Linkpost for April

Effective altruism, law, economics, science, rationality, short stories and poetry, fun.

Effective Altruism

New bipartisan legislation will likely increase the effectiveness of USAID. It mandates the existence of Development Innovation Ventures, which experiments to discover which programs are the best. It requires USAID to hire innovation fellows, whose job is to make sure the agency will adopt programs that DIV found were successful. USAID has chronically insufficient resources to decide how to spend their money, so the hope is that the innovation fellows would make effective aid programs the path of least resistance.

Advocating for evidence-based tobacco control policies has relatively little money spent on it for the high benefit; the effects are also relatively easy to measure. However, cigarette smoking is declining regardless. It’s relatively expensive to change policies, because of tobacco industry lobbying. The countries that don’t have tobacco control policies yet are generally more difficult to affect and also smaller. Encouraging e-cigarette use may be neglected, but the effects of this intervention are highly uncertain.

StrongMinds is probably less cost-effective than GiveWell top charities. The Happier Lives Institute overestimates spillover effects on household members, the duration of the effect, and the effect of depression treatment on life satisfaction. They also fail to properly adjust for social desirability bias, publication bias, and programs working less well in the real world than they do in trials. The Happier Lives Institute also has beliefs about to what extent improving someone’s life is better than averting a death that GiveWell thinks don’t match the beliefs of its donors or recipients.

A groundbreaking paper showed that building more primary schools increased income in adulthood. A careful reanalysis suggests that there is no detectable effect. One problem is wage scale dilation. People in richer areas, which would already have primary schools, earn more money than people in poorer areas. Because the wages in high-paying jobs increase over time, and the wages in low-paying jobs don’t, you see a larger old-young income gap in areas where fewer primary schools were built.

Access to Sesame Street makes children more likely to be on grade-level, especially if they’re poor, black, or male.

Giving people cash immediately before natural disasters can help them bounce back from the disasters.

Electronic payments didn’t work well for Haiti, because SIM cards are difficult to distribute and Internet access and cellphone coverage are limited. Many corrupt government officials steal SIM cards with aid money on them and give them to family, friends, or supporters.

The best ways to encourage people to eat less meat are social media or blog posts, classroom education, news articles, and meat-free challenges.

Lapsed vegans and vegetarians named motivation, support from loved ones, easier recipes, and more plant-based food options as the key blockers keeping them from returning to their diets. The most common obstacle they faced while vegan or vegetarian was food dissatisfaction, including cravings, feeling hungry, and boredom. Older lapsed vegans and vegetarians had a more satisfying experience with being vegan or vegetarian than younger ones.

Most consumers believe that humane labels on animal products indicate a higher level of animal welfare than they actually do.

The factory farming of chicken began when, by mistake, a farmer received 500 chickens instead of the 50 she ordered and decided to raise them all anyway.

The plant-based meat industry may be in a decline.

Alternatives to rodenticides are expensive, inconvenient, hard to use, and less effective. Rodenticides are the best way to deal with public health emergencies and severe infestations that pose a risk to occupant health. Reducing rodenticide use is likely to be difficult and expensive.

Longtermism is surprisingly irrelevant to whether we should prevent human extinction. Standard cost-benefit analyses suggest investing far more in it than we already are.

AI companies should prioritize alignment research, security, and safety standards, while avoiding unnecessary hype or speeding up AI research. They should also prepare for weird situations in the future, such as by planning for the future or setting themselves up as public-benefit corporations.

Protest movements lead to a small net-positive change in public opinion and public policy. They also lead to a moderately sized increase in how much attention people are paying to the issue. Protests vary wildly in effectiveness. The best protests are large; are nonviolent; have unusual participants like children; gain media exposure; are costly to politicians or other people whose behavior you want to effect; and convey a clear frame for how the issue should be discussed.

A lovely set of personal essays from a Singaporean EA about how her background reflects her experiences in effective altruism

A former employee writes about his experience at Alameda Research and discusses what we could have known about Sam Bankman-Fried when.

April Fools post about Dustin Moskovitz wrapping up OpenPhil.

Law

The “reasonable person standard” can be hard to implement given how subculturally diverse America is.

David Friedman gives ideas for better assessment of law schools and how to make them cheaper.

A black writer for a police procedural talks about the ethics of his job.

The Oakland Police Department covered up a hit-and-run by a sergeant so they wouldn’t have to fire him.

Heartwarming: the Alliance Defending Freedom, Americans for Prosperity, and Project Veritas all file amicus briefs defending a left-wing citizen journalist who focuses on police misconduct.

Affirmative action did not cause a police shooting.

Economics

Ireland pilots a basic income program for artists.

Related: the stories of four people who participated in an unconditional cash transfer pilot.

Reducing the regulation on specific medical devices leads to more innovation, lower prices, and in some cases even more safe products (because you can more easily sue manufacturers of less stringently regulated medical devices).

An explainer about the FairTax proposal.

Fascinating article about why we should build more housing and buildings in downtown Berkeley. I absolutely agree that Shattuck has gotten cooler while I’ve been living here, and eagerly anticipate more cool businesses.

Science

Most scientific fields have a suspicious pattern of p-values that points to p-hacking. Economics is by far the best; the runners up are ecology and environmental science; psychology and sociology; and zoology. The most suspicious fields are complementary and alternative medicine; plant biology; chemistry and geology; and informatics, mathematics, and physics.

In humans, epigenetics makes your bone marrow cells behave differently from your liver cells. It does not pass down intergenerational trauma.

It’s not clear that lab animals or wild animals have been getting fatter. Differences in obesity rates across US states are probably explicable by demographics and lower atmospheric pressure at higher altitudes.

Learning styles theory is wrong. People can’t be reliably divided into the categories of visual, auditory, or kinesthetic learners. Even if they could, this just reflects a preference; many people learn best in the way they don’t prefer to learn. Nearly all information is best conveyed in a particular way: you wouldn’t want a purely auditory explanation of the geography of Africa, a purely kinesthetic explanation of how to pronounce Spanish words, or a purely visual explanation of how to play tennis. This effect dominates over the effect of people’s preferences.

Rationality

Hosting events with less stress: host events you find easy to host; plan the environment well, then get out of the way; keep track of what works; practice!

It’s useful to distinguish motivation (what causes your action), volition (whether something feels intentional), and control (whether you expect things to turn out as you meant them to).

The origin of everything from species to ideas to technologies is that they evolved through a process of natural selection.

Fully polyamorous relationships and fully monogamous relationships are equally happy. However, most poly people are slightly poly, which tends to lead to worse outcomes; people who aren’t embedded in a poly community are more likely to be slightly poly. As such, people with bad relationships are overrepresented in many people’s anecdotes about polyamory.

Superforecasters are most useful when trying to figure out exactly how likely a pretty likely event is. Regular people are incapable of distinguishing probabilities above 75%, unlike superforecasters. However, superforecasters have a high variability in accuracy between questions, which limits the usefulness of their predictions.

Short Stories and Poetry

Mandragora: A Harry Potter fanfiction about sympathetic people missing ongoing moral horrors.

The Daemon Lover: Creepy, unsettling short story by Shirley Jackson of The Lottery fame.

Godzilla’s Twelve Step Program: Godzilla is in recovery from stomping on buildings.

Uncleftish Beholding: The first chapter of a chemistry textbook from a universe where English only has Anglo-Saxon roots.

Different Kinds of Darkness and comp.basilisk faq: Two excellent short stories by the same author set in a world where certain images cause the human brain to shortcircuit.

The secret history of Wakanda: The history of the legendary land of Wakanda, from Pliny the Elder to modern-day conspiracy theorists.

Horror Story: An uncanny take on creepypasta-style haunted house stories.

Degustation: Body horror about autistic mushroom people. Content warning for consensual cannibalism.

All Watched Over By Machines Of Loving Grace: Lovely poem about transhumanism.

Fun

Aella gives the results of her Chaos Survey.

A very fun rant about why it would make you unhappy to win the lottery.

Actually, artificial ice was extremely popular in the 19th century; the term was not an insult.

The implosion of a queer banking startup.

The Fujoshi Guide To Web Development. It’s real. Git wears a maid outfit.

If you read one link in this post read this one: an extremely detailed review of TikTok YA novel Lightlark, which is one of the stupidest YA books to come out in a while.