Linkpost for August
Effective altruism, social justice, psychology, life advice, tech, fun.
Ozy Elsewhere
Over on the fiction substack: I have a new short story, Mea Maxima Culpa, in which gay space Catholics argue about absolution. (Mind the warnings.) I have decided to post occasional blog posts over there in the hopes that reminding people to buy my book would be less annoying if a blog post is attached, so we have me being unhinged about John Wick, a free version of the previously paywalled post Sad Showtunes Shit, and the Hard Humanities Fiction manifesto.
Speaking of me and fiction: hey! New people who came from Scott linking to me a bunch! Have you heard I have a novella about a girl in a troubled teen camp acausally dating a girl in a space imperial Chinese harem? I recommend reading it if you would like to cry about lesbians!
Alexei Cifrese has kindly recorded audio versions of my blog posts Against Blanchardianism and Bad Therapy Review: Privilege. Check it out!
Effective Altruism
Global Poverty
What Open Philanthropy’s Global Health and Wellbeing grantees accomplished in 2024. Progress towards a strep A vaccine! Lead elimination in developing countries! Technology that lets us image the brain and spinal cord at the same time and thus improve our understanding of chronic pain! Fish welfare inclusion in seafood sustainability standards! Replication games!
The Center for Effective Aid Policy has shut down. Convincing policymakers to use aid money more effectively is hard. Much of a country’s aid budget has to be allocated a certain way, because of binding commitments and political realities (e.g. the popularity of aid to Ukraine). Nonprofits that successfully get one of the remaining slivers of the already tiny budget are really really good at getting grants, and just coming in and saying “we know how to allocate money effectively!” won’t work. I applaud the cofounders for realizing that what they were ing wasn’t working and wish them luck in their next endeavors.
Nigerian researchers are developing a technology that uses CRISPR to detect epidemics quickly.
Humane-certified sugar often comes from fields where workers were coerced into hysterectomies. I really like how this article emphasizes that humane certification is really hard. Bonsucro, the sugar certification organization, started in Latin America, which has relatively large farms; everyone agrees its efforts in Latin America hold farmers accountable. But India has very small sugar farms, and it’s nearly impossible for a thirty-person organization to check that all of the thousands of farms that supply a mill are behaving ethically.
Animal Welfare
Predictors of temptation to eat meat among conflicted vegetarians and vegans: positive associations with meat in general; remembering their positive associations with meat when they have internal conflict; general ambivalence about eating meat; responding to temptation the previous day by downplaying the positive aspects of meat, instead of using some other strategy.
Cities could adopt policies that would promote wild-animal welfare, such as requiring bird-friendly windows that birds wouldn’t run into, planting trees that provide food and shelter, and avoiding habitat fragmentation.
Actionable: Wild Animal Initiative, which focuses on building a robust academic field of welfare biology, will no longer be funded by the Open Philanthropy Project, which previously provided half of its funding. This is a real opportunity for small donors to make a big difference for wild animals.
Existential Risk
An economic model of when robots will take your job. Lots of good stuff, but one thing I found interesting is that not all jobs have to be automated for a “wage collapse” scenario. Instead, a wage-collapse scenario happens when there’s more demand for work either machines or humans can do than for work that only humans can do.
Leading AI labs fudge their benchmarks to make it look like their models are better than they actually are.
Many AI risk arguments depend on “value fragility”— the idea that if you optimize for a value system slightly different from human values, you end up getting something that humans find valueless. But it’s unclear how intensely a superintelligence would have to optimize to make value fragility true, how much a superintelligence would optimize in the real world, and whether it would be possible to program a superintelligence not to be ambitious or to be nice to other value systems.
Actionable: Arkose provides support for mid-career machine learning professionals considering getting into AI safety.
Other
Countries don’t differ very much in terms of how much people care about human babies, human children, human adults, various kinds of animals, Siri, trees, or rocks; they do differ in terms of how much people care about embryos. People in all countries value human babies and human children more than human adults. Across cultures, people weight sentience highly when deciding how much to care about an entity, but don’t care much about agency or skills in social cognition.
Public transit expansion means you can build housing more densely without making traffic more congested. A major problem is that American public transit construction is absurdly expensive compared to anywhere else in the developed world. Public transit advocacy could be cost-effective if you think YIMBY activism is cost-effective.
Meta Effective Altruism
“Warren Buffett pledges $100 billion for nothing in particular”. Instead of giving money to the global poor, Buffett’s new charitable trust, the largest in the world, will make decisions based on the consensus of his offspring, none of whom seem to care about helping those most in need.
Principles of truthseeking: don’t look for an “epistemic daddy” who will save you from having to do your own thinking. Create information. Don’t make it harder for others to think clearly. “Contact with reality should (mostly) feel good.” Share information publicly. It’s a red flag if you don’t believe anything strange to your social circle. Be willing to hurt people’s feelings (but not unnecessarily).
Politics
Calling Trump and Vance “weird” is a powerful attack line because it ties them to the chaos of the 2010s while signalling that the Democrats intend to be boring and normal going forward. “You guys tried to overthrow the government,” while true, also implies that it’s going to be nonstop apocalyptic rhetoric. I like this analysis, but I think a lot of it is just that Trump is marketing himself as a rebellious badass and “you are a threat to democracy” fits in much better with his image than “lol cringe.”1
Inside Trump’s anti-vaccine beliefs.
In general, the executive branch tends to exempt some companies from new or raised tariffs. Normally, this process involves a lot of technical analysis (could the company buy the product from an American company?) and politicization (how good are the company’s lobbyists?). Trump is absolutely going to give exemptions to companies that stay at his hotels.
In Bullshit Jobs, David Graeber argues that “rich people create fake jobs in order to have an impressive-looking number of economic dependents”. But the example bullshit jobs he gives are usually at least one of: explicable with standard economic theory like rentseeking; ordinary profit-maximizing jobs that Graeber thinks hurts society; or jobs that seem kind of fake because they’re new and distant from value-creating activities but that actually do make our lives better (like marketers or app designers).
Social Justice
Particularly Good: An explainer of which Indians immigrate to the United States and how they’re different from non-immigrants.
I really liked this close-reading of the influencer Ballerina Farm, a Mormon homesteader and former Juilliard-trained ballerina. It doesn’t focus on what we can figure out about Ballerina Farm (after all, we don’t know her) but what her popularity says about gendered expectations, the appeal of the traditional life, and the concept of “getting back to nature.” Also there’s a tangent about the history of stoves!
Psychology
In perhaps the least surprising turn of events imaginable, MDMA therapy poses risks of psychiatric abuse. Your cool new radical alternative to therapy as usual? Has the same risks as therapy as usual!
Antidepressants actually do help prevent depression. The small apparent effect in randomized controlled trials is partially because not all antidepressants work for all people. Even so, antidepressants consistently show an effect similar to standard depression treatments.
It’s meaningful to say that depression “causes” the traits it’s correlated with, even though depression is ‘just’ a way to refer to a cluster of symptoms that tend to occur together. We can use “depression” to refer to either the description or the hypothesized cause. Further, in English, “cause” is often used to refer to the cluster something’s in: “depression causes irritability” can mean “this irritability is part of the depression cluster and not the mania cluster.”
Being diagnosed with borderline personality disorder can cause people to be discriminated against and denied needed care—and many patients find it a helpful way of understanding their own experiences.
Elementary school students still go to lessons if lessons are optional. (The essay features a cameo from my kid!)
Life Advice
Tech
This Crowdstrike explainer is really cool but let’s be real here I’m mostly linking it because it taught me that information security people constantly talk like they’re from a Tom Clancy novel.
If you inject “enzymatically digested pig brain tissue”, will your brain work better? No.
Particularly Good: Someone interviewed a bunch of engineers who had worked both with atoms and with bits to learn about how the two kinds of engineering are similar and different.
Fun
Particularly Good: Someone at the New Yorker got paid to infodump about sharks and I’m all for it.
The rooms in Open Philanthropy’s offices are named after “big wins” for philanthropy, including The Pill, Marriage Equality, and Sesame Street.
Particularly Good: The pitch meeting for Animaniacs. “Animaniacs will not make sense to them now. It will make the world make sense to them later.”
If Jane Austen wrote Star Wars.
Mr. Beast is fine: “Who, if they had the ability, wouldn’t want to do real life Squid Game? Or try to blow up a yacht while their friends tried to protect it? Who wouldn’t buy a train and then drive it off a cliff? Are you saying that given unlimited money, you wouldn’t build giant obstacle courses like a demented Willy Wonka? Liar.”
“Bounce” is not a word that should ever appear as a description of what a plane is doing, especially near the adverb “twice.”
Chinese Doom Scroll: high-class hotel puts antibiotics in food to hide that the meat is expired. Cute STEM flirtation. Apparently American Olympic athletes are “blood doping.” Chinese people pay attention to murders in Kenya?
Ideally, we would not call anyone cringe, but I’ll save my appeals to the better nature of the American voter for when the American voter isn’t considering electing someone who tried to do a coup.
Blood doping is a real thing some athletes do! (It is considered cheating, it is not allowed in competitions.) Maybe the quotes were not intended to imply "these people made up this term", but like, it is a real thing and not new.
Huh, I tried every trick in the book to avoid going to school! Yours must be really cool. It was so boring to listen to someone droning on about what some dead old poet thought.
I think becoming "cultured" was largely a signalling system, but this rapidly lost its utility. People do not assign much status today to quoting Juvenal.
It is an agonizing question whether reading Juvenal should be abolished, since so much of the identity of the Western culture depends on matters like this. But it is also clear today people assign status to people who quote Acemoglu, and that is more useful and relevant anyway.