Linkpost for February

Effective altruism, life advice, detailed reality, fun.

Effective Altruism

Global Poverty

286,000 women died last year from pregnancy or childbirth. If maternal mortality rates were the same as they were in the vast majority of history, 1.2 million women would have died. If all countries had the same maternal mortality rate as Europe, only 11,000 women would die each year from pregnancy or childbirth.

Conditional cash transfers and unconditional cash transfers both reduce poverty. However, conditional cash transfers lead to better health and educational outcomes, probably because people are directly incentivized to send their kids to school or perform health behaviors.

GiveWell has made its largest single grant yet— $96.3 million to the Against Malaria Foundation. You can see the grant in GiveWell’s grants database: as of the time of writing, GiveWell has not put up their grant writeup. Very cool!

GiveWell from A to Z.

USAID is the test run for Trump’s plans for cutting other agencies: freezing funding, getting rid of people, and causing fear.

A heartbreaking story about the costs of cutting USAID. A doctor divides therapeutic food (given to malnourished children) into thirds to try to make it last, while already purchased therapeutic food sits in USAID warehouses no one can open.

Reflections on ordo amoris, moral demandingness, and foreign aid.

Animal Advocacy

The most recent EggTrack report, which tracks compliance with cage-free egg pledges, has been released. The headline news: 40% of U.S. eggs are now cage-free! :D

Anima International chose to give up its successful campaign to make institutions like schools and universities serve more vegetarian food, upon realizing that replacing meat with eggs potentially increased animal suffering. Anima currently works on cage-free egg campaigns. Good job, Anima!

Effective animal advocacy charity Sinergia Animal allegedly claims credit for animal welfare commitments that don’t exist or that already existed before its advocacy. Extremely concerning allegations; I await Sinergia Animal’s promised response.

The federally funded U.S. Meat Animal Research Center conducts research into increasing farmed animal productivity with a shocking disregard for animal welfare or research ethics. Even animal farmers are opposed to many of its programs—such as a breeding program intended to get cows to produce more twins that caused nearly 1 in 6 calves to die.

Particularly Good: A trialogue about what animals are moral patients.

Emerging Technologies and Existential Risk

Trends in the AI industry make it far more difficult for researchers to figure out if AIs are moral patients. Increased secrecy means that researchers have to look at behavioral outputs instead of being able to understand how AIs think. New AI architectures may be adopted so quickly that researchers don’t have time to figure out whether they’re moral patients or not. We might soon develop AI agents that are superhumanly good at AI research, but we have no reason to believe they’d be superhumanly good at moral philosophy (AI research has “one right answer” which makes reinforcement learning much easier)—new developments would outpace our understanding of AI moral patienthood.

According to standard economic theory, it’s unclear whether AIs/robots being able to do everything more cheaply than humans can would increase or decrease wages. Most arguments about this subject don’t address the main important issues according to economic theory.

Meta Effective Altruism

I continue to work my way through Joe Carlsmith’s blog archive; he’s rapidly becoming one of my favorite writers. A few I particularly enjoyed last month include: Problems of evil, about atheist spirituality and the project of developing an attitude towards Reality Itself; On sincerity, talking about the virtue of not playing pretend with yourself; Care and demandingness, on whether it is reasonable to expect morality to be nondemanding given that non-moral goals can be nearly arbitrarily demanding.

Other Causes

Five days under Trump. Trump is really doing an overwhelming number of things. I have been trying to keep my priorities straight: I care only about preserving American democracy and PEPFAR.

Trump’s plans for the FBI fit in well with the agency’s current culture: illicitly spying on peaceful members of minority religious and political groups; missing major terror attacks; violating civil liberties; and maintaining dossiers on people without reasonable suspicion of any crime.

Literally the most valid DOGE supporter. “I wanted to have one form "wizard" that would allow a veteran to enter their information once, and automatically apply for all the benefits for which they were eligible. OIRA told me that to do this, I would first have to submit every possible permutation of this wizard for approval—a request I would have found delicious to comply with, were there enough trucks on the planet to deliver that amount of paper.”

Prediction markets underperform conventional forecasting because people invest in them as an expression of identity rather than as rational bets on the future.

Life Advice

Mid-conditional love: “if you love me unconditionally, presumably you love everything else as well, since it is only conditions that separate me from the worms.” Therefore, we should think about loves in accordance with how many conditions there are on them.

House rules for Baba Is You that help you practice making plans in situations where you’re very confused about what’s going on.

Reality Has A Surprising Amount of Detail

The January 2025 jobs report showed a spike in the labor force and employment, which wasn’t real at all and was caused by adjustments to be more in line with the Census’s population estimates. Even though the increase in the Census population is almost entirely immigrants, because of how the jobs report is calculated, it will look like an increase in native-born employment. I love this kind of weedsy statistics stuff.

Lungless salamanders! No one knows why they stopped having lungs!

Particularly Good: The history of relationships between writers and editors at The New Yorker.

Discourse

Oliver D. Smith is a poorly hinged individual who targets intelligence researchers through SLAPP suits and by using sockpuppets to edit their Wikipedia and RationalWiki pages. Smith even managed to get an article criticizing him removed from the Wayback Machine; fortunately, the article still exists on archive.li. What strikes me about this is that everyone knows Smith is a bad actor—he’s been banned from both Wikipedia and RationalWiki. And yet he was still able to run successful cancellation campaigns, including against prominent and completely unobjectionable paleoartist Emily Willoughby.

Feminism went mainstream in part because of very positive coverage in women’s magazines.

The core beliefs of tobacco harm reduction advocates: most people are misinformed about tobacco; nicotine isn’t very dangerous; products that reduce the risks of tobacco are the best way to reduce smoking; tobacco consumers should be more involved in tobacco policy; it’s better to give people choices than to coerce them.

Fun

Chinese Doom Scroll: Fines for going over or under word count on your mandatory book report on the CEO’s speech (shades of the Snow Crash toilet paper memo). America will never succeed at AI because English is a shitty clunky language that doesn’t have a word for ‘undershirt.’ Second-language Chinese speaker makes an accidental BDSM innuendo.

Particularly Good: This close reading of Goodnight Moon, which for such a popular book is remarkably uncanny.

Particularly Good: Snow plow names. My particular favorite is “Ice Fought the Thaw and the Thaw One.”

The mathematical constant of pi is playing Pokemon.

“The exponential dependence of resistivity on temperature in germanium is found to be a great big lie. My careful theoretical modeling and painstaking experimentation reveal 1) that my equipment is crap, as are all the available texts on the subject and 2) that this whole exercise was a complete waste of my time.”