't understand them, you probably shouldn't be making arguments (especially in favor of something as complex as God!) that rely on the mathematics of probability theory.
That seems like a pretty big overreaction. I updated a small amount away from Cremieux, but:
- That chart wasn't the only bit of evidence Cremieux cited.
- I don't even know what Higney could have meant when he wrote "For some reason [Cremiuex] cut off the bottom of that chart" - Higney posted a different chart! (https://x.com/AnthonyHigney/status/1867163809267507628) If there's room for legitimate disagreement about the relative value of PCCs and elasticity, the entire basis for Lexer's (and Higney's) criticism vanishes in a puff of smoke. As an outsider, I asked some LLMs about it and Claude said "If lead exposure leads to small PCCs but high elasticities, it suggests the relationship might be nonlinear, confounded, or noisy, leading to weak correlations." If that's correct, it's not necessarily diabolical to downplay elasticities, particularly if you think they are due to confounders.
- The interpersonal drama Lexer brought up seems completely irrelevant to me. But at best it's a draw, in my book. Higney called Cremieux "a bad faith actor", which is a rather damning thing to say about someone, imo, so even though I don't fully understand Cremieux responding by calling Higney "surly, reactive, and untrustworthy", it seems in-bounds given the level of interaction they were having.
#10 is unsurprising; male socialization for teenage boys is even more fucked up than it usually is on the interpersonal level (source: anecdotes and vibes) and there was a big organic push from anti-woke influencers over the past X years which the party and the money put their shoulders behind, from hoi-polloi types like Rogan and Walsh to intelegencia blogs like this one.
This seems like the result one would expect if they knew 12 years ago that the anti-woke side of the culture war would be making gains.
#8 I prototyped a video game with some friends when I was in college where you would earn karma by performing good deeds and lose karma by performing evil deeds, and whenever your character dies she'd reincarnate as a different animal, with different abilities, according to the current karma level. The backstory was that the protagonist was a member of an order of monks grinding good deeds to attain enough karma to ascend to a mythical creature after reincarnation, but was tricked into deicide, the most heinous crime, by a demon king, and lost enough karma to be on the threshold between humans and monkeys. But since she got the power to immediately reincarnate and retain her memories she's the only one who can use the powers of all the creatures to resurrect the godess she had killed and get the karma back to ascend to the higher entity and defeat the demon king.
The gameplay loop was full of contrived trolley-problem-style situations where you'd gain or lose karma and then kill yourself (or allow yourself to get killed) to reincarnate as whatever creature the platforming situation called for.
If I had known at the time about this Ming dynasty guide to karma bounties, I would have used it verbatim, this is hilarious.
> 30: Twitter and Bluesky both seem to be losing users...
On X, I've noticed that lately I'm being swamped with posts from obnoxious rightwing crazies. I block and block, but the X algorithm keeps throwing them at me.
OTOH, everyone on Bluesky is just too fecking nice. I need a little more drama to keep me engaged, but not toooo much drama!
I miss the old Twitter where was a nice balance between interesting people and assholes.
And my Facebook feed is filled with AI-generated groups that swamp the posts from my friends. I'm wondering if there's any future for social media.
#6 Cate Hall's conclusion strikes me as very strange. She were deficient in testosterone for her own sex. Once the levels were brought up to what was typical for her sex, she did much better.
Men don't automatically feel 10 times sharper and more energetic because they have 10 times the testosterone. They feel similarly sluggish if their T is low for their biology (which would still be a very high number for a woman in most cases).
I think the conclusion here is "having a hormonal profile that's out of whack as compared to what's normal for your sex will have negative effects", not "men run life on easy mode".
> Re 45: Conception beliefs among Australian aborigines. Did you know that pre-contact aborigines didn’t know that sex caused conception?
It wasn't until the 1920s that Western medicine had a complete and testable hypothesis about the role of ovulation and the necessity of fertilization with sperm at the correct time in the menstrual cycle for conception to occur. And that hypothesis wasn't tested and confirmed until a decade later.
Although 19th-century scientists had come to the conclusion that sperm was somehow necessary for conception, in the 18th century, there were angry debates between the Spermists and Ovists. Neither side would concede that both were necessary for conception. So, it's no surprise that pre-Enlightenment societies didn't understand conception.
From the chapter you mentioned...
> By contrast Arnold van Gennep, the eminent French folklorist, considered it to be little different from the ignorance of procreative mechanisms still prevailing among the masses of Europe.
> Nuno Sempere (maybe with the rest of the Samotsvety team?) has launched askaforecaster.com
Nice! Yes, confirming that some Samotsvety people will help me answer if there is too much demand. In the meantime someone is asking good questions for $150:
> Change in non-engineering product-related roles in tech by 2030?
>
> Current AI tools are augmenting work for engineers at many technology companies. The same thing is happening for non-engineering product-related roles as well -- for instance, Product Designers are having AI included in design tools to create interface mock-ups faster; Product Managers are using LLMs to write drafts of briefs and memos; etc. Presumably, as these tools improve employee efficiency, companies will need to hire fewer and fewer employees to get the same amount of output.
>
> Specifically for these non-engineering product-related roles (which, you may have guessed, I fall into), I wanted a quick forecasting take on how employer demand for these roles may fall in the next 5 years. Is it more likely to expect a modest 10% fewer positions? Is a 30-50% re
duction possible? An industry-shattering 80-90%? And what factors might make a position more or less resilient?
#7 - the author's admission that they were taking ketamine throughout the "trip report" seems like a major confounder.
"Throughout"? Nay, that was weeks beforehand. It takes time to book an appointment with a psychologist, get the gender dysphoria diagnosis, go to the pharmacy, etc.
Additionally, the contents of the post detail collected impressions over quite a long period of time, from late 2021 to present. It would have been impractical/expensive to confound this with something short lasting like ketamine!
17. Seems hard to draw much of any conclusions about when the trend started given the massive gap in data before 1910 and between 1930 and 1980. What the hell happened to poetry at those times?
25. You will however go to hell for being a Cubs fan.
Yeah, the data in #17 seem incredibly fishy. Looks like just fitting to artefacts.
Looks like a random spread before 1930 (though with suspicious peaks at 0/25/50/75/100), then virtually nothing until 1980, then a vast amount of un-rhyming poetry collected since ~2010 or so pulling the average down. And why was there loads of poetry in one year (maybe 2015)? It's all so bizarre.
Like, look at the actual data from 1930-1980. In what world is that red trend line a fit to those dots?
The dots are individual poems. Poems with strict rhyme schemes score round numbers. ABAA is 75%. 25% is weirdly low, though.
It would be good to look at other averaging methods, like dichotomizing < 25% or maybe <= 25%, which is probably what people mean, not a switch from ABAB to ABCB.
Here's the economist article the graph is from: https://archive.is/CP8rW
Copyright issues for the 1930-1980 gap? And maybe people are just less likely to put those pre-1910 poets like Robert Bridges online than they are Pound and Eliot.
To Scott's point, non-rhyming doesn't equal modernism. Regular meter might be an interesting thing to measure like this, although it might be tough. Poems in iambic pentameter are going to have plenty of lines with variations, so it might be harder to see a pattern over time.
Gap is likely explained in this way: Before a certain year, poems are in the public domain and easily findable via book archives. After a certain year, poems are commonly published online. Between that is all the stuff that was published in print, but is still under copyright, and so is less available for automated searches.
Not hell for the Cubs fan, only 108 years in purgatory.
17 seems like a good opportunity to pound my hobby horse that song lyrics are by far the most popular and pervasive form of poetry from the 20th Century on and it's silly to talk about poetry in general and popular poetry especially without including them.
I've heard this point made by an actual published poet, and it makes sense to me.
Bob Dylan's Nobel prize should really have hammered that in.
(and some of what I'd consider among the best poems in any language ever are Pink Floyd lyrics)
28 is Stephen Fry, not Steven
I should just split the difference and call all these people "Stevhen" as a penalty for being unable to make up their minds.
You just convinced me to watch "Steven and the Stevens" again (Steven Universe episode; yes I'm autistic).
Why did the seek and then include a blurb from Mark Ruffalo? Who exactly is the intended audience for this book?
Well he did help fight Ultron.
#23. This is also true of skateboarding and has been for over two decades. Jordan’s shoe deal reworked this category and sneaker heads amplify. Consider only two people have dedicated shoe facilities at Nike. Jordan and Janoski.
It’s an interesting meeting point for urban culture and sport.
#10: Frankly this isn't very surprising to me. I'd imagine this graph correlates very strongly with the popularity of figures like Andrew Tate and etc with this age group. The amount of internet celebrities popular with teenage boys that express this kind of sexist attitude really started to take off ~2020. Would be interesting if there is any data on the popularity of those figures over time.
Supporting a society in which women get paid less than men for the same work is one thing, but actively believing that it’s actually better to pay women less for doing the same thing seems more egregiously sexist than I can believe 20% of people are.
I feel like that question is really just measuring how important boys think it is to virtue-signal that they aren’t gay liberal cucks, and apparently it has become noticeably more important to them since 2018.
2018 was the peak of MeToo and my guess is that flipped many males from "sure, feminism has a point" to "they're my enemy".
Women don't get paid less for the same work.
Even if there is such a correlation, it wouldn't answer whether those figures are causing these beliefs, or becoming more popular because they're validating the beliefs people are picking up for other reasons.
I wonder if based on the questions some of it is "fighting" with the survey team. I'm reminded of the classic Onion article "1 in 5 Americans believe Obama is a cactus": with some of these questions you can recognize "ah this is one of those *woke* questions, I'm gonna give them the *based* answer just to piss them off!" I remember reading a while back on how ~10% of high school boys report being certified submarine captains.
A way to confirm or refute this theory would be with a somewhat more oblique question or experiment that tries to get at someone's "real" views on gender. Maybe one of those money-negotiating experiments (like the Ultimatum Game) where you experimentally manipulate the name + photo of the person you are negotiating with, to see if these "I hate women" survey boys actually try to give less money to girls (even at the cost of money for themselves.
Kind of a tangent, but when they actually interviewed a panel of official submarine captains and asked them what they most fear down there, the number one answer (by far) was, of course, “dogs”.
It's not surprising, just like it's not surprising that figures like Tate are popular when we just spent 20 years telling boys (especially white ones) that they are evil, useless, and should be discriminated against. Of course they are going to lash out at the people vilifying them.
There's a question here whether social media algorithms - whether intentionally or not - are partly to blame.
47. But did you know that this effect comes almost entirely from daughters?
I really really struggle to believe this. Unfortunately the post is paywalled, but I'm pretty sure I did read it, or a similar post; and the study it's based on is not very big- around 400 women total (200 bearing daughters).
I assume it's this study: https://docs.iza.org/dp17817.pdf
Sample size: 441 mothers
They look at only the gender of the first child, but ignore the evidence (which they do cite, but don't account for) that first born girls are more likely to be followed by a sibling than first born boys. That is, the effect could be entirely that the mothers of first born daughters have larger families, and hence reduced employment. They kind of sweep it under the rug by saying "if parents anticipate building a larger family, they may be more likely to specialize, with one more focused on market work and the other on home production."
Haven’t looked at the paper yet but that seems like a huge confound (fitting username btw).
I don't have time to read the paper, but do they control for how long the mothers left the workforce? How about for multiple kids? People always say daughters are easier to raise when they are young, so I could see mothers of daughters being more likely to spend more time with them before returning to work. I could also see families that start with a daughter being more likely to have extra kids because the father wants a son.
#5 Would you let a dead spider, controlled by AI robotics, perform surgery on your eyeball? If the alternative was going blind? No. I would rather go blind.
By this power, you too will know man-made horrors beyond your comprehension.
#9 - Come back 20 years from now and the following sentence will have aged like milk: “Looking at . . . the current state of Meta’s VR universe, it is absolutely impossible to imagine a world where this kind of thing is ever going to succeed.”
Once you spend enough time trudging through the internet's gamer communities, you realize that there are millions of people out there who would drop everything real in their lives if they could "live" in virtual worlds, embodying a perfect version of themselves (often an all-powerful character) and creating romantic links with other people's avatars. Meta's strategy with VR has been to buy everything and everyone who does anything remotely interesting in that space - until the technology catches up (GPUs that can render realistic worlds, systems that suppress motion sickness and improve real world mobility -> virtual mobility) that technology looks like a massive failure on all fronts because there's almost no revenue to be found, and people who already have a dislike for video games find the whole concept extremely virgin-coded (they aren't wrong).
> Meta's strategy with VR has been to buy everything and everyone who does anything remotely interesting in that space
Turning a creativity based space into a monoculture is a surefire way to kill that space. Were they aiming for that?....
I bet not. They want the money and control, and don't care about the creativity.
>Once you spend enough time trudging through the internet's gamer communities, you realize that there are millions of people out there who would drop everything real in their lives if they could "live" in virtual worlds, embodying a perfect version of themselves (often an all-powerful character) and creating romantic links with other people's avatars.
Second Life? Meta seems to have never heard of it. Probably just as well. But for everyone else, here's a film about it: https://draxtor.substack.com/p/our-digital-selves-my-avatar-is-me
ETA: Second Life is still there, alive and well, in case that wasn't clear, and currently celebrating its 22nd birthday. The film I linked to is a celebration of what is, not a lament for what was.
A guy decided to check out an old virtual reality game from the mid-2000s that was still running. At first it seemed like no one was there, but then he ran into some users. It went places I didn’t expect and in the end I got a little verklempt.
https://www.youtube.com/watch?v=QSMDb1CWD6Y
Second life is still alive and well, but has mostly been totally supplanted in the VR space by VRChat! VRC is basically the modern day second life, with a few million people owning it and tons and tons and tons of active users. It is what i imagine Meta wanted it's metaverse to be, but i don't think Meta wants all the gay furries.
Of course the concept of living in cyberspace has been around since at least William Gibson. But most people don't want to just sit and interior decorate, and definitely not with PS2 level assets. They want something to do, a world to explore and a game to play. Then they also want to be doing the popular thing that other people are doing. What does it mean to be "an all powerful character" in Horizon Worlds?
In their defense most of the money is probably going to hardware development of the headsets which have real engineering challenges. But if you wanted to make a virtual world that people live in, really you don't need VR, you need to seamlessly merge the gaming space with a discord-like multimedia sharing app, with social media, and then with outside corps and organizations. People already basically do live in that cyberspace it just isn't integrated as it likely could be.
The thing is VRChat is still just chugging along - I just wander around weird worlds random people have created about once a week. It's an amazing experience that's completely at odds with Meta's culture and business model, and I say this as a former Meta engineer.
The metaverse kind of exists - but only for a small population of weirdos. Something like 60% of them are furries, and there's not much money in it aside from independent digital creatives selling custom avatars.
I think VR will persist as a niche weird thing for the foreseeable future. If Meta abandons it I expect a few other companies to fill the gap, albeit with less excessive subsidies. Valve is a private company beholden only to the whims of a handful of nerds with access to a free money printer and a demonstrated interest in making stuff that they want to make in VR.
If meta made VR chat with the GDP of the country I grew up in behind it instead of worse second life, I would really feel the urge to get inside the GiveUp-Machine.
If Meta did that it wouldn't have the qualities that make VRChat what it is. VRChat works because it's completely dominated by a small set of weirdos and has little mainstream appeal. This protects it from the profit incentives that would lead to metric-maxing optimization, and then you get what every mainstream social media converges towards.
You want enough capitalism to push people to make better stuff but not so much that "exploit every weakness of human psychology at industrial scale" becomes a strategy worth pursuing.
It’s “Meta” because “Dystopia” is too hard to spell.
==============
Decentraland started in 2015. The following from 2023 shows how well it has "progressed".
https://youtu.be/EiZhdpLXZ8Q?si=7NTUVCLWjKtWX41w
(kinda long.)
The idea behind gaming is to escape the real world (in a certain sense). I don't pay close attention to Meta tbh, but from what I remember, they seem all too eager to bring the real world with all its irritations to the virtual one.
Doesn't it have huge implications for non-virgin coded non-video game applications? Enterprise stuff, realistic essential worker trainings and so forth.
I'm actually long on this. I think it'll pan out.
"you realize that there are millions of people out there who would drop everything real in their lives if they could "live" in virtual worlds, embodying a perfect version of themselves "
Until that becomes an actual option, we really don't know. Millions of hippies swore they were willing to drop anything to live in a commune *if only*...
"Meta's strategy with VR has been to buy everything and everyone who does anything remotely interesting in that space - until the technology catches up"
Except, there is no guarantee the technology ever catches up. Also having a VR set, even much better than the current ones, is still not quite "Ready Player One". How many technologies have been 2 years away for 30 years? And how many technologies that were believed to be the next Industrial Revolution just became a niche hobby little influence on most people's lives? 3D printing is an obvious examples
#26: perhaps link to the story on substack: https://alexsalvinews.substack.com/p/the-election-of-pope-leo-xiv , to avoid the dozens of clicks required to read it on X? I'm sure there's a way to "unroll" the thread on X, but a webpage where that's already done for you seems good.
Thanks, fixed.
Another linking problem:
#18: the third and the fourth link are identical; I think the third link should have been this: https://maximumeffort.substack.com/p/a-measured-response-to-benthams-bulldog
Thanks, fixed.
In general, it’s really annoying to see a link to a Tweet, figure there’s probably more to it than just one Tweet, and then realize that there’s probably a thread, and you’re just not allowed to read a thread if you aren’t logged in to a Twitter account.
"Then universities borrow money now against that future windfall. What goes wrong?"
The same thing that goes wrong if I go borrow a million quid based on "next week for sure I will win the lottery so I definitely will be able to pay you back, honest!"
"I thought you weren’t allowed to do any of this, and certainly not talk about doing any of it, on pain of excommunication?"
We're Catholics. You seriously expect us to stick to the rules? 😁
The difference is that you probably aren't going to win the lottery, but it's overwhelmingly likely that Democrats will have unified control of the U.S. federal government at some point in the next several decades. Not saying there aren't reasons why this might not work, but that isn't one.
The right will call it a "bailout for Harvard!"
Of course they will, the right hates everything about this. The whole argument (which, again, is not entirely free of holes) is that the Democrats will eventually have the opportunity to pass legislation on a party-line basis over the right's objections.
It won't just be the right, I wouldn't like to be the Democrat official charged with explaining why a moneybags Ivy League college is getting taxpayer money to repay their borrowings, while [insert issue here: housing crisis, potholes in the roads, systemic racism, healthcare costs, etc.] are crying out for funding.
Oh, the Democrats are likely to get back into power. But when they do, will the promised money be there? Or will it be spent on other, more "worthy"/"needy" causes? Suppose Harvard borrows against the promised payout, do you really think there won't be outcry from the more lefty side of the coalition about rich endowment white privilege colleges getting a government bailout instead of state schools., or black colleges, or any of the rest of the list of "more urgent social needs"?
See Biden's student loan forgiveness, yes it did happen, but not as broadly as people thought it would:
https://www.pbs.org/newshour/politics/despite-collapse-of-his-forgiveness-plan-millions-had-student-loans-canceled-under-biden
"Despite failing to deliver his promise for broad student loan forgiveness, President Joe Biden has now overseen the cancellation of student loans for more than 5 million Americans — more than any other president in U.S. history.
In a last-minute action on Monday, the Education Department canceled loans for 150,000 borrowers through programs that existed before Biden took office. His administration expanded those programs and used them to their fullest extent, pressing on with cancellation even after the Supreme Court rejected Biden’s plan for a new forgiveness policy."
I seem to recall he was able to cancel loans for those who had entered into government employment, which may have satisfied his base.
> Or will it be spent on other, more "worthy"/"needy" causes?
Perhaps if we all started tweeting "giving money to Harvard is racist", the universities would realize this risk and stop borrowing?
1. Loan forgiveness was severely curbed by the courts (fortunately, imho). Unlike the current admin, the previous one was in the habit to respect sentences
2. We are not talking about huge sums, at least when compared to the national budget. For the left to be able to block a 0.x% addition to the budget, that the center-left campaigned on and publicly promised, they'd need a level of control not even remotely close to the one they have
3. The left (as in, the Squad) has been pretty vocal about the selective grant cancellations and has repeatedly framed it as an attack on research rather than on Harvard. Nobody guarantees consistency obviously, but to completely renege their current position for no appreciable reason would be deeply misguided and there is no reason to think they would.
> Democrats will have unified control of the U.S. federal government at some point in the next several decades
I disagree. While I do agree that *some* left-leaning party will have unified control at some point in the future, the present-day Democrats have decayed past any point of recovery. Their putatively successful political descendants will have to deliberately distance themselves from present-day Democrats to such an extent that making high-confidence predictions about their policies is difficult (if not outright impossible).
Are you saying that there'll be an ideological realignment like in the 20th century, or are you literally saying that the United States Democratic Party will never again have unified control of the federal government? If the latter, well, I think that's less than 20% likely and would bet accordingly.
I wouldn't say "ideological", I'd say "political" or perhaps "messaging" realignment (as I doubt that politicians have anything resembling a consistent ideology). And there will likely be an organization *named* "US Democratic Party" that may, in the next 10..30 years, gain full control of the US Government. But that organization's internal structure, policies, messaging, strategies, financing priorities, etc., will be so different from the current Democratic Party that making confident predictions about their behaviour is premature. Otherwise, they'd have no hope of taking any kind of control over anything.
This is what people thought of the Republicans in 2013 and the Democrats in early 2005.
The problem is planning for a sure thing, and planning for a time period. If the Democrats regain control of the federal government in 2028, that looks very different for a university carrying large loads of debt than it does if Democrats regain control in 2036. Also, maybe the Democrats don't back pay any grants, or at a lesser percentage than full, or even decide that it worked out pretty well and don't even add new grants. Then the universities are in a really bad position. This would be a very bad bet to make.
The question is whether any banks or hedge funds or whatever would be interested in bearing that risk for the right price.
Any bet made at the whims and capability of elected officials to do something that they may or may not even want to do is a bad bet. There is no guarantee that Democrats as a whole would make such a change, let alone individual Democrat politicians. There is even less guarantee that they would have the ability to do it, such as failing to get all three branches, having too small of a majority to pass disputed legislation/filibuster, etc.
Hedge funds are in the business of betting on things that aren't guaranteed. It's just a question of at what price.
Maybe, but there is nothing stopping them from selling state-contingent claims (x$ will be paid to the bearer *if* Dems win AND make good on their word). If they trade at 10c on the dollar, too bad, it's not Harvard's problem as long as they have buyers.
Also, once these state-contingent claims are on the market, they create a natural, powerful constituency pressuring Dems to in fact make good on their word.
This depends on a much longer chain of hypotheticals than that.
For this scheme to work, it's necessary that 1. Ds regain unified control sometime in the next decade or two (pretty likely), then 2. those Ds retain enough respect for commitments made today that they make a serious effort to pay out the (probably very large) back funding, and 3. that effort successfully beats out all other competing demands for funds, in an increasingly strained budget environment.
Then there's the whole question of timing. This bet looks very different depending on whether you anticipate the payout in 4 years, vs 8, 12 or 16. The further into the future it is that Ds regain control, the less likely that points 2 and 3 above will succeed. Then, of course, there's the time cost of money to consider.
Meanwhile, this policy requires a very public commitment and funding negotiations; thus, it can't be done quietly. Trump has all kinds of ways to interfere with this process, from going after the lenders in various ways (all banks are extremely vulnerable to USG pressure) to using other levers against the universities.
Think about this as if you are the prospective private lender. Would you be willing to pay out $billions to universities, with repayment contingent on someone else fighting and winning a thankless political struggle in an unclear number of years? I am pretty sure that would not pass any conventional underwriting process.
Whereas my comparison would be paying in US Dollars during the War for Independence, or in Confederate dollars during the Civil War. It's yet another dangerous escalation in the political landscape.
#4 – this was likely Moderna's CMV vaccine candidate.
I'm glad they can detect this, because engineered CMV is one of the things that keeps me up at night.
44: Ultimately, I think the most concise way to think about "The purpose of a system is what it does", is to see that most people are split into two diametrically opposite interpretations of it:
1) What a system does must be the purpose that was intended by its designers
2) The "purpose" of a system is a concept that can, and should, be described independently of any intent or design
And the way to resolve this is to realise that POSIWID itself is a very confusing sentence out of context, and Stafford Beer (presumably) never intended it to confuse people, so:
Under interpretation 1), POSIWID is a self-contradiction;
Under interpretation 2), POSIWID proves itself - its a meme that perpetuates discussion of itself by the conflict between both interpretations, and it's useful to imagine that as its "purpose" in the context of asking how it sticks around.
If a system is sufficiently "bad" it will be torn down. I think that's the best way to hear POSIWID. Democracy, worldwide, is used as a tool to let the CIA/United states government influence elections and policy in other countries.
It seems like what people ordinarily point to when using the word "purpose" is the ways in which a thing promotes some given terminal goal- so not just the intent of a designer, but also not really independent of intent.
Most of the people doing the pointing, however, don't actually have a solid grasp of terminal goals as a property of agents and distinct from instrumental goals, so the way people end up conceptualizing purpose tends to be more like "the ways in which a thing promotes [unnamed, half-formed concept]". People aren't sure whether that half-formed concept is an objective property of reality or something subjective, so we get lots of very confused memes like "the meaning of life" that assume purpose can exist independently of any agent.
Reading through the linked blog post, it sounds like those early cyberneticists noticed that anything a thing does could constitute its purpose to some hypothetical agent with some particular terminal goal. But they also didn't really have a solid concept of terminal goals and were operating under this confused cultural assumption that purposes are properties of things that can exist independently of agents, so the observation just looked to them like "anything a thing does could constitute its purpose"- phrased even more confusingly as POSIWID.
#15: yeah, someone whose funding gets cut and loses their academic position is most likely to look for a job in industry, not to start a startup. Launching a startup (at least in biotech) usually requires preliminary data which you can't get if you don't have a lab.
When we recently hired for a PhD-level position we got over 300 applicants the first day alone.
#16: an interesting idea, but what if the Republicans win in 2028?
#28 - they also got George Church! https://intelligence.org/2025/06/18/new-endorsements-for-if-anyone-builds-it-everyone-dies/
28: I'm not as surprised at that one, he seems like the kind of person to be into this.
I don't know, Stephen Fry also seems like the kind of person to be into this. The surprise is that he chose to spend his weirdness points this way given that he has fewer to spend as a much more mainstream figure.
#8 - I can get 1/100th of the karma of saving a life by simply praising good deeds? This does not seem remotely play-tested. Surely praising good deeds a hundred times is easier than saving someone's life. If paying someone to praise good deeds on my behalf counts, I'm sure I can hire someone to do that at a much more effective rate than it would cost to spend the money saving lives.
And "adopting an orphan", something that brings with it daily obligations for years, is just 50 karma! I'm never going to adopt any orphans with this scoring scheme. (Though maybe it made more sense in Ming China, where children may have been economic investments rather than money sinks. But still.)
On the other hand, saving lives is *also* surprisingly cheap. GiveWell generally gives estimates of $3500-$5000 in the modern world, and perhaps it was even cheaper in Ming China due to more people near you living close to the subsistence threshold.
On the other, other hand, you don't merely have to pay someone to praise good deeds; praising is the easy part. You have to pay them to identify good deeds! Which is such a thorny proposition that modern philosophers still have not solved it.
I think it's about $2 to keep a child from going blind. That may not be saving their life (or it might be, consider orphans and other destitute), but it's certainly keeping them as "valued members of society."
I always like to run that as a metric for "How Much does this charity deserve to be funded?"
This seems to good to be true, but in case it is, I'd be up for curing a number of kids.
Look at the link below... Looks like we still need more money/effort on this front, hundreds of thousands of children are still going blind (and that's with a 50% mortality rate after blindness).
Do you have a citation? Usually when cost numbers are that low, someone is conflating 'cost per treatment' with 'cost per preventing negative outcome', which are only the same if 100% of treated people would have the negative outcome without the treatment and 0% would have it with the treatment.
https://pubmed.ncbi.nlm.nih.gov/35294044/
We have upwards of 200,000 children a year going blind (with a mortality rate of 50% in the first year alone after blindness). This is after distributing 400 million doses per year.
https://en.wikipedia.org/wiki/Vitamin_A_deficiency
Quoting from wiki:
"Globally, VAD is estimated to affect about one-third of children under the age of five, causing an estimated 670,000 deaths in children under five annually.[8][9] It is most prevalent in sub-Saharan Africa (48 percent) and South Asia (44 percent).[8]"
Yes, this is cost per treatment (vitamin A shots). To "fully prevent" Vitamin a Deficiency you need a shot a year...
OK, let's assume that it takes 4 shots then, to fully prevent VAD (since this is largely a <5-year-old issue.)
Your paper doesn't give blindness stats (which I know was the initial claim), but it does give all-cause mortality improvement. It drops all-cause mortality from 2.6% to 2.3%. That's 3 lives per thousand treated children, or per 4000 doses, or about $1,333 per life saved. Very, very efficient!
But... while this doesn't give blindness data, note that 50% of kids die within a year of developing blindness. That caps the # of blindnesses prevented at twice the number of deaths prevented (which would be true if all the mortality improvement was due to preventing blindness; unlikely.) That means that you have to pay at least $666 per case of blindness prevented (and probably quite a bit more.)
Still a bargain; but not wildly implausible like $2 apiece would be.
GiveWell's rough estimate is more like $6,000 apiece.
Each treatment costs $2.70 ($1.35 per shot, but you need two shots). Divide that by the probability that the child would have gone blind without the treatment but won't with it, which they calculate by multiplying the following numbers together:
- 1.3% prevalence of night blindness in Guinea, Burkina Faso, and Mali
- 10% of cases of night blindness progress to full blindness (there apparently are no studies of this so this number is a guess but they think it's a generous one)
- 68% chance the treatment works
- 50% chance of surviving early childhood after going blind (as your comment accounts for)
They then go on to calculate DALYs (divide by 0.187 disability weight for total blindness, and by 25 more years of average life expectancy for a blind child who doesn't die right away) and conclude that blindness prevention through vitamin A supplementation costs about $1,200 per DALY. This is worse than giving cash, which they estimate improves lives at a rate of $1,000 per DALY. So ultimately not the most good you can do.
They recommend Helen Keller International because of the mortality benefits, for which the numbers are better ($3,500 per life saved, not sure the exact conversion factor to DALYs but it's obviously way better than cash).
Source: https://docs.google.com/spreadsheets/d/1tIK6aJEzQReTq8PveyKk3o-JLwwWXquFr67_WrARXDc/edit?gid=373284575#gid=373284575, which is linked from https://docs.google.com/document/d/1wuZC00Llw3SpvNFGCadCHNFBSqwo4wU3Elu80n6Cl6k/edit?tab=t.0#heading=h.6icmli4ndtv4, which is linked from https://www.givewell.org/international/technical/programs/vitamin-A#footnote248. Cost-per-life-saved number is from https://www.givewell.org/impact-estimates#Impact_metrics_for_grants_to_GiveWells_top_charities.
Thanks for running the numbers! I'm definitely on board (and helping) with yellow rice and other tricks to no longer need this particular charity.
is there somewhere I can donate for this cause?
So I'm lazy:
https://www.givewell.org/charities/helen-keller-international
I think it's clearly written for a different society, where words had more weight, anonimity was not an option, and if you told a butcher to their face they had to change job you could have been in for an unpleasant retort.
Presumably, as both the material impact and the cost of praising good deeds and harassing butchers decreased, the karma points would decrease in proportion
#1 - Scott, you say that post helped you crystallize some thoughts on what good writing and good taste are in general. Sounds like the perfect topic for you to write about - will we get a glimpse?
I hope so!
#52 - I wrote about an earlier version of this therapy here: https://denovo.substack.com/p/base-editing-coming-soon-to-a-liver
Despite having a successful therapy, VERV stock really didn't do well over the last 4 years. They recently got acquired by Eli Lily for $1.3B but this is down from about $2B in 2022 (and recently it was as low as $300M).
I personally lost about $15k on this. Just goes to show that biotech is hard to make money in.
That is odd. It seems like the stock should be higher than that. Any idea why it's so low? Even on the day the results were announced (April 14, 2025), the stock only went up ~$0.90 to ~$4, off from a peak of >$70 four years ago.
> looks like the rhyme decline started well before then.
Paradise Lost, an epic poem published in 1667 by John Milton doesn't rhyme. I gather many people were critical of it at the time for that reason. My only complaint about Milton's poetry generally is the plethora of ancient Greek characters, divine or otherwise, who pop up in it. Presumably readers in the 1600s would have been familiar with these references, but most people today find them obscure and baffling. Piers Plowman, written towards the end of the 1300s, doesn't rhyme either.
Doesn't Milton treat rhyming as newfangled? My understanding is that English poetry didn't rhyme early on (Beowulf used alliteration, Shakespeare used patterns of stressed syllables), and that rhyming was introduced via French poetry.
Yes, it's possible that thyme in English is a cuiltural.cringe towards.The greeks.and romans.
AFAIK, classical Greek and Roman poetry doesn't rhyme either -- it's based on metric structure, like Old English and Sanskrit poetry (which suggests it's an Indo-European thing). Poetry in Latin starts rhyming consistently in the Middle Ages.
Shakespeare doesn't predate the initial burst of English rhyming poetry, he came in towards the end of it - his sonnets with their ABAB CDCD scheme were a compromise away from the ABBA ABBA Petrarchan sonnets that had a vogue in the 1500s (which are far easier to write in the very rhyme-friendly Italian lexicon than the English one).
Yes, rhyming was from French poetry. As with all barbaric, agglomerating English -- rhyming is much, much more difficult in English than languages that do it natively.
Relevant smbc: https://www.smbc-comics.com/comic/rhymes
If you count Shakespeare plays as poetry (most of them are in meter) you get a very small fraction of rhyming lines.
#48: Bret Deveraux notes that the "strategic" effects of the air campaign against Iraq, i.e. strikes on leadership, communications, electric power etc., were minimal.
https://acoup.blog/2022/10/21/collections-strategic-airpower-101/
> The DoD’s own report, issued in 1993 (Cohen and Keaney, Gulf War Air Power Survey (1993)) noted first that only some 15% of strikes were against ‘strategic’ targets, while strikes against Iraqi ground forces consumed 56% of strikes; as Coalition air forces exhausted their list of strategic targets, they switched over to strikes against ground forces.
> Despite effectively running out the entire list of strategic targets, Cohen and Keaney nevertheless conclude that strategic effects were broadly not achieved. Despite striking the Iraqi communications network with more than 580 strikes, “the Iraqi government had been able to continue launching Scuds” and “sufficient ‘connectivity’ persisted for Baghdad to order a withdrawal from the theater [Kuwait] that included some redeployments aimed at screening the retreat.” Consequently, “these attacks clearly fell short of fulling the ambitious hope” to “put enough pressure on the regime to bring about its overthrow and completely sever communications between Baghdad and their military forces.”
So no, strategic air power was neither a necessary nor a sufficient reason why the Iraqi army crumbled. They crumbled because the Coalition forces represented the undisputed peak of modern warfare on a perfect battlefield (flat desert) whereas the Iraqi forces had neither the equipment, nor the training, nor the doctrine to resist effectively. There is no reason to assume that an invasion of Taiwan would be the first proof of the effectiveness of strategic airpower after a century of failed tests.
I would generally caution against using Deveraux as an authoritative source outside his area of expertise, i.e. Roman era military history. The post about strategic bombing draws on WWII for a lot of the material, where bombing was highly inaccurate and the belligerents had strategic goals not clearly linked to military objectives like "bomb the civilians to make the enemy regime collapse" (which definitely didn't work). Anyway, what #48 is proposing isn't strategic bombing in the sense Deveraux is talking about, where air power alone is the deciding factor. He's talking about a combined arms approach where the enemy forces are crippled by precision strikes taking out key assets - so that a ground force can then overwhelm the disorganized enemy. The ground force is key here. But it doesn't matter for Taiwan because the ground force has to be lifted across an ocean for a contested amphibious landing, which is an entirely different type of operation.
I'm not sure what the rest of the ACOUP post has to do with my objection. I presented a specific quote from the post, not the entire post, because that quote dealt with the question of whether you can use air power to disrupt an army to the point that it simply breaks down at the slightest pressure. That was the theory developed for the 1990 Gulf War, it was tested in that war, and unless you're saying Deveraux misquoted or misunderstood the DoD report, the answer seems to be "no, that theory didn't work". The #48 tweet repeated that disproven theory, elevated it to fact, and builds its argument on that faulty foundation. That is why I object to #48.
Didn't tactical rather than strategic aircraft play a big role in destroying Iraqi armored vehicles? The A-10 "Warthog" comes to mind IIRC.
Yes, that's the point. That is very different from the imagined strategic use of airpower, i.e. to basically win the war entirely from the air rather than by providing boring old ground support.
A humble thread image suggestion: the greatest logo in Finnish team sports. https://en.wikipedia.org/wiki/Ilves
That is amazing.
#9 - Reporting on how much money Meta has lost on "the metaverse" often conflates their crappy-looking VR game Horizon Worlds with their entire Reality Labs division. It's the division that's lost $70 billion, but it also makes VR headsets and does all sorts of pie-in-the-sky VR/AR research.
Horizon worlds has by all accounts been horribly mismanaged, with developers that refuse to actually put on a headset and test the game, but not quite to the level of spending 50 Star Citizens worth of money.
The brute fact of software engineering being extremely inconvenient to do while wearing a VR headset may be the single biggest obstacle the technology faces.
# 17, the poetry data from the economist could equally support the hypothesis that rhyming poetry has survived better. They clearly have a lot more data for recent poetry, and the dataset is short on unrhyming poetry from the past. They also clearly have a lot of rhyming poetry from today, but it’s getting overwhelmed by the unrhyming stuff.
In other words, # 17 is survivorship bias.
#31. That comment surprises me: as I read AI 2027, I thought that if it was intended to be read by any one man, that man was the Vice President.
#1 - Some points seem similar to the old discussions about AI Dungeon prose quality and the new sampling tricks to force LLMs to stop repeating the same cliches. I wonder if there are more in-depth writings about both history and state of the art of AI writing, including the new experiments and tricks.
I'm curious about the karma thing. Why is it virtuous to persuade someone not to abandon their village during a famine?
Seems to me that if food is scarce and someone leaves, there's one less mouth to feed and more food to go around for the rest of the village. What is the moral calculus that deems this not a good thing?
Perhaps because this person most likely become a highway robber.
Just a guess but it might be something to do with family obligations especially to ancestors. I believe making offerings to ancestors graves was considered important at the time and this might not be possible if you leave.
#19 - Mifepristone is the "abortion pill", not the "morning-after pill". The morning-after pill (Plan B) is a large dose of (iirc) estradiol - basically a large dose of regular birth control.
Thanks, fixed.
Morning after pill is levonorogestrel, a progesterone receptor agonist. Birth control is reliant on progesterone to limit ovulation, although most formulations contain some estrogen.
Thanks for the correction!
> I’d always heard modernism was a reaction to the trauma of the World Wars, but it looks like the rhyme decline started well before then.
It's also well to keep track of the language; I think it's not a universal phenomenon, but a regional one. At least, intuitively the more modern Russian-language poetry is much more rhymey than its English counterpart. (If you suffer from a need of rhyming poetry, go ahead and learn Russian and I'll share some of my favorites.) I've been re-reading Sorokin's novel "the norm" (1980-ies) recently, which is all postmodernist / countercultural / in-your-face-reader, and the verses there rhyme — that's the one tradition Sorokin doesn't break. (Although of course Russian vers-libre also exists.)
(Also, how would war trauma induce rhymeless verse? Did the wars include rap battles?)
I think Brodsky, who was both a Russian and an English (well, American) poet, had quite a few things to say about that - in particular, not just that Russian was linguistically more suited to thymes, but also that Russian poetry was much younger and still hadn't grown out of its formal stage.
I'm not even sure what it means for poetry to have a formal stage. A stage where significance is placed upon the form of a poem? As opposed to what, not noticing form, not having form? The words in which a poem is written are still part of it, it doesn't matter what conventions it does or doesn't follow — you still can't get away from the fact that the words are light or heavy, or curt, or clumsy, or whatever. If you happen to speak Russian (which I though you might?), this is one of my favorite Russian free verses (by Ivan Akhmetyev):
русская поэзия катилась
на ритмических салазках
пока не выехала на советский асфальт
веррлиббрр
No rhyme, no regular rhythm, and it wouldn't work without the form. When Kharms and Vvedensky were dismantling all previous literary conventions of their literature, were they writing rhyming verses because they hadn't outgrown a stage? What does it even mean? If they had a working tool that contributed to their unique styles, why were they supposed to outgrow it?
***
Actually, now I'm a little worried and confused: when a connoisseur of English poetry reads, I don't know, Walt Whitman, do they just not notice the texture of the material that their hand touches because it's just form?
I don't think it's a technical term, but I meant with it poems with a specific rhyme scheme and rhythm, yes. I assume most of the rhymed poems in 1900 in that Economist graph also had a fixed rhythm.
Honestly, I'm not a connoisseur of poetry, English or otherwise, and am not able to answer any of your questions (or even understand some of them). Except that I think it's generally accepted that in most of art, there are periods when certain things are done, and then periods when they are cast out. Scott has written a lot on architecture, for example, and you really couldn't write a novel with Victorian-style chapter summaries nowadays without it coming off as a parody. Of course most manifestations aren't as clear-cut, but I think it's pretty clear that the corresponding graph for Russian rhymed poetry would be shifted right by at least 50 years, probably more.
I see. My point is that the history of Russian poetry includes some events of throwing out all previous conventions and "things were done this way" (the Oberiu group is a salient example), but the fact of rhyming remains preserved because it still works as a tool in terms of the artistic impression it makes on the reader, even when other conventions are broken, so it can't be thought of in terms of "someone just reusing the same tools because they didn't think to try anything else". Breaking with tradition is something of a common theme in the history of Russian literature.
I don't disagree with that - I just vaguely remembered Brodsky saying something on the lines that Russian poetry hasn't been using rhymes as long as English poetry, and so they still worked just fine, which I think is what you say as well.
Sorry if I did a bad job putting that into words...
No worries, sorry if I came off as snarky or anything.
In any case, modernism in all the arts clearly starts before the world wars. In music, there’s clear breakdown of tonality and meter in Debussy at the turn of the century, and even some Wagner and Liszt in the 1880s, and you have very obviously modernist music like Pierrot Lunaire and The Rite of Spring just before the First World War. In visual art, you’ve again got the post-impressionists at the turn of the century, the impressionists a couple decades earlier (who are clearly modernist, even if they no longer trigger anti-modernist viewers) and before the First World War you’ve got works like Les Demoiselles d’Avignon and Kandinsky’s Improvisations. Literature may have been slower to turn to clear modernism, but people like Dostoevsky, Kafka, and Jules Verne all have elements that seem modernist well before the world wars.
#31 Ross Douthat interviews J D Vance. - Well, I guess I guess I can quit wondering if I’m being locally transgressive by linking to a piece in that particular source.
#49 - I couldn't find any evidence this happened, other than that one unsourced tweet. A reply to that tweet linked a blog that supposedly documented it, but it's behind a paywall (https://quillette.com/2025/05/16/is-the-university-of-austin-betraying-its-founding-principles/).
I think Quillette paywalls articles after a time delay, or something like that. Which means the Wayback Machine has it: https://web.archive.org/web/20250531165142/https://quillette.com/2025/05/16/is-the-university-of-austin-betraying-its-founding-principles/
It seems they didn't "fire someone" exactly, they terminated their funding for that person's organization (the Mill Institute). Still pretty clear that UATX's dedication to free speech means "free speech for people who agree with me." i.e. they are just picking the opposite pole in the culture war from where most of academia sits, not actually trying to create an open ideological space.
Thanks for the link. Disappointing stuff.
Strange that folks can believe that father's could be biased toward sons but not that mothers might be biased toward daughters. That would also explain the difference: mothers preferring to spend time at home with daughters more than with sons.
54 - The FDA regulation in question just lets you avoid FDA regulation. You would still be violating the US patents, and a court can (and would) order the importation to cease.
>It would be convenient if there were seem deep synergy between the Jewish religion and psychedelia. But there isn’t. So they can either drop the issue, or else confabulate something.
I am kinda surprised. To me Judaism would be the second least psychedelic religion after Islam. First because it revolves heavily around "lawyering". Second because it focuses on text, not visual imagination.
Of the religions popular around here, I would nominate Orthodoxy as the most psychedelic - the concept of theoria.
A large mainline branch of Orthodox Judaism, Hasidism, is heavily grounded in mystical works attempting to describe the relationship between Man and G-d and the nature of reality, a lot of it using geometric symbology. Fertile psychedelic ground, and there is in fact a growing movement on the fringes to incorporate psychedelics into religious practice.
I'm guessing it's IQ mediated. Psychedelic users skew high-IQ and Jews are high-IQ.
VR: I thought nobody is using this anymore, and then I found out it still has a fringe following among people who like to see themselves in non-human bodies. Which means... Zuck spent $50Bn to build... Furry Heaven.
Beyond the humorous aspects, can someone tell me how are people like this rich? Muh super efficient meritocratic markets?
> VR: I thought nobody is using this anymore,
It's decent for (regular) porn, I guess.
> Beyond the humorous aspects, can someone tell me how are people like this rich? Muh super efficient meritocratic markets?
I use whatsapp, facebook and insta every day, many times a day. It's reasonably earned.
And he's not spending his pocket money, AFAIK he's CEO of Meta, so he's controlling more than his own purse. With the blessing of the stock owners.
> Beyond the humorous aspects, can someone tell me how are people like this rich? Muh super efficient meritocratic markets?
He's spending Meta's money... and I think this is a pretty reasonable choice for humanity, if not necessarily for Meta. They've pushed the frontier on VR/AR tech quite a bit, both in fidelity and cost-efficiency. I don't know how that pays off for humanity in the future, but it seems like a pretty good bet that it will, eventually.
I know multiple people that use a Quest instead of a monitor when working because they like being able to have a variable number of "monitors". Also, a lot of kids have a quest and play the games, so it's something that I think will gain user share as people that grew up with it age.
#43 I have always thought militant anti-colonialism was odd and not in tune with the public in most places. The implication of anti-colonialism is that pre-colonial feudal kings who might still be foreign were better than European liberals who built roads, trains and had some form of rule of law.
Note that the question seems to be asking about the modern influence of the former colonizer, not about whether the colonization itself was good or bad. You can believe that the 1930s Germany that invaded your country was literally Hitler while still believing that modern Germany has a positive influence on your country.
The two countries with the highest positive view of their former colonizer, Cabo Verde and São Tomé and Príncipe, have a very large migrant community living in the territory of that former colonizer, Portugal.
A quick back of the envelope calculation tells me that the equivalent of 20% of São Tomé and Príncipe's population (40k compared to 200k) lives in Portugal, and the figure is 17% for Cabo Verde.
I suspect that for countries with a large community living in the former colonizer, and where most people would know at least someone living there, the answer might have more to do with the general image they have of the country (Portugal) than with the actual political and economic influence of the former colonizer.
Caveat: I'm including acquisitions of citizenship by Cabo Verdeans and Santomese (?) : https://en.wikipedia.org/wiki/Demographics_of_Portugal#Immigration
Cabo Verde was uninhabited before the Portuguese showed up. No Africans at all are indigenous to it. The island was colonized, but no group of people were "colonized" in the political sense of the word. There's only African-descended people there at all for the same reason they're in Alabama. That's a pretty unique case, it appears to have been swept up by the marxist pan-African fantasies of some guys in the 60s but it was never really of a kind with the other nations in that poll.
The world had a log-normal distribution of income in 1820, but had formed a pretty dramatic bimodal distribution by 1975 between colonizer and colonized nations, and then since 1975 has returned to a log-normal distribution: https://ourworldindata.org/the-history-of-global-economic-inequality. This strongly suggests that colonialism did hold back the Global South on average. The absence of colonialism doesn’t necessarily mean those countries would have had feudal kings forever; they could have modernized better through trade with the developed countries, rather than being colonized, like Meiji Japan did.
I disagree, Finland, S Korea, Taiwan, Norway, Poland, Canada etc were colonies, the richest states in Europe like Switzerland, Sweden, Luxembourg had no colonies.
There is no correlation between colonialness and current GDP. We can easily test the hypothesis with neighbouring countries with different histories.
8) A karma system that punishes/rewards for unrealized thoughts is a pretty stupid karma system. It's more like a juju system.
#2 I love the fact that 664 Texans apparently weren't satisfied with the choice on the ballot presented to them (Allan Shivers the Democrat nominee vs. Allan Shivers the Republican nominee) and went for a write-in candidate, Allan Shivers.
Maybe they were those for whom it was the only way to preserve the sense that they "voted for the man, not the party".
#18: I don't really understand what you got out of the Dylan Black articles. Your argument about Tegmark's Mathematical Universe was the correct response to the fine-tuning argument. What Dylan Black says seems between straight-up wrong and missing the point.
He's using a very specific formalization of probability theory from Andrey Kolmogorov from 1933. I think there has in fact been some progress in the field of probability theory since then. For example, it's totally possible to have a uniform prior over the natural numbers (or even the real line): it's called an improper prior [0] and can be rigorously defined. The Beta(0,0) distribution is another example of an improper prior which is often used.
Or take this example in his article:
"Problem Statement: You're presented with an urn. I tell you nothing about this urn. You reach in, and draw one red ball. What is the probability that it contains 50% red balls? If you feel uncomfortable, good. You should. This isn't a hard problem or an imprecise problem. It's not a problem at all. It's a grammatically correct but meaningless arrangement of mathematical symbols, it is ill-posed."
I mean sure, the problem is underspecified if you want to calculate an objective number, but one of the central points of books like E.T. Jaynes' *Probability Theory* (2003) is that it all depends on what prior information you have. A human does have prior information about what kind of things are stored in urns in mathematics departments. And so, gun-to-their-head, they could give some probability for the above question. A superintelligence could use some approximation of Solomonoff induction to give an even better probability in answer to this question.
[0]: https://en.wikipedia.org/wiki/Prior_probability#Improper_priors
K complexity isnt even the only version of computational complexity, And assuming probability is computational complexity begs a lot of questions.
I didn’t read the Tegmark or Black, but the post you’re replying to doesn’t mention Kolmogorov complexity in any way. Kolmogorov’s foundational work on probability comes years before Turing’s idea of computation was described, let alone Kolmogorov and others using that to develop theories of complexity.
Is the right photo of Allan Shivers more faded or is that an optical illusion?
Illusion, I checked the page history and they're the same image.
#56: that paper is ancient. But so's this other one I found on Gwern's server somewhere, one of my favorites next to that German one about fucking vacuum cleaners: https://gwern.net/doc/technology/1958-bain.pdf
This initially seems like an industrial design study, but soon gets into unintended psychology: some children, when trapped in a fridge, do nothing. Plus the summary in Interlingua is a cute historical touch.
16. That assumes the Democrats regain power soon. Historically, parties could be out of power for decades. This sounds a lot like like Gambler's Ruin.
That is rather an overstatement, but regardless, presumably that would be factored into the interest rate.
Not an overstatement. Before Gingrich's revolution in 1994, the Republicans hadn't had the House for the previous 40 years.
And yes, it would. I'm not saying they wouldn't be able to get loans: it's the repayment that risks ruin.
Yes, I am aware of that example. That’s what makes your claim "rather an overstatement", not a falsehood.
I wonder if banks would be willing to allow them to wait until the windfall comes in before repaying.
Probably, but I expect they'll eventually stop extending them further loans, and keep tacking on interest to what they already borrowed.
They can just trade state-contingent claims. The neat thing is that since the state they are contingent on is a political decision the bearers of the claims can influence, the more claims are emitted, the more likely the state is to be fulfilled.
Even if some non-Trump Republic will succeed him, it's hard to imagine he'd resist the pressure forever
I write, and almost exclusively read, poetry that rhymes. I also spend a decent chunk of time looking for modern poets who do the same, people like Chelsea Rathburn, A. E. Stallings, and Faith Thompson. I know there's a famous tweet (once linked to on this newsletter) that says something like, "People want poetry that rhymes. But we won't give it to them." So why aren't the three people I just named more famous? They're brilliant. Maybe it's the lack of interest in poetry in general. It's certainly not a new state of affairs.
36. This are valid points but his waving away the concerns about diversion doesn't quite convince me
"Well, the Israeli government says it's because of large-scale aid diversion by Hamas. But they have never once provided evidence of that, publicly or privately"
(notice "large-scale" btw)
There is some evidence but I don't want to get into that discussion. From a purely Bayesian perspective, what are the odds that Hamas didn't exercise some kind of control over the old process? Does he seriously think that they didn't get food for their fighters?
I couldn’t believe when I got to that part. He doesn’t think the government of a polity which is currently being starved-out in a war for survival would commandeer resource shipments? If Hamas *wasn’t” using food aid to feed its troops, shouldn’t every Hamas soldier have starved by now? It’s absurd.
If hypothetically Hamas wasn't stealing food, couldn't Hamas members just show up at the aid distribution site dressed as regular people and get food like any other person?
1. One has to show that there is still a significant number of Hamas "soldiers" around after 2 years of war
2. Even then, they are not really soldiers, they are irregulars. They can very well just tell their people to get food on their own, much like any other guerriglia force did.
3. At this point (actually, it's been more than one year), Hamas cannot keep criminals in prison, cannot prevent them from raiding armories, and formely incarcerated gangsters such as Yasser Abu Shabab are openly claiming territory for their gangs and looting aid on their own. It is not hard to believe that Hamas might be unable to seize aid simply because other local actors have disloged them
I'll add that GHF workers have been killed or tortured by Hamas (https://www.fdd.org/analysis/2025/06/29/hamas-places-bounty-on-aid-workers-ghf/), which (even aside from any Israeli evidence) implies pretty strongly that Hamas has reasons to not want aid distributed through them. I can't think of an explanation for this that isn't that it's because they control the alternate means of aid distribution.
It's also worth noting here that UN (and affiliated aid orgs) have been suspiciously quiet about this (see e.g. https://www.jns.org/ghf-lack-of-un-condemnation-of-hamas-killing-of-aid-workers-shows-bodys-underbelly/ ). There's a range of how we can interpret this - from somewhat benign ("they don't want to piss off Hamas in order to keep their neutrality) to amorally greedy (they don't want anyone else cashing in on their aid bucks project) to outright hostile (they hate Israel and don't want anything that can help Israel reduce Hamas's power).
But whichever interpretation you go with, the underlying fact that UN aid orgs are incredibly politically hostile to GHF stands, and UN aid insider criticism of them should be seen in that light.
I could think of any number of explanations - they see GHF as enemy combatants (as they are run and mostly manned by the US government which is also supplying Israel with weapons) while the UN not, or they (like many humanitarian orgs) see the GHF as an attempt at ethnic cleansing, or they fear the GHF will shift the populace away from Hamas strongholds and deprive them of human shields...
What's the risk in Hamas diverting some but not most of the aid? They're trying to shoot them, not starve them.
#13. As of an hour ago, Reuters is reporting also that the AI state regulation provision is stripped:
https://www.reuters.com/legal/government/us-senate-strikes-ai-regulation-ban-trump-megabill-2025-07-01/
And Bloomberg:
https://www.bloomberg.com/news/articles/2025-07-01/senate-removes-tax-bill-provision-limiting-state-ai-regulation?embedded-checkout=true
Kind of surprised this isn't getting more coverage!
It just passed the Senate 51–50, with Vance breaking the tie. I expect it'll start getting coverage now that it has to pass the House again, since many prominent people highlighted this is reason they'd vote it down.
#3 - I think this is a good test between people who truly understand physics, biology, chemistry, etc, and people who simply memorize facts.
If a chess master is burning as much chemical heat energy as a professional athlete, then where does all the energy go? Dry skin in a room temperature environment is unable to radiate heat at a 6000 calorie/day burn rate. Chess players would have to sweat like basketball players if the claim were true.
I don't think this is necessarily true, because there's a lot of variability in how concentrated heat-generating activity can be to achieve a rate of 6000 calories/day. Basal metabolic rate is roughly 100 calories per hour. Working out in a cool room, I sweat heavily at caloric expenditures of 800 calories per hour, but not at 400 per hour. A chess tournament can last much longer than a basketball game, and with the activity distributed enough, the participants would not necessarily become particularly sweaty. Also, my understanding is that some chess players do become quite sweaty at tournaments, if not as sweaty as basketball players. Personally, I tend to sweat when I concentrate intensely, so I'd be surprised if that weren't the case for at least some chess players.
The issue I think is harder to square is that chess players don't seem to *breathe* as heavily as is associated with even modest cardio exercise. They might sweat heavily due to stress, but if they're engaging in increased respiration, they'd have to take in more oxygen.
"The issue I think is harder to square is that chess players don't seem to *breathe* as heavily as is associated with even modest cardio exercise. They might sweat heavily due to stress, but if they're engaging in increased respiration, they'd have to take in more oxygen."
Did you actually check out the link? The whole thing started with someone observing rapid breathing from chess players in tournaments (for short periods of time), and someone else extrapolating (poorly) on how many calories they could burn with that much extra oxygen.
To be slightly more precise they report the chest movement rate (from impedance pneumography) and interpret it as breathing rate- since your chest can move for reasons other than breathing and they do not specify a method for filtering the data it is not clear that the maximum signal they reported should even be considered a breathing rate in the first place
Exactly. Weirdly on-point anecdote but my ex had taken Sapolsky's class at Stanford and casually repeated this claim to me many years ago. I have a physics degree and my immediate reaction was "no f-ing way, show me the evidence" and made exactly that waste-heat argument.
Ha! Thank you for this comment. I didn't think to explicitly search through Sapolsky's lectures, but sure enough, it's in lecture 1 of the recorded version of his Introduction to Human Behavioral Biology (albeit just "thousands", not 6000). Note that this video has over 18 million views.
https://www.youtube.com/watch?v=NNnIGh9g6fA&t=2262s
> You have two humans, two individuals, who are going through a ritual. They are sitting at a table. They're absolutely silent. They're making no eye contact, and they do nothing more physically taxing than every now and then, one of them picks up their hand and moves a little piece of wood on the table. And if these happen to be the right two individuals in the middle of a chess grandmaster tournament, these people are maintaining blood pressure for six hours running that you only see in a marathon runner. **These people are going through thousands of calories a day doing nothing more than thinking. And this is outrageous.** Because you look at one of these chess grandmasters who's just taken down an opponent, took down their queen or whatever, and they will have the exact same physiology as some male baboon on the Savannah who's just ripped the stomach open of his worst rival. And we're doing it there just with thought.
#50 - the decline seems to have begun in 2022: https://www.fbi.gov/news/press-releases/fbi-releases-2022-crime-in-the-nation-statistics
My city experienced a 61% increase in homicide in 2016 over 2015. I looked up to see what the explanations then given were.
They said they didn’t know.
Covid wasn’t ready to hand as an answer.
#39 I wrote about the relation between FLFP and (higher) GDP in Latin America a couple of years ago, if anyone is interested.
"For Latin American countries it appears to be the case that the higher average income is, the less women feel compelled to work."
https://www.mangosorbananas.com/p/a-comparison-of-gender-equality-between
#13 Note that the huge income tax "cuts" in the BBB is mainly just maintaining the current tax rates! No one actually gets lower taxes than they currently pay.
This counts as a tax cut because the original bill only lowered taxes for the next 10 years, when rates would go back up.
This is done to make the official forecasts look good, but the intention was to keep the new rates in the future, which is now.
I'm not saying this is good (or bad) policy, but it's important to understand that taxes largely stay the same, even though it is in a technical sense a tax cut.
Aren't the "no tax on tips" and "no tax on overtime" new?
They are. The higher SALT cap is also a real tax cut.
These are comparatively small changes though.
Yes, though those don't favor the rich (the SALT changes do though).
Thanks for posting this. I came to learn only this week that the 'tax cuts' which people have spilled a lot of ink over are 'the same tax policy we've had for the majority of the last decade'. Maybe I'm dumb but this feels like an underdiscussed point.
It's still important though because, when the original tax cuts were passed, the entire deficit/impact forecast and subsequent conversation at _that_ point was under the assumption that would expire the way that they were meant to. This is the bait and switch that is continually getting played (by both sides). Get what you want now, minimize the projected impact by giving it an expiration date in order to reduce public backlash and criticism, and then, when it comes time to actually let it expire, convince everyone that "oh this is just the status quo, it's not that big a deal to let it continue and cancel the expiration". So now, we are at the point when the lie of the expiration is finally being made explicit. So, relative to what was promised when they got passed, this _is_ additional tax cuts.
If you want to argue from the perspective that this is nothing new and just status quo, then I hope that the next time someone proposes something with such an expiration, you demand a forecast assuming it won't expire, and take your view based on that forecast.
I am, if anything, inclined to penalize this even more than usual specifically to try and discourage this bait and switch tactic.
I don't disagree that the TCJA's original proponents made their arguments in bad faith, what has been weird for me is the framing. I feel like both sides are selling this as a major tax cut, with the republicans hoping that people will like that and the democrats hoping that people won't like that because it includes/favors rich people. But, unless you are on medicare or frequent a medical provider with a high percentage of medicare beneficiaries, nothing about your life is likely to change if this bill passes. Like, I don't understand how the bill's authors expect it to win them in any votes in the midterm, given that it makes no-one's life better and some people's worse than the status quo. Democrats can and are pointing to people losing out from benefits being cut, but they will not be able to point to a rich CEO making more money from the bill, because the CEOs will be paying the same as they were in 2019.
If the republicans were not so committed to donor enrichment they could have just let the 199A deduction expire, not touched benefits, and said "SEE DEMOCRATS WANT TO RAISE YOUR TAXES" and the democrats would have had to get an argument with the ref, both in the style you've gestured at. But no one has done either of those things, so the rhetoric hasn't matched the situation and I don't see how it will pay off for them a year from now.
10: The explanation for this that I've always heard is that sometime during Trump's first term, social media algorithms changed to start pushing the manosphere much harder to teenage boys, realizing it was a good way to get them really hooked into the social media systems. I haven't done any research into whether or not this is true, that's just what I've heard.
33: "I guess this is good (the verdict wasn’t dependent on a few ignorant or dishonest people), but maybe also bad (shouldn’t being familiar with the best evidence for one side or the other make you believe that side more?)" Isn't this what you should expect from a healthy scientific field? The experts in a subject matter who don't know the details of a specific topic still know whose opinion they should trust, and thus end up converging on the truth? This sort of shortcut is exactly how science is supposed to work, right?
#20 Wonderful timing. The Roman Catholic lectionary cycles through much of the Old Testament every few years, and today's Old Testament reading features the destruction of Sodom and Gomorrah!
#1 - The way that the author identifies the relatively small set of emotive words the AIs consistently alight on (ghost, whisper, echo, buzz, hum, pulse, etc.) reminded me of a common observation about another poorly-trained artificially intelligent storytelling entity, the modern Walt Disney Corporation. It's an in-joke in theme park fandom that Disney only knows about a dozen words (dream, fantasy, wish, enchantment, adventure, magic, etc.) and combines them in various ways to serve every possible purpose.
Other corporate entities surely have their own favored vocabularies (synergy, sustainability, difference, etc.) just as AIs bent toward nonfiction do, but something about narrowing the domain to the production of fiction makes them even easier to surface.
Hey, if Taco Bell can do it with beef, cheese, and tortillas, Disney should get to do it too.
I for one embrace the budget entertainment that is adventure tortillas.
It makes me think of the "start to Clench" races for Stephen Donaldson books, where everyone opens a different book to a random page and sees how long it takes until the word "clench" is used.
(In trying to find that article again, I've discovered Google lists no blogs or articles called 'Rickety Fiction', which seems like a fun enough name that SOMEONE would have used it.)
Re #3: Man, that's really disappointing about Sapolsky. I really respect him and though he was a responsible scientist, but this is pretty irresponsible. Now I will have to take every claim from him under suspicion. I wonder how many of the stories in A Primate's Memoir were embellished or even made up.
I just want to add that one thing I really like about Scott is that he is not a bullshiter. I really dislike bullshiters. I think Scott is about as honest as people come, and I really respect that.
Thanks. I think relatively few people are bullshitters, and it's more of a question of how careful your back-of-the-envelope calculations are. It looks like Sapolsky's issue was that someone said chessmasters breathed three times as fast (at one point), and he seems to have assumed breathing was a perfect correlate for metabolic rate so you could multiply daily metabolic expenditure by 3x. I can imagine making this mistake if I filed this fact away in my head as "something something 3x something metabolism chess" and then remembered it while writing an essay about something else and rolled a critical failure while trying to recall it.
I have probably done things approximately this stupid; when I don't, I'm protected less by careful source tracking, and more by a good enough mental model of how metabolism works that if I felt tempted to say something like this, I would think "no, that's crazy, I know that an active brain doesn't consume that many more calories than an inactive brain" and then check my source.
I don't think Sapolsky deserves too much disapproval for having gotten the calculation wrong. I think he deserves a lot of disapproval for, once it became clear that it was wrong, going around falsely telling people that it was a journalistic misrepresentation of what he said, instead of admitting that he made a mistake. I think that you would not have done that.
Also condemn him for blaming the original source.
I think bullshitting is a spectrum, and I think almost everyone does it to a certain extent. I also think that there is a continuum between genuine mistakes and intentional bullshitting. The question then is, how likely are you to make a mistake about a fact or an inference from given facts, based on your commitment to truth and your love of bullshitting? In this case, and of course this is speculation, but as an intellectual, would it not be seductively attractive that thinking burns as many calories as athletic competition? Someone committed to truth would be wise to stop and take a step back, when such a seductive fact presents itself. But if you are prone to exaggeration, you will dismiss such thoughts.
Perhaps I am being too harsh, and I am sorry about that. But even in the most generous interpretation, this reveals a carelessness that I find off putting in a scientist. I still like Sapolsky and will continue to enjoy his books. I will just approach his writing with more skepticism, which I suppose is not a bad thing.
https://m.youtube.com/watch?v=VzykUGonAyg
17) "I’d always heard modernism was a reaction to the trauma of the World Wars," No, it was a reaction to the glutinous over-reach of late 19C art. The more every new building is ornately Gothic, the more you're likely to dream of clean Bauhaus lines merely as a refresher. The problem is that then the reaction went even more overboard than what it was reacting to.
Useful to remember that Virginia Woolf wrote, "On or about December 1910, human nature changed."
56) I also hated dogs as a child, still do. And not for any psychological displacement reasons. Dogs attack strangers, they bark, they bite, they knock you over. Nothing used on the child in this story would have worked on me at all. Isn't there a reason for signs reading "Beware of dog"? You don't see any signs reading "Beware of cat" or "Beware of goldfish."
36 & 44: the gaza aid distribution failure seems like a great example of POSIWID
You'll have to spell this out so I can see how intensely I disagree with you. If you mean "even though the people who distribute the aid really wanted it to go well, I am going to say the 'purpose' of the aid was to have it go badly, to express my disapproval of the process and my belief that good intentions aren't enough", I continue to think that's just confusing matters. See https://www.astralcodexten.com/p/come-on-obviously-the-purpose-of and https://www.astralcodexten.com/p/highlights-from-the-comments-on-posiwid
I'm not Big Swan, but I assume they meant to imply that the (radical?) left's support of Palestine is mostly about "decolonization", i.e, that they're fully on board with terrorism against Israel, support 10/7, etc. And that this has nothing to do with the recent occupation of Gaza by Israel.
ETA: to be clear, I'm merely relaying claims I've heard from rightists about the supposed true motives of leftists.
I do think it's a vibe based question (IMO most teenagers are going to answer surveys based on vibes and will probably have a higher lizardman constant), but there are definitely enough people to hit 20% on that question. The replies on that link are completely full of people arguing the premise - IE "if women did the same work they should get paid the same, but women don't do work as well as men" and things like that. The age they're surveying is particularly susceptible as well, I personally remember being pretty shitty at that age and growing out of it in college with exposure to more people and becoming friends with more women.
13. Trump's BBB "lowers taxes (especially on the rich)"
1. Most of the so called lowering consists of making permanent provisions that were enacted in 2017 and were scheduled to expire later this year. I have been a tax professional for 50 years, during that time there have been major revisions to the tax law in almost every administration. Any thing that is in effect for 8 years should be viewed as permanent. Taxes are imposed on a dynamic economy. We do not know what the effect of most changes are until years after they are put in place. A real reform would be leaving things the same for long stretches.
2. It is a tax on income. People who do not have much income do not pay much taxes. The standard deduction for a married couple filing jointly is not $29,200. Any decrease is mostly going to be a decrease for those who are better off.
3. The US has had an income tax system for more than a century now. it has not affected the distribution of wealth or income in any meaningful way. It probably cannot. The distribution is determined by factors that are not related to taxation. The wealthiest people in the country, Musk, Bezos, Gates, Ellison, Page, Brin, all made their money by exploiting a new technology. In the gilded age it was railroads, now its the internet. The real impact of taxes is imposed by the forces of supply and demand. Any tax that on a growing segment will simply be passed along as a consumer cost. Taxes should be designed to collect revenue cheaply and and easily. Getting the most feathers from the goose wit the least amount of hissing.
#10: Coincidentally, Matt Yglesias wrote about this today: https://www.slowboring.com/p/progressive-millennials-are-old-now
It's unfortunately paywalled, but here's my summary of his argument: It's a cohort effect. Most millennials came of age in the late 2000s or early 2010s, which was when the coolness of progressivism peaked. The conservative establishment had three huge high-profile political failures in short order (the Global War on Terror, the Great Recession, and same-sex marriage), which created a sense that the right was just bad and wrong about everything. The election of Barack Obama was the most exciting thing in ages and put tons of cultural momentum on the progressive side. And the cultural hegemony of scoldy Christian moralists was in just the right phase of its decline: recent enough that millennials remembered it as the uncool establishment to rebel against, but too advanced to push back effectively against that rebellion.
Now, as Yglesias's title says, those progressive millennials are old and uncool. The new cohort that's now coming of age is doing so in an environment where the left hasn't been doing so hot, and grew up being scolded by *woke* authority figures and so thinks of wokeness as the moralistic establishment that's cool to rebel against.
Someone in the comments linked https://slatestarcodex.com/2014/04/22/right-is-the-new-left/, which is now on Scott's Mistakes page, and made a point that I hadn't heard before: Maybe Scott wasn't wrong, maybe he was just early.
Also Democrats used to benefit from the decreasing share of the white vote, but that's stopped working as the right has gotten better at appealing to racial minorities. I'm not sure to what extent it makes sense to think of that as the same trend, though.
Your summary of the social media consensus is wrong (to be clear, that's because the reporting and tweeting on this by the authors was misleading). I'm not sure what the plot is from (not the preprint), but the study did a Delphi process and it looks like this is the stage 1, not the final vague statements that they had consensus on (see third link).
I wrote about how the consensus sausage got made:
https://rubenarslan.github.io/posts/2025-05-20-consensus-how-the-sausage-gets-made/
Some SM scholars explained why they didn't engage in this process:
https://www.science.org/content/article/social-media-consensus-paper-causes-social-media-uproar
A representation of the final consensus, although the first author of the preprint (Capraro) says the best representation is the table in the preprint.
https://bsky.app/profile/ruben.the100.ci/post/3lprsihzzn22l
#10: Although the woke movement peaked in 2022, specific lines of woke argument peaked long before then. One of these is the alleged 23% gender-based pay gap. Although it was debunked by mainstream outlets in 2012, the myth persisted among feminists for *years*. I remember seeing new videos and articles decrying the pay gap in 2017, and perhaps as late as 2019.
For some data points about the history of this debate, see section 6 on this page:
https://becauseits2015.wordpress.com/2016/08/06/a-non-feminist-faq/#power2
My first hypothesis was that some boys took the "reversed stupidity" response to the pay gap claim. Feminists claimed "Women should not earn 23% less than men", and although many boys inverted the claim correctly ("Women *do not* earn 23% less than men"), some of them inverted it incorrectly ("Women *should* earn 23% less than men). This explanation seems plausible in principle, but I think the timing is wrong.
My next hypothesis is that it's related to the MeToo movement, which began in 2017. The movement may have contributed to the proliferation of workshops and other programs about sexual consent. (Presumably, these would have began 1-2 years later.) Many of the boys would have felt like they were being blamed, profiled, or punished for bad things that they didn't do.
In addition to stoking a general feeling of resentment in the boys, I suspect that some of the boys also responded with solidarity toward adults who actually did mistreat women. Both of these psychological factors could contribute toward the development of sexist attitudes.
My bet is on MeToo. It peaked in 2018 and I believe exhausted many men's tolerance for modern feminism. For me personally I know that's when I switched from considering it a little nutty to outright adversarial and in bad faith.
13. What I find interesting about this is that it was largely the religious right that drove this over concerns about pornography and child protection. It's also a good reminder that one reason prediction is hard is that success or failure can come from basically exogenous factors to the narrow area you're examining.
16. Decker is not as smart as he thinks he is. No bank would lend on the promise of a political party, in part because the Democrats can neither credibly promise to win the next election nor that they will pass that bill. You could calculate the risk discount rate but it would be huge. You'd need to discount that they'd win the presidency and both houses which, even if they're individually favored in all three, would net out to a lower than half chance. Then you'd need to discount they actually fulfilled the promise.
Giving some fairly favorable assumptions:
P(Presidency win) = 60%
P(House win) = 55%
P(Senate win) = 55%
P(Passing the bill even with control) = 70%
You get about 87% discount rate. That is insanely high. That's 4x the discount rate that VC gets for investing in companies that don't exist yet. That's higher than the discount rate given to Zimbabwe during its crisis.
That math isn't right because it assumes that the three electoral outcomes are uncorrelated, and because it represents the probability of *any given* election resulting in unified Democratic control, without accounting for their getting another chance every two years.
The math is simplified for sure. If you want to assume they're positively correlated and produce a P(Unified Gov) that does increase your chances but it still ends up at like 40%. Which produces a discount rate of like 73%. Which is still gigantic. And that's in a good year, if they have a less than 50% chance of winning it drops.
They get a new chance every four years, not two, because they need the presidency. But you're right that's the biggest simplification: you'd need to calculate the time value. You get an expected wait of 3.7 cycles through that, rounding to 16 years. Multiply them together and you get 89% even with the higher correlated number.
I'm not sure I'm following your discount-rate math. If (we assume for the sake of argument) the banks will have to wait 16 years in expectation to get paid back, then can't you just apply the normal interest rate to figure out how much they have to charge? The concern would be the possibility that they never get paid back at all, and I'm not sure how your back-of-the-envelope calculation purports to calculate that.
No, because the bank is not trying to give the universities the lowest possible interest rate. It needs to take into account default risk, political risk, time horizon uncertainty, etc. This means the cost of capital converges to the risk adjusted rate, not the prime interest rate.
What I think you're kind of feeling is the universities have other, cheaper sources of capital already. Which is true.
Somewhat that, but also I'm literally just trying to understand your math; I don't work with financial math that frequently. When you say "89% discount rate", does that mean that a lender would have to charge 669% APR to break even on expectation? If not, what does it mean? And can you show how that number is derived from the probability of unified government per election?
36. While I'm sympathetic to the idea that experts tend to know things, I'm also sympathetic to the idea that experts have a vested interest in finding whatever replaced them failed. There's no mention of the reason the GHF concentrated in a few checkpoints: because they were under military guard after Hamas supporters attacked those more distributed checkpoints.
He also does not mention the reason the UNRWA and other aid organizations were able to disperse aid checkpoints is because they paid off Hamas. Now, maybe he just doesn't think this is worth mentioning, but I think he knows the optics of "we paid Hamas so they would allow us to feed starving Gazans" would harm already waning support for UNRWA. And so on and so forth. This is a narrow selection of facts that veer into speculation, memetically fit but not a mature discussion of the situation.
48. The Gulf War was fought over land though which meant that the US could sweep in hammer after the degredation including actual, hard fighting. China's hammer would need to go over a rough sea. There's also the problem that the Gulf War, which is widely regarded as rapid, took six months. That's plenty of time for even a relatively unprepared America, Japan, etc to get itself together and intervene. And once they decide to intervene China has no option but a slugging match with no clear diplomatic exist. Of course, "a slugging match with no clear diplomatic exit" is also how I would have described Ukraine.
Another important fact is that neither Ukraine nor Iraq could retaliate against the home territory of the attacking country. Taiwan absolutely can.
Sure, maybe China can hope for the victory to be so quick, and the Taiwanese military so rational and dispassionate, that they just negotiate a dignified surrender rather than ravaging Xiamen in defiance. But it's an additional risk they're taking compared to Russia and the US.
55. China has had various institutes for understanding and controlling the effects of AI. It was seen as a natural extension of their control of other forms of technology and information production. I would bet the point of making an AISI that's modeled on the western ones is not to signal a new commitment to AI safety but to provide a Chinese voice in such conversations abroad. They do this a lot, having an internal version that controls domestic affairs and an external version that's meant to push the Chinese point of view in international organizations or create commitments China itself is not bound by.
A key tell: can you get commitments from actual Chinese regulators and are they sending over top party officials? Do they actually make laws and regulations that are enforced internally? Are they allowed to advocate for it openly inside China in Mandarin? Or are they just sending over people who are mostly specialized in dealing with foreigners with little domestic power? So far the CnAISDA style outreach has been the latter.
That said, I stand by my comment China invests more into AI alignment research than anyone else in the world. They just want to align it with the CCP.
Compare the image manipulation retraction of the Sodom & Gomorrah paper to the image manipulation retraction of the Alzheimer's amyloid beta paper. The archaeology paper was published in 2021 and immediately had concerns about repetitive elements. The Alzheimer's paper also had concerns in 2021, but was published in 2006. And my understanding is that no one looked at it merely because it was an influential paper, widely known to be unreplicable garbage, maybe widely believed to be fraud, but only because they were doing general sweeps, having discovered > 1% image fraud rate.
Why didn't the Alzheimer's paper get any attention to its images? Because no one thought it would be easy to prove fraud? Or because no one thought proving fraud was worth the personal career risk of being a whistle-blower? If image manipulation fraud is so common, why aren't there any whistle-blowers testifying to having witnessed it? Do the people doing systematic studies of fraud have a secret origin story of witnessing fraud but don't want to throw their mentors under the bus?
22. So it’s sorta like Cameo, but for dorks like us?
23. I’m surprised it’s only 32 players. I thought the general practice was to take any player of prominence and slap his name on a random athletic shoe design.
#22: Where are you finding the list of paid reviewers on here? I clicked around a little and couldn't.
#12 they already signed a contract with the military...
#1 - Is the quick and easy answer for why AI writes like that simply because the training data for fiction writing is overwhelmingly things like tumblr posts, mediocre mass-market stuff, and, perhaps most significantly, fan fiction? When I was reading the AI output it reminded me most of all of cringe-inducing tumblr posts and that time in high school (I've fruitlessly tried to erase from memory) when I went to see what all the fanfic fuss was about.
Aren't there lots of traditionally published books and journalism and stuff in the training corpus? Also, I don't see why fanfic in particular would be more than a small percentage of it, most user-generated text on the public internet isn't fanfic.
I'm not familiar enough with what's in the AI corpus, nor do I understand how much training on non-fiction (most of the internet, I assume) informs an AI's fiction stylings. But a very quick search indicates that fanfic, specifically, numbers in the lower 8-digits, and the number of traditionally-published fiction books is in the high 5-digits yearly. How many of those are available for AI training? I have no clue.
Self-publishing (mostly e-books), which probably has some overlap with fanfics, as I assume people often publish to both fanfic sites and create e-books of their works, numbers roughly 3-4x traditional publishing. I'd hazard to guess most of that, in terms of quality, is closer to the fanfic average than the traditional average. But, of course, the traditional publishing average is probably closer to the fanfic average than it is to "great" writing.
But it seems obvious to me that it will be extremely hard or impossible to get AI to produce great fiction, given that it's something of an aggregation machine. Nearly all humans can't write great fiction either!
That said, the particular stylistic quirks shown in the article resemble a particular style of internet writing, at least to my eyes. I'm not confident in this argument at all, as you can probably tell.
I'm not terribly familiar with Cremieux's take on the lead/crime hypothesis -- I think he has mostly argued that cohort studies are heavily confounded by race and class, so you can't just compare violence rates in high blood lead neighborhoods against low blood lead neighborhoods.
Meta-analysis of such studies and their confounders is difficult, but there are many other reasons why the lead/crime hypothesis fails to be some kind of unified theory for the rise and fall of crime across the western world in the second half of the 20th century:
https://medium.com/p/949e6fc2b0dc
In short, the lead/crime hypothesis fails for many other countries: Japan, Hong Kong, Mexico, Brazil, etc. The lead/crime hypothesis also fails in the US, if you divide the data up by age groups -- the rise in violence in the 1960's was a period effect, not a cohort effect, as you'd expect if this were an issue of subsequent generations getting lead poisoned in youth.
The rough synchronization of crime trends across western countries is fascinating, but may ultimately have simpler explanations. Demographics is one of them -- these countries all had a post war baby boom and then a decline as birth control became more widely available in the 60's/70's. A proper analysis would, at the very least, age adjust the data before just noticing the bump in the murders per capita graph.
There does remain a plausible theory that blood lead levels are one small factor in the bigger picture of crime rates, and that leaded gasoline did raise crime somewhat (10-20%?), all else equal. A sufficiently careful analysis of cohort studies and their confounders might be able to find whether this effect exists, and how large it is. But that's a hard thing to tease apart from the data and separate from other variables that did have an effect on the 80's crime wave and subsequent decline. Other strong factors may include: demographic changes, crack cocaine, mass incarceration, changes in policing, and medical changes that reduced the lethality of gun shots:
https://medium.com/@tgof137/explaining-the-crime-wave-of-the-1980s-d98395133dfc
Wow, wasn't expecting to see myself here! Thank you ACX, you're all wonderful.
I am confused about the karma point chart, I don't see it anywhere in the linked paper.
"Related: Cube_Flipper on his (?) experience taking estrogen. “What did change was my sense of space. This one’s quite subtle – it was the kind of thing that was more noticeable when I experimented with deliberately spiking my hormones. I’ll do my best to explain. It’s as if I took the entire volumetric representation of the space around me and increased the degree to which every point within that could influence the location of every other point, recursively. This allows everything to elastically settle into a more harmonious equilibrium. This effect is basically identical to what a small dose of psychedelics can do, specifically a tryptamine like psilocybin or DMT.”"
Might be an absolutely wild thing to say but I've been thinking this for a while, a big difference between women and men is that women are basically constantly microdosing on LSD/psychedlics. Thats there experience of the world
17. My pet theory is that the "decline of rhyming poetry" is largely an illusion caused by the rise of sound recording and broadcast media. There is still a ton of rhyming poetry being written, and much of it is extremely popular and commercially successful, but we call it "song lyrics" rather than "poetry" because it comes to us as audio/video recordings of the poem being performed set to music. Even for genres where the lyrics are often spoken rhythmically than sung (certain subgenres of folk, blues, and rap), we still think of it as music rather than poetry.
Before phonographs and radio became widespread, lyrics were commonly composed and published without any particular music being intended, and people could read them as poems or set them to whatever tune they liked that fit the meter of the poem. And poetic forms were standardized enough that you had a large menu of tunes to choose from for any given poem. Especially for "common meter" or "ballad meter", which is both very specific and very widespread, and is why you can sing the lyrics of the Gilligan's Island theme song to the tune of Amazing Grace and vice versa. One major and very popular form of poetry/music in the 19th century was "broadsheet ballads", poems written in ballad meter and published in broadsheet format (basically a single sheet of a full-sized newspaper).
Copy pasting my comment about #3 here
"Chess Grandmasters Do Not Burn 6000 Calories Per Day"
There is a reddit post with the exact same title from a deleted account, is it your post? Also, more generally how good is substack for search engine optimization. I dont think its very good, I cant find your post on google, limiting the value of correcting this misninfo. I've had this problem with other substacks as well
17: "the decline of rhyming poetry"
I think it depends on how you define poetry. If you limit yourself to the academic journals, maybe yes. But, I think that is snobbery. I think the great mass of poetry these days comes from in the form of lyrics and rap, both of which are deemed to be musical genres.
Didn't the august body that awards the Nobel Prize for literature concede my point by giving the laurel to Bob Dylan? He characterized himself as a poet when young:
Though you might hear laughing, spinning,
swinging madly across the sun.
It's not aimed at anyone.
it's just escaping on the run.
And before the sky there are no fences facing
And if you hear vague traces of skipping reels of rhyme
To your tambourine in time
It's just a ragged clown behind
I wouldn't pay it any mind
It's just a shadow you're seeing,
that he's chasing
Rap of course rhymes. I detest it because of its coarseness and vulgarity, but it does often rhyme.
The thing about musical lyrics is that they have an unfair advantage. There's something about hearing words set to a tune you like that changes how you process them. For instance in song, phrases or sentences are often repeated, sometimes over and over with no other phrases intervening, and not only is that not tedious, it is often very powerful. For instance I love hearing the Talking Heads' sing "take me to the river, drop me in the water" over and over. Or there's the end of a hallelujah chorus -- "forever and ever" & "hallelujuh" over and over. But it's not just repetition we experience differently in song. Cliches are much more tolerable -- can sound heartfelt and obviously true -- .and so are phrases that don't make sense. I've often looked up the lyrics of songs I love and felt taken aback by how lame they are with the tune stripped away.
Re Metaverse:
Ahem: AHAHAHAHHAHAHAHAHAHHAHAHAHAHAHAHAHAGAHSHAHAHAHAHAHAHAHAHAHAHAHAHA (etc)
"this post helped me crystallize some thoughts on what “good writing” and “good taste” are in general."
Care to elaborate? Would love to hear your thoughts on this.
Allan Shivers did not "beat himself" in the election. His vote share from each party line was pooled together to make his total vote share. It wasn't like there was a competition over whether he gets elected as a Democrat or as a Republican. It's also not unprecedented for candidates to be nominated by both major parties. e.g. Earl Warren won both major party's nominations for CA Governor in 1946. Robert Heinlein wrote a book about his experiences in politics, in which he complained about how he had very narrowly lost an election in which everyone told him they wouldn't bother helping him because he had no chance; I fact-checked this and discovered that he had in fact narrowly lost the Democratic primary to the Republican incumbant. In New York, candidates still can get listed on multiple parties' ballot lines, and it is common for the Democratic and Working Families parties to nominate the same candidates. It's no longer common for both major parties to endorse the same candidate, but Simcha Felder did get both major parties' endorsements for his NY State Senate run in 2024.
Until the Top 2 Primary ballot measure took effect (in 2012, IIRC), California still had a weak back-door ballot fusion system. You could only be a listed candidate on the primary ballot for one party, but you could register as a write-in candidate for other party primaries for the same office. I don't remember if you got separate ballot lines if you won more than one nomination, like in NYC, or if you had to choose one (so the benefit would just be less competition on the general election ballot), or if you got one line with both parties listed on it.
I know about this because in 2010, I was peripherally involved in a Congressional campaign in a safely Democratic district where there was a relatively strong Libertarian candidate and no qualified candidates for the Republican nomination. The Libertarian, Edward Gonzales, tried to use the process to also get the Republican nomination. It didn't work: the local Republican party organization recruited and endorsed a different write-in candidate who won the primary by a large margin, and the Democratic incumbent (Zoe Lofgren) was reelected by an enormous margin (69% to 24% to 7.8%) in the general election.
#10 (fewer boys believe in gender equality): the explanation seems clear to me, 2018 was the year #MeToo peaked. Weinstein got arrested, the Golden Globes was a public MeToo rally with celebrities wearing black, the phrase "believe women" gained currency, and the Kavanaugh hearings happened. It pushed a certain subset of men to conclude that the 'equality' movement was just rhetorical cover for a zero-sum political gender war. So "gender equality" got (appropriately, in my view) re-coded as "anti-man" by many men and so they stopped supporting it.
46. Very interesting. This offers a lesson about prior restraint reverberating through the ages. José Clemente Orozco famously self-sabotaged in the same fashion when painting his tremendous mural "Prometheus" (aka "the hunk without the junk"), located in Pomona College's Frary dining hall.
When I was a student, I heard that Prometheus' penis had been removed or obscured by a third party. It was understood to have been due to puritanical inclinations, but one can't dismiss considerations of appetites might have played a role.
At any rate, I haven't thought of Prometheus in a while, so I looked it up and was surprised to learn that Orozco had actually omitted the titanic genitals himself!
However, seemingly filled with regret, "He attempted to add one when he visited Pomona several months after initially completing the mural, but it did not adhere properly to the wall."
Per the Wikipedia article here: [https://en.wikipedia.org/wiki/Prometheus_(Orozco)], based on a lengthier piece in Pomona College today: [https://magazine.pomona.edu/2012/summer/orozco-at-the-border/]
Edit: the original link was almost number 47, which would have mystically tied it back into Pomona College lore. Alas, so close.
#43: So portuguese have been pretty cool about their colonialism? Good catholics who know about humility. At least, by a glance, germans seem to score better than english and french. I wonder what african view on china may look like a century from now.
#35 - I'm impressed by the problem solving strategy the LLM is using. It seems like it's either a novel strategy that others aren't using, or it's using one of the most advanced existing strategies. Are there other domains where LLMs are coming up with novel ways to go about solving a problem? Right now it seems like LLMs serve as quickly-responding encyclopedias of known problem-solving strategies. Are there domains where LLMs are producing novel problem-solving strategies without expert prompt engineering? If/when this starts happening, I wonder which domains it will happen in first and how it will spread across domains/topics.
48: "T. Greer on China’s strategy for invading Taiwan."
The Chinese strategy described sounds just like the Russian strategy for invading Ukraine. The Ukrainians are a lot more like Taiwan than the Houthis, who are tribesmen wandering around a desert.
But the all time champions of fighting off invading armies of larger countries are the Israelis. And that is who the Taiwanese should model themselves after, if they are serious about staving off the CCR regime. Of course, it would need to triple its spending, start producing its own armaments, and institute mandatory universal military training and reserve service.
But, if they do something like that, their prospects aren't that bad. The Island is 380 km long and only 130 km wide, but, the eastern two thirds of the country is quite mountainous, The mountains can be a shelter from attack and a defensive redoubt. Given time and money the Taiwanese can defend themselves. they have the money. The questions are whether they have the time and the will.
I object that Bentham got "Eulered" there. If you're going to try to make an argument that is so heavily mathematical in nature, it is entirely correct to point out any issues with the math. Math provides a high level of certainty (or at least, the impression such) *due to the unrelenting demand of rigorous argument.* The issues that Dylan Black raises are pretty core to the use of probability on infinite sets. If you don't understand them, you probably shouldn't be making arguments (especially in favor of something as complex as God!) that rely on the mathematics of probability theory.
#28. Jeez, the Yudkowsky-Soares book needs a catchier title! Anyone have suggestions? Here are a few of mine.
DEATHCODE!
When AI codes Its Own Genius, It's Code Blue for Humanity
------------
THE DIRTNAP IS NIGH
-------------
HUMANITY IS CIRCLING THE DRAIN . . .
and AI Holds the Plunger
43.
The few North African/Muslim countries included are more down on former colonial powers than Black/Non-Muslim countries. That tracks.
Separate from the merits of question being argued over, #41 makes me wonder why to take any of Cremieux's takes seriously.
That seems like a pretty big overreaction. I updated a small amount away from Cremieux, but:
- That chart wasn't the only bit of evidence Cremieux cited.
- I don't even know what Higney could have meant when he wrote "For some reason [Cremiuex] cut off the bottom of that chart" - Higney posted a different chart! (https://x.com/AnthonyHigney/status/1867163809267507628) If there's room for legitimate disagreement about the relative value of PCCs and elasticity, the entire basis for Lexer's (and Higney's) criticism vanishes in a puff of smoke. As an outsider, I asked some LLMs about it and Claude said "If lead exposure leads to small PCCs but high elasticities, it suggests the relationship might be nonlinear, confounded, or noisy, leading to weak correlations." If that's correct, it's not necessarily diabolical to downplay elasticities, particularly if you think they are due to confounders.
- The interpersonal drama Lexer brought up seems completely irrelevant to me. But at best it's a draw, in my book. Higney called Cremieux "a bad faith actor", which is a rather damning thing to say about someone, imo, so even though I don't fully understand Cremieux responding by calling Higney "surly, reactive, and untrustworthy", it seems in-bounds given the level of interaction they were having.
#10 is unsurprising; male socialization for teenage boys is even more fucked up than it usually is on the interpersonal level (source: anecdotes and vibes) and there was a big organic push from anti-woke influencers over the past X years which the party and the money put their shoulders behind, from hoi-polloi types like Rogan and Walsh to intelegencia blogs like this one.
This seems like the result one would expect if they knew 12 years ago that the anti-woke side of the culture war would be making gains.
#8 I prototyped a video game with some friends when I was in college where you would earn karma by performing good deeds and lose karma by performing evil deeds, and whenever your character dies she'd reincarnate as a different animal, with different abilities, according to the current karma level. The backstory was that the protagonist was a member of an order of monks grinding good deeds to attain enough karma to ascend to a mythical creature after reincarnation, but was tricked into deicide, the most heinous crime, by a demon king, and lost enough karma to be on the threshold between humans and monkeys. But since she got the power to immediately reincarnate and retain her memories she's the only one who can use the powers of all the creatures to resurrect the godess she had killed and get the karma back to ascend to the higher entity and defeat the demon king.
The gameplay loop was full of contrived trolley-problem-style situations where you'd gain or lose karma and then kill yourself (or allow yourself to get killed) to reincarnate as whatever creature the platforming situation called for.
If I had known at the time about this Ming dynasty guide to karma bounties, I would have used it verbatim, this is hilarious.
> 30: Twitter and Bluesky both seem to be losing users...
On X, I've noticed that lately I'm being swamped with posts from obnoxious rightwing crazies. I block and block, but the X algorithm keeps throwing them at me.
OTOH, everyone on Bluesky is just too fecking nice. I need a little more drama to keep me engaged, but not toooo much drama!
I miss the old Twitter where was a nice balance between interesting people and assholes.
And my Facebook feed is filled with AI-generated groups that swamp the posts from my friends. I'm wondering if there's any future for social media.
Try the substack feed. It's like Twitter-lite. There are way fewer users but they're generally higher quality.
#6 Cate Hall's conclusion strikes me as very strange. She were deficient in testosterone for her own sex. Once the levels were brought up to what was typical for her sex, she did much better.
Men don't automatically feel 10 times sharper and more energetic because they have 10 times the testosterone. They feel similarly sluggish if their T is low for their biology (which would still be a very high number for a woman in most cases).
I think the conclusion here is "having a hormonal profile that's out of whack as compared to what's normal for your sex will have negative effects", not "men run life on easy mode".
> Re 45: Conception beliefs among Australian aborigines. Did you know that pre-contact aborigines didn’t know that sex caused conception?
It wasn't until the 1920s that Western medicine had a complete and testable hypothesis about the role of ovulation and the necessity of fertilization with sperm at the correct time in the menstrual cycle for conception to occur. And that hypothesis wasn't tested and confirmed until a decade later.
Although 19th-century scientists had come to the conclusion that sperm was somehow necessary for conception, in the 18th century, there were angry debates between the Spermists and Ovists. Neither side would concede that both were necessary for conception. So, it's no surprise that pre-Enlightenment societies didn't understand conception.
From the chapter you mentioned...
> By contrast Arnold van Gennep, the eminent French folklorist, considered it to be little different from the ignorance of procreative mechanisms still prevailing among the masses of Europe.
> Nuno Sempere (maybe with the rest of the Samotsvety team?) has launched askaforecaster.com
Nice! Yes, confirming that some Samotsvety people will help me answer if there is too much demand. In the meantime someone is asking good questions for $150:
> Change in non-engineering product-related roles in tech by 2030?
>
> Current AI tools are augmenting work for engineers at many technology companies. The same thing is happening for non-engineering product-related roles as well -- for instance, Product Designers are having AI included in design tools to create interface mock-ups faster; Product Managers are using LLMs to write drafts of briefs and memos; etc. Presumably, as these tools improve employee efficiency, companies will need to hire fewer and fewer employees to get the same amount of output.
>
> Specifically for these non-engineering product-related roles (which, you may have guessed, I fall into), I wanted a quick forecasting take on how employer demand for these roles may fall in the next 5 years. Is it more likely to expect a modest 10% fewer positions? Is a 30-50% re
duction possible? An industry-shattering 80-90%? And what factors might make a position more or less resilient?
>
> 2025-07-01 15:40:56.188382,PRICE150
9: Meta has spent $50 - 70 billion on the metaverse over the past five years.
How do you spend that much money on, essentially, a Second Life clone?
9: Meta has spent $50 - 70 billion on the metaverse over the past five years.
How do you spend that much money on, essentially, a Second Life clone?