Linkpost for December
Effective altruism, subcultures, mental illness, life advice, policies, short stories, fun
Effective Altruism
Global Poverty
Statistics in developing countries are really, really, really bad: “Many African states are failing at the basic task of knowing how many people live in their borders—let alone accurately measuring their economic activity. The vast, unobserved informal sector (which includes subsistence farming, and something like 60% of working people) is usually estimated just as a direct function of population. Lacking direct harvest yields, estimates of agricultural output are often produced using FAO models based on planting-season rainfall data. Even the minimal task of measuring the goods traveling across borders—in theory, the easiest thing for a sovereign state to accomplish—is occasionally beyond the reach of statistical agencies. Until 2008, landlocked Uganda only collected trade data on goods that eventually passed through the Kenyan port of Mombasa, ignoring the four other countries on its borders.” This presents a serious issue for evidence-based health and development policy: the evidence we rely on is very thin.
About 25% of people worldwide have anemia, usually caused by iron deficiency. Anemia is preventable through nutritional supplements, hormonal contraception (which reduces blood loss from heavy periods), prenatal care, and prevention of infectious diseases.
Jimmy Carter, a mediocre president and one of the greatest former presidents of all time, is dead. The guinea worms outlived him. :(
Animal Advocacy
Sentient Media’s most surprising stories of 2024. I myself was surprised to learn that monkeys are enslaved to make coconut milk. (For those concerned, Sentient Media includes a list of coconut brands that don’t enslave monkeys.)
Existential Risk
Particularly Good: I’ve been reading Joe Carlsmith’s series ‘Otherness and Control in the Age of AGI,’ which I have been avoiding for a while due to my deep-seated bigotry against people who use Magic: the Gathering card colors in serious conversations. More fool me; the series is great. The first piece, Gentleness and the artificial Other, has a clear-eyed view on cultivating the virtue of wonder when interacting with digital minds. About a week ago, I was trying to articulate some thoughts along the lines of ‘An even deeper atheism’ in conversation, and then I read the article and was like “Oh! That’s exactly what I meant, I’m glad someone else said it so I don’t have to.”
Some people worry that models trained on synthetic data (i.e. data generated by AIs) will be worse than models trained on human data. This probably isn’t true, as long as synthetic data is curated. Synthetic data is useful to fill in gaps in human data (such as multilingual exchanges).
Everything that happened with large language models in 2024: GPT-4! GPT-4 on your laptop! Lower prices, but also you have to pay more for the best models! Images, audio, video, voice chat, and live video! No agents! Evals are very important and hard! Chain-of-thought reasoning models! Synthetic training data is fine! AI slop! LLMs are hard to use and most people don’t know how to use them! We didn’t start the fire
Actionable: promising funding opportunities in AI safety.
Meta EA
You don’t have to justify all your life choices as the maximally EA thing to do.
How to attend an EA conference without burning out: set your goals for the conference and say “no” to activities that don’t fit with your goals. Don’t push yourself beyond your limits, and take breaks. Schedule important activities for your prime time, and plan the rest of your schedule around making sure you’re OK during prime time (e.g. no late nights if you have early-morning meetings).
A nice little post from Giving What We Can explaining how some charities can be thousands of times more cost-effective than others. Good examples.
GiveWell staff’s personal donations for 2024. The most popular donation target continues to be GiveWell Top Charities or All Grants Fund. Many people also gave to GiveDirectly or animal welfare causes. One person donated exclusively to local causes, while six (out of fourteen) made at least token donations to local causes. Two people made donations to help Palestine.
About four-fifths of applicants to the Effective Altruism Global conference are admitted. In general, the conference admits almost everyone who seems to understand what effective altruism is and to be in a position to improve the world using effective altruist principles. Therefore, we should all have less woe about EAG admissions.
Open Philanthropy has stopped funding university effective altruism groups and passed its portfolio to the Centre for Effective Altruism. The major changes: part-time university organizers will no longer be paid; full-time university organizers may or may not continue to be paid; funding for group expenses is now less flexible and (reading between the lines) less generous.
People who don’t natively speak English can come off to native English speakers as less smart and articulate. This bias means that effective altruist leadership is disproportionately made up of native English speakers, who often lack expertise in dealing with non-Anglosphere countries.
Other Causes
Trump’s immigration team is currently divided. One side wants mass deportations, which would require a massive expansion of state capacity, including building dozens of new detention centers and calling up the National Guard. The other side wants to basically continue business as usual, but with more anti-immigration rhetoric. Trump also plans to hold people in detention instead of letting them go until their trial date. Since that policy would lead to massive overcrowding of immigration detention centers, it’s likely a bunch of people would go back to their original countries—even if they were legally allowed to be in the country—just to get out of the detention center. The policy also pretty intractable to fully implement: ICE is aware of 7.4 million people it believes it is currently allowed to detain, which is more than three times the current U.S. prison population.
Subcultures
Sarah Constantin on the underlying ideology shared by “reasonable” people around her. These include: caring about Goodhart’s Law; trying to outperform the normal way of doing things; avoiding busywork; not caring too much about social approval; caring about doing good work; and including outsiders and newbies. I like pieces like this that try to make concrete the idiosyncratic ideas and worldviews that feel like “common sense” or “just the way people like us do things.”
Building a subculture: Pick something to rally around. Create symbolism of your subculture. Choose an outgroup to be against. Make spaces for your subculture to congregate. Build connections to existing groups. Let go, and let the subculture evolve for itself.
Although this post is called The Abuser’s Guide to Transmisogyny, it’s a well-observed takedown of the ways social-media mobbing is used against marginalized people. “Really, literally anything you can imagine anyone having any conceivable level of problem with can be reframed into your victim having committed some unforgivable sin only solvable by exile... If you lean into it enough, anything can work for your desired life destruction, as people are itching to get rid of trans girls in the first place, and at its core, you can tap into that with anything.”
Mental Illness
We should frame mental health, not around “what’s wrong with you?” or “what happened to you?”, but around “what matters to you?” That is, mental health shouldn’t focus on people’s past experiences or pathologies, but on their needs, desires, and goals. People can work through their trauma or seek medication, if they want. But if people care about practical matters like getting a job and housing, or finding solidarity with other people like them, or the spiritual nature of their experiences, mental health programs should also support them.
Particularly Good: I loved, loved, loved this article about palliative care for anorexia: that is, for people with very treatment-resistant anorexia, giving up on recovery and setting a goal of making the patient as healthy and happy as possible in the time left for them. (As you might expect, the article contains a lot of explicit discussion of eating disorders and suicide.) I’ve never seen an article this supportive of mentally ill people’s autonomy in a mainstream publication! As the author points out, mental illness is treated differently from physical illness. If you have cancer, you can refuse treatment because you think the benefits aren’t worth the side effects, or for that matter because you want your naturopath to treat it with green juices. If you have anorexia, you can’t. I (and the author) agree that involuntary treatment of anorexia is sometimes justified: often, people only want to be anorexic when they’re starving, and when adequately fed they don’t want to starve to death. But if a person can never refuse treatment for anorexia, we’re saying that (unlike any physical illness) treatment for anorexia always passes a cost-benefit analysis for everyone, no matter how invasive the treatment or small the chance of success. The piece also points out the downsides of palliative care for anorexia: for example, anorexics are competitive, and if palliative care for anorexia is allowed many of them will try to starve enough to qualify. It’s a good piece. Check it out.
Related: a heartbreaking personal essay about the author’s childhood friend, who died of a heroin overdose. The author points out that heroin addiction is often a rational response to having no prospects if you get clean.
Particularly Good: A beautiful piece about the children of writer Alice Munro coming to terms with her covering up her partner molesting them. I particularly liked the discussion of Munro drawing on her experiences covering up molestation to write stories. It raises questions about what material it’s all right to make art out of (although I guess maybe the answer is “anything, and also don’t cover up your kid being raped”).
Life Advice
A review of the evidence about social contagion and quitting smoking, with an emphasis on its application to animal advocacy. Romantic partners are most important: people are between 1.3 and 11.8 (!) times more likely to quit if a partner quits. Friends are the next most important, followed by coworkers, then by siblings. The effect goes all the way out to friends of friends of friends, although by the fourth degree of separation it fades.
People who have important messages to convey often fail to convey them because they’re bad writers. Consider learning to write before you have an important message.
Breathing weird can give you a psychedelic trip????
The apparent decline in testosterone is caused by changes in how we measure testosterone. We are presently measuring exactly how much testosterone people have. In the past, we had to estimate and there was a lot of noise and mismeasurement.
Policy
Clinical trials are inefficient because of a cultural problem: the culture around clinical trials tends to be risk-averse, slow, and paternalistic. “But messaging around the actual steps that are needed to accelerate [medical innovation] has for a long time been contrary to progress, such that whenever there is a trade-off between accelerating innovation and literally anything else (e.g. a perceived risk of undue influence), it seems that the other concern wins. In practice, society is pro medical innovation to the extent that such innovation is carried out by self-disinterested, non-profit making entities with absolutely zero risk whatsoever when it comes to any privacy related issue and given maximum paternalism.”
X’s Community Notes feature involves trusting users, unlike most social media fact-checking, which involves trusting a small group of expert elites. X’s Community Notes were carefully designed to give reliable information without X’s direct oversight. For example, a note is only shown if a bunch of people who normally disagree with each other agree it should be shown. Further, X has no ability to remove individual community notes, except by turning off the whole feature: if the people wish a community note to be on a post, it’s on a post.
Particularly Good: A few days in the life of a Texas ob/gyn as he decides whether to leave the state.
Particularly Good: I don’t agree with the politics of the author of Confessions of a Bastard Cop, but I still recommend the essay. The author has a clear-eyed view of the problems with police officer culture. They don’t assume that cops are just bad people, so there; they explain why cops behave badly. This essay makes police brutality and violation of civil liberties understandable as a result of a dysfunctional culture.
Short Stories
Because of the Yuletide small-fandom fanfiction exchange, Her Voice Is A Backwards Record has two, count em, two pieces of fanfiction. I am ecstatic. The author is from the “fanfiction should give you something you didn’t get in the book” school of fanfic writing: she gave us Yuya and Ana having a nice time post-book.
A World Away (A Step Apart): Superhero romance with the main pairing “Male Supervillain with Day Job at a Coffee Shop/Tired Male Superhero In Need of Caffeine.” The worldbuilding about how superhero RPF would work is sharp, well-observed, and very funny.
Xeno ISO Synth for One-Time Encounter: Very NSFW. A bug alien writes a personal ad seeking a robot for a BDSM scene. I enjoyed the subtle inclusion of the worldbuilding; it evoked a world much larger than the one we see in the story. Plus, in-universe documents are always a win for me.
Particularly Good: 17 Last Words from Hunger Games Tributes That Are STILL Echoing In Our Heads. Yes, Still.: The pitch-perfect Buzzfeed voice combines with the author’s eye for the most precise fucked-up details to make the most disturbing Hunger Games fanfic I’ve read in a while.
Particularly Good: Effective Altruism and Rationality meet at a Secular Solstice afterparty: What if we talked about philosophy and the end of the world (and we were both girls).
Fun
Chinese Doom Scroll: Techniques for accountants to avoid being prosecuted for fraud when (as is surely inevitable) their bosses try to get them to commit fraud. Stealing clothes from bars. A Daoist master’s snarky answers to questions (are all Daoist masters like this? Xianxia writers take note). The Untamed!!!!! Why are all these people literally starving? Teachers are so mean to kids! Oh my god Chinese therapists why are you like this. Chinese Olympic athletes doing online sex work; plus, what English words Chinese people think are fun to say. The worst ending for a romance novel is when they break up and are happy with other people. Why is a Japanese government official doing sex work????
Relatedly—Chinese Doom Scroll is in Asterisk Magazine with a guide to Chinese Internet slang!
Things to argue about over the holidays other than politics.
I think that minimum wages make sense when an employer is paying below the (middle of the range of) the market wage, because the employer has more leverage. (Or occasionally other effects, but that's much more common.)
I agree with that essay that if prices between worker and customer are liquid, fiddling will only make it worse.
I think the question should be, how often are wages NOT a fair market value. I think that's common but I don't know how to tell how common it is. I'm interested how much people familiar with market arguments agree. And I wonder how much it varies by country, both in fact and in opinion.
> I don’t agree with the politics of the author of Confessions of a Bastard Cop, but I still recommend the essay.
This essay reminds me, more than anything else, of what I've heard about fake conversion stories in american evangelical christianity. I'm told the demand for actual conversion stories far exceeds supply, and so people sometimes pretend to have once been wretched sinners who saw the light and have now reformed into model christians. These stories serve both to confirm the badness of people the audience already thought were bad, and to provide the positive affirmation of someone on the other side agreeing they're actually right about everything.
The author of this essay claims to have once been a "bastard cop", and assures us that all the other cops are also bastards, in exactly the way you, the presumed cop-hating audience member, already thought they were. Since then, they have somehow transformed into a perfect doctrinaire ACAB leftist, repeating seemingly every talking point I've ever heard from that group, with no caveats or remaining points of disagreement. No explanation is given for this radical transformation, and the overall message seems to be that the assumed audience is completely right about cops, both in general and in every particular, and that the experience of being a cop would only confirm that.
I am accordingly suspicious of both the author and of the process that lead this particular essay being shared around.