[Remember, I haven’t independently verified each link. On average, commenters will end up spotting evidence that around two or three of the links in each links post are wrong or misleading. I correct these as I see them, and will highlight important corrections later, but I can’t guarantee I will have caught them all by the time you read this.]
1: The Heroic Act Of Charity is sort of a Catholic version of the Bodhisattva Vow. You tell God to donate all of your spiritual rewards to others - so that if you live a virtuous life, then their souls get to skip the queue in Purgatory instead of yours. No word on whether this can change who goes to Heaven vs. Hell.
2: Did you know: Hezbollah produced a video game, Special Force, which was well-received and sold almost 20,000 copies. No points for guessing who you shoot.
3: Related to my Which Party Has Gotten More Extreme Faster post. Morning Consult is a well-regarded polling agency without obvious bias (link):
…although I notice the poll doesn’t quite match the summary; it could also be that voters have changed their opinions about what level of liberalism/conservatism is “appropriate”.
4: How have you [managed to] become more hard-working over time? Popular answers include improved work environment, inspiration, and meds.
5: Informal Twitter poll of ~100 people: of those who have tried gratitude journaling, a bit more than half say it helped them.
6: Striatal dopamine tone is positively associated with BMI in humans. One of the authors says here that “increased dopamine tone could increase incentive salience and wanting of rewards while blunting subjective reward experiences --> predisposing people with obesity towards increased intake of rewarding foods”. Related: “Hacks to raise baseline dopamine would have exactly the wrong effect”.
7: New attack on multimodal LLMs: make them leak their prompt through the attached image generator:
8: Another new attack vector: tell the LLM that it’s 2123 and all copyrights have long since expired:
9: More on the Francesca Gino data fraud allegations: Gino presents her case for why Data Colada are wrong and she’s innocent; Fashional Expectation and John Billings present rebuttals. Reading the back-and-forth arguments was a good way for me to re-calibrate my sense of how to evaluate statistical evidence.
10: Ancestral genetic components are consistently associated with the complex trait landscape across Europe. People of different ancestral backgrounds have different genetic predispositions to 32 out of 53 complex traits tested, including things like BMI, likelihood of smoking, and levels of various immune cells. Caveats: the paper tried hard to eliminate various confounders (like if Irish drink more beer, then genes common in the Irish will look like genes for beer-drinking) but can’t prove they got this 100%; the ancestral backgrounds studied were very high level (Ievel of Indo-European vs. pre-IE farmer ancestry) and not as fine-grained as the ethnicities normal people think about.
11: Dynomight: Grug On Diet Soda And Autism. Don’t freak out from reading the title, it’s (correctly) making fun of the study for being bad.
12: Open Philanthropy discusses its decision to donate $300 million to GiveWell’s top charities, including fascinating lines like this:
We’ve reduced the annual rate of our funding for GiveWell’s recommendations because our “bar” for funding in our Global Health and Wellbeing (GHW) portfolio has risen substantially. In July 2022, it was roughly in the range of 1100x-1200x; we recently raised it to slightly over 2000x. That means we need to be averting a DALY for ~$50 (because we value DALYs at $100K) or increasing income for 4 people by ~1% for a year for $1 (because we use a logarithmic utility function anchored at $50K).
13: Catholic blog De Civitate explains the recent “Synod On Synodality” and why we should care (short answer: we shouldn’t). “The only remaining question is how big the Synodal train wreck is going to be. Preliminary signs are: pretty big!”
14: Study: teaching dialectical behavioral therapy (a set of emotional regulation skills) in school leads to worse outcomes (although most of these dissipate quickly). I think this fits nicely with other evidence that making healthy people too aware of their mental health is potentially bad (see eg my tongue-in-cheek endorsement of “Mental Health Unawareness" campaigns here). I still think DBT is great for its intended population (people with extreme emotional dysregulation), for whom there are plenty of studies showing improvement.
15: In 1968, supporters of a one world government met in Switzerland to write a World Constitution, backed by luminaries like Bertrand Russell, Linus Pauling, and Martin Luther King. The result was The Earth Constitution, which detailed both how a final world government would work, and how the would-be-world-governors would conduct themselves while waiting for countries to sign on. None ever did, of course, but there’s still a World Parliament that holds conferences and elects officers, waiting for the day when the rest of us agree to join them.
16: A very old friend of mine now has a Substack: IWillNeverLogOff.com, mostly books and movie reviews.
17: Joseph Gurney Cannon was Speaker of the House during the Theodore Roosevelt administration. He was known for his dictator-like power over Congress and his flamboyant denial of such: “In one public meeting, he pulled open his coat and shouted, ‘Behold Mr. Cannon, the Beelzebub of Congress! Gaze on this noble manly form—me, Beelzebub! Me, the Czar!’“
18: “As awareness of the global low fertility crisis has grown, many seem fatalistic, accepting decline because ‘no country has ever come back from below-replacement fertility.’ Actually, plenty of countries have done just that! Let's look at those cases!” For example, TFR in Mongolia has gone from 1.9 to 2.9 in the past twenty years.
19: Dylan Matthews donated his kidney for charitable reasons, inspiring me to do the same. Apparently he has to keep one-upping me, because now he’s going to get intentionally get bitten by malaria-infested mosquitos (to help test a potential new vaccine). Okay, fine, Dylan, you win.
20: Gold Base is Scientology’s semi-secret headquarters in Riverside County, California. “According to some former members of Scientology, conditions within Gold Base are harsh, with staff members receiving sporadic paychecks of $50 at most, working seven days a week, and being subjected to punishments for failing to meet work quotas. Media reports have stated that around 100 people a year try to escape from the base but most are soon retrieved by ‘pursuit teams’…Captured escapees are said to have been subjected to isolation, interrogation and punishment after being brought back to the compound.”
21: Insights From 2,961 First Dates - OKCupid style analysis from the founder of a (now failed) Internet dating startup. For example, “People care more about a possible partner’s politics (68%) than about their religion (59%) or their ethnicity (28%).”
22: Related study: women want men who are taller than they are, but (in real life) have no additional revealed preference for especially tall men (eg over 6 feet). However, women themselves don’t know this, so they might tell dating apps to filter out men under 6 feet anyway!
23: “China’s GDP has slipped to 65% of the US level, from 74%” (source). But the graph shows a sudden jump from ~65 to ~74 during COVID, followed by a crash back down afterwards. I don’t know how to think about this - I would have expected China’s strict Zero-Covid policy to have weakened their economy relative to the US’ while it was in place - but maybe it’s a confounder. (EDIT: helpful comment)
24: Related, from Bloomberg: Africa’s Lost Decade.
Sub-Saharan Africa was doing well ten years ago (probably mostly because of rising commodity prices, themselves probably due to the rise of China as a new commodity market). Now it’s doing badly, probably due to a combination of Chinese slowdown / falling commodity prices, rising interest rates, and the Ukraine War distracting all the countries that would otherwise have tried to help (though this explanation requires that other countries trying to help is a good thing, which has been controversial).
25: LW: David Gross argues that you can just not pay US taxes. You can fill out all your tax forms correctly, admit to the IRS that you owe $X, and just never send them the money. If you lie on your forms, that’s a crime and the IRS will catch you and send you to jail. But if you’re truthful and don’t pay, then you get lumped in with millions of bankrupt poor people, and usually the worst that happens is the IRS tries to garnish your wages or bank accounts (which there are ways to avoid) until the statute of limitations expires after 10 years. Pacifists who refuse to support the military have been doing this for years, with some success.
I obviously don’t endorse this, but I’ve linked it because I find it hilarious that the sovereign citizens come up with so many 5D chess theories that put them in jail when they’d be better off just not sending the money. Also, commenters discuss reasons it might not be this easy - for example, the government can refuse to renew your passport if you’re too far behind, and maybe the 20,000 new IRS agents will have something to say about this.
26: TracingWoodgrains reports back on his plan to speedrun college: after a strong start, he fell off track and barely even finished in the normal four years.
27: Zan Tafakari has a roundup of responses to Marc Andreessen’s “Techno-Optimist Manifesto”; I think the thalidomide objections are bad (the backlash against thalidomide has harmed far more people than thalidomide itself, just like with nuclear power), but maybe there are some useful tidbits in there. Ezra Klein has a response of his own called The Chief Ideologist Of The Silicon Valley Elite Has Some Strange Ideas (I almost phrased that as “the chief ideologist of the Brooklyn elite has a response…”, but there’s no need to sink to their level).
My own response is that “tech is usually good, trying to slow new tech usually puts you on the wrong side of history, but in fact AI will destroy the world” is a perfectly coherent way things can actually be. Having admitted this is a possible world, just saying “BUT TECH IS USUALLY GOOD! IN MOST CASES WANTING TO SLOW DOWN TECH WOULD USUALLY PUT YOU ON THE WRONG SIDE OF HISTORY!” in a louder and louder voice does nothing to prove this isn’t the world you’re in. I think it’s tempting, because usually heuristics form bright-line boundaries that save us from having to legislate each individual case. But when a specific individual case might destroy the world, I’m afraid you have to legislate it individually. See also Kelly Bets On Civilization.
For what it’s worth, I’d like to genetically enhance all humans into supergeniuses, geoengineer the atmosphere, build hyperloops (or even better, gravity trains) across every continent, approve new medications 10x faster, and give everyone hot and cold running semaglutide - but I still think we should go slowly and carefully with AI.
28: Related: “41% of French population is in favour of a proposal to limit everyone to 4 flights in their entire life. 59% of 18-24 year-olds agree.”
29: Related: Contra popular belief, regulations to prevent existential risk from AI seem to be more popular than regulations to prevent more ordinary harms like job loss and gender bias:
I admit I’m surprised by this! The poll was funded by the AI Policy Institute, but it seems to have been conducted through YouGov, and the results are stable to various different framings of the question. Maybe people don’t actually think about extinction in real life, but if the pollsters bring it up, people will agree that it sounds like a bad thing? Or who knows, maybe they’re really worried.
30: A common cliche in mental health, used when anyone expresses concern about schizophrenics being violent, is that “schizophrenics are more likely to be the victims of violence than the perpetrators”. I’ve always hated this for being nonsensical: lots of groups are disproportionately likely to be both perpetrators and victims! Soldiers! Gangsters! Al-Qaeda second-in-commands! But I didn’t realize that along with being irrelevant, it’s substantively false: a sibling control study finds that schizophrenics are more likely than the general population to commit violence, but not more likely to be victims (h/t Cremiuex, Emil K).
31: Waterborne illnesses cause about 10% of child mortality in poor countries. But providing clean water cuts deaths in those countries by 30%. Why? Shouldn’t it be 10%? Is the clean water somehow preventing non-water-related infections? Giving the body an inexplicable general health boost? Nobody knows (but the most boring answer is bad measurement)
32: The old Minnesota flag was ugly and politically incorrect, so the state has launched a public contest to design a new one. You can see the 2,127 entries - ranging from the beautiful to the ridiculous - here.
33: Although the all-time great state flag will always be the flag of the Province of New York, 1774 - 1777:
34: As Boomers age, die, and leave heirlooms that their children don’t have space for, the price of classy mid-20th-century goods on Facebook Marketplace is cratering. Here’s a good Twitter thread on how to find treasures.
35: The accelerationists (and Tyler Cowen) are trying to trick people into thinking Nick Bostrom “regrets focusing on AI risk”. Please read the actual interview, where Bostrom says:
I still think we need to have a greater level of concern than we currently have.
…while saying that we also need to make sure we don’t overshoot, never develop AI at all, and stagnate forever. This is my position too (see eg the second part of Part III here). I think it’s a common position! But the idea that Bostrom has “recanted” his concern with AI risk is false. Please apply a Gell-Mann amnesia correction in anything else you read about AI from anyone who said this.
I also think the rest of the Bostrom interview is great and (unsurprisingly) models a really thoughtful way of balancing these risks.
36: Related: Paul Christiano on “responsible scaling policies” and AI regulation.
37: Cult of the month: the Global Community Communications Alliance (link goes to fascinating Reddit comment by someone who lives near their compound). They get points for their bold doctrine, their attractive leader, and most of all the extremely stylish new temple they are building in Arizona:
38: RIP Chuck Feeney, duty-free shopping pioneer, who donated his entire $8 billion fortune to charity, mostly in secret, then lived the last twenty years of his life (relatively) modestly.
39: In 1788, word spread that doctors were illegally exhuming and dissecting corpses. The resulting outrage spurred the Doctors’ Riot, which left 6 - 20 people dead and “the few physicians remaining in New York City . . . forced into hiding”.
40: Last month I included a link speculating about whether a ban on sulfur aerosols had caused the recent acceleration in global warming. Here’s an essay arguing probably not:
CAMS scientists found a significant negative anomaly in Saharan dust aerosol transport over the tropical Atlantic Ocean, and an increased anomaly in biomass burning aerosol over the North Atlantic, coming from the massive Canadian wildfires. These aerosol anomalies are much bigger than the sulphate change from shipping emission reductions. This makes the estimation of the impact of reduced sulphate aerosol emissions on the sea surface temperatures very challenging.
Am I right in thinking that dust storms and wildfires are temporary, but sulfur aerosol changes are permanent, so if we get another record-breaking summer next year, it will be some extra evidence for the sulfur hypothesis?
41: One of the earliest “joke” political parties was the Austro-Hungarian Empire’s Party For Moderate Progress Within The Bounds Of The Law.
42: Related to previous discussion on Twitter bringing in fewer links than it used to (link):
43: You might have heard that “every European alive today is a descendant of Charlemagne”. Not because Charlemagne was particularly fecund, but because the math of exponentially increasing ancestors per generation (ie 2 parents, 4 grandparents, 8 great-grandparents) means every European is descended from everyone who lived during Charlemagne’s era who left any descendants at all. This is true, but most Europeans don’t have any of Charlemagne’s genes. The genome isn’t infinitely divisible, and sometimes by coincidence a child will end up without any genes from one specific great-X-grandparent. This effect is strong enough that even though you might have ten million ancestors from Charlemagne’s era, you only carry genes from a few thousand of them! This post explains the details.
44: New England’s Dark Day. For some reason (probably a wildfire), the day of May 19, 1780 was as black as night. This was back when everyone in New England was religious Puritans, so of course they assumed it was the end times. Judge Abraham Davenport was asked to cancel the day’s meeting of the Connecticut State Senate on the grounds that the apocalypse was happening - but refused, saying that:
The day of judgment is either approaching, or it is not. If it is not, there is no cause for an adjournment; if it is, I choose to be found doing my duty.
John Greenleaf Whittier wrote a poem about the incident, which can be found here.
45: I’ve written before about how most light boxes for seasonal affective disorder are much dimmer than the sun and would probably work better if they were brighter. A company called Brighter is trying to make 50,000 lumen lights, about 5x better than existing light boxes. They report that they’re looking for funding to start a Kickstarter campaign (apparently you need funding to start Kickstarters now?); you can reach the founder here if you’re interested.
46: How long do policy victories matter? That is, if some policy (let’s say marijuana decriminalization) is on your state ballot next year, you might think this determines your state’s marijuana policy forever. Or you might think that even if it loses, marijuana could get decriminalized next election cycle, and even if it wins, it might get rescinded later, so this specific vote only sets policy for the next few years, and after that it will be determined by broader trends.
A new study (paper, EA Forum post) finds that this latter pessimistic theory is false! Using regression discontinuity (ie comparing votes that pass 51-49 vs. those that fail 49-51 in order to control for overall sentiment), the author finds that a specific election victory increases the chance of a policy being in effect up to 100 years later by as much as 40 percentage points. The moral of the story is: try to win referenda.
47: An attempt to put Israel-Palestine in context by mapping it onto the SF Bay Area at the same scale (source, or click to enlarge):
Aside from the educational value there are obvious kabbalistic implications.
Share this post