I live for a high disagree-to-upvote ratio
huw
Karma: 2,788
I don’t know how you get around this, but the framing of ‘lives saved’ feels a bit off, if you’re not going to count ‘deaths caused’. For example, there’s a guy on there who’s probably responsible for a few million deaths, give or take, because he pretty much single-handedly shut down USAID. It feels icky to see him up there tbh, given the harm we know he’s done.
Why is Australia not an EA ‘hub’ in the same way even
lesserless-hubby-than-the-UK European countries like Germany or the Netherlands are?
I downvoted your post because I think this is really unhealthy, and that no amount of caveating can save it from promoting unhealthy behaviour to the rest of the community.
Frankly, if you can’t measure your work in a that an independent person could evaluate, I would be very skeptical that you’re having any impact at all.
The kind of grifting that most concerns me within EA would result in a lot of fudging and inflation around reported impact in order to gain money & status. I think the hardest thing is fake, and the main thing we care about protecting, is real, verifiable impact. If there is an influx of money into EA, I would want the first thing to be done is to undertake a lot of new, third-party evaluation with depth. This is not happening today and I have not seen great plans to do so right now.
G’day Madeline, I run an EA mental health org in India. The reason for this is simply that existing mental health interventions do not compare with GiveWell’s grants on a DALYs/$ basis. In my opinion, the reasons are:
DALY moral weights may be biased against depression
The moral weights of different diseases in the Global Burden of Disease study, which informs DALY estimates, are determined by asking the general public whether they’d prefer to have one disease against another. When you do this with depression, people who haven’t experienced it tend to prefer to have it to many other conditions. However, when you ask people who have experienced it, they choose many very painful conditions over depression. This is one of the widest gaps in the moral weight data. See Pyne et al. 2009 and this post.
Psychotherapy is usually modelled as a short-term effect
Psychotherapy is typically modelled as a treatment, and not a ‘skill’. What I mean by this is that a dose of psychotherapy is assumed to only have effects that decay over a period of time and zero out after that in most CEAs, including those from the Happier Lives Institute. However, many psychotherapy patients will tell you that they learned skills that were useful long after the therapy ended, and there is some limited evidence that psychotherapy’s effects may last decades, or potentially never zero out. If this were true, the effects could be very long-lasting and therefore it would be much more valuable to treat a case of depression.
Suicide prevention isn’t cost-effective if it’s only a short-term effect
Consider that for most of GiveWell’s top interventions, the bulk of the DALYs averted come from ‘saving’ a life—i.e., preventing a death from a disease in a way that allows the person to then go on to live a healthy life, such as preventing a malaria case in an under-5 (which they might die from), even if they go on to catch it after 5 years old.
As a short-term effect, psychotherapy can only postpone a suicide by the length of the treatment effect. But if it were a skill and had some durable long-term effect, it may genuinely prevent one, which would tremendously increase the value of suicide prevention interventions.
Existing interventions haven’t been cheap enough yet
With the exception of some incredible policy work in, for example, reducing toxicity of pesticides commonly used for suicide, existing interventions are still quite expensive. The Happier Lives Institute’s top charities cost ~$40 to treat a single person, while a bednet costs $7. I’m fudging the numbers a bit here, but if we stick with DALYs, psychotherapy is still about an order of magnitude more expensive than it needs to be to look great for EAs.
However, there is work being done to improve that! My charity, Kaya Guides, treated people for $20 each in April, at what we estimate is a similar effect size to the best charities, and we’re confident we can get below $10. We’re using a technique called guided self-help that allows us to dramatically reduce contact hours per participant (and being all-digital helps a lot, too).
Conclusion
Orgs like the Happier Lives Institute have done a lot of advocacy work too, to raise the profile of mental health within EA, and there are plenty of funders that take mental health seriously (in a way that apparently wasn’t true a decade ago). It is, after all, still a nascent space.
(Just FYI, the current forum policy is to downrank LLM-written posts, if you want more visibility on this—which is undoubtedly useful work—it currently reads to my eyes as very LLM-written so could be adjusted a bit)
Just as some feedback, I don’t think the usual objection to working on longtermism or GCRs is ‘it’s not valuable enough’, even from a general audience. I think it would be more persuasive to explain somewhere why working on these causes would achieve anything at all. It’s not clear from your post that there would be a relationship between effort and rewards, and I suspect a lot of people share the intuition that any counterfactually valuable work on the very long term future would be washed out or easily reversible.
Always struggled to understand who these are supposed to be for. Politicians and policymakers don’t know or care about what an acausal trade is, and I suspect a lot of this piece is just too ‘out there’ to be persuasive to someone who spends most of their day doing ordinary politics. Is it supposed to be for the people who will then persuade the politicians? Something else? Very confused
It somehow completely fails to cover Leverage deploying spies into other organizations and trying to take over CEA
I’ve heard this claim from Habryka and others many times and nobody seems to be willing to go on the record to back it up. He does not really think to ask that perhaps Lydia also heard about these rumours and was unable to substantiate them with the rigour required of a published magazine article either.
At least to just satisfy my curiosity, can someone involved privately DM me and explain what this means and provide some evidence for it?
(I otherwise fully agree with Habryka’s assessment that this article feels waaaaay too credible about Leverage and seems to be mostly doing image rehabilitation, from someone who is simply too personally close to the story to be believable as a third-party observer)
(Just realising I am developing an eye for AI-generated graphics now, to complement my eye for AI-generated text, images, and videos. There’s something so familiar about this style.)
I was never accepted into the Charity Entrepreneurship programme, but I was able to land a high-impact role at one (now leading it!), by getting in touch with other founders. For me, this is no less rewarding than being the real founder, and enough people treat me as if I went through the programme that it’s usually a surprise to people that I didn’t.
So I would urge some people to reconsider if founding is exactly what you want, you may be able to derive almost all of the benefits via a slightly different path, and CE’s charities would very much welcome talented founder-type generalists in high-level leadership & ownership roles!
🔱 Please choose a slightly heavier weight for the font 🔱
We’ve started the process of creating a brand new codebase for the forum built on NextJS
I initially balked at rewrite syndrome, but thinking again, I think LLMs probably do really change the calculus of it. I like that you announced this by showing the proof of it in the new post page design!
A lot of modern training data isn’t stolen, though. There are organisations which recruit people to do their jobs normally and screen share, or provide worked-through examples of their work, and this is increasingly making up the bulk of the data that’s used to pull frontier models ahead of others on work benchmarks. People are being paid for this and do it willingly, usually with knowledge of where their labour outputs are going!
So really, the problem is a subset of workers in each field are ‘defecting’ (to use a rat term I kinda loathe). How do you create solidarity among groups of workers to prevent a small number of them from putting the others out of work? Or, if technological progress is to be necessary, how do those groups of workers politically agitate for a welfare state and good ongoing education?
The left solved this problem two hundred years ago, but I suspect EA won’t like the solution…
(Just contributing by adding that yeah, GiveDirectly’s fund works because Yemen is so dominantly muslim that the odds of going to a non-Muslim are basically nil). Even in Nigeria, the additional cost of trying to achieve Zakat compliance for a non-cash-or-in-kind-transfer charity may hamper expansion beyond the supermajority states enough to not be worth it. Unsure.
Always loved your work, but fully understand the decision. Thank you for it!
One thing which is unclear to me is, why aren’t these users counterfactually using free commercial offerings? Price is clearly not a barrier, is it just language? And why, then, wouldn’t a frontier lab be well-positioned to capture that market?
During the OpenAI board fiasco, we saw a large number of employees exert pressure on OpenAI to re-form under Sam Altman, suggesting labour power in AI labs is real and effective. Companies do have a hard time shutting themselves down for no reason, but have a much easier time scapegoating labour unions and strikes. And it’s relatively inexpensive to build a labour union, and easy to do when talent is scarce. Just sayin’
Like, if your goal with this project is to inspire rich people to think about doing good, I wouldn’t include an obviously really bad guy on the list.