Software Developer at Giving What We Can, trying to make giving significantly and effectively a social norm.
Lorenzo Buonanno🔸
Karma: 5,678
They seem to have made a lot of money by investing in his fund, even after this week, so I find it hard to criticize them for it
Aschenbrenner didn’t launch the fund alone, his cofounder Carl Shulman had previous experience managing money
We don’t know the rest of their portfolio and how situational awareness fits into that (e.g. maybe it’s partially a small hedge against AI making Stripe worthless, or who knows)
it is objectively nuts to put a 24-year-old with no experience in charge of a $50B hedge fund
That would indeed be nuts, but I think that’s also clearly not what happened. My understanding is that he founded the fund himself, and he initially raised $100-200 million from a bunch of non-EA investors. See https://www.forourposterity.com/
Thanks to their spectacular returns they had a lot more assets under management, but my sense is that this is from normal investors who buy into the “AI is going to be big” thesis and were trying to make a profit.
I also found this article interesting, as well as the fact that Bloom Energy, IREN, Nebius, SanDisk, Micron, CoreWeave, and so on are up 10% to 30% since Wednesday
They certainly lost a lot of money in July and by being forced to sell, but per the WSJ Aschenbrenner wrote that the fund is still up 80% on the year.
The fund doesn’t seem to be closing, and it looks like they weren’t forced to sell their stake in Anthropic, so it’s not clear to me how much of a failure or a success the project as a whole is so far.See also this from the Financial Times.
Edit: see also this tweet claiming to be the full letter Aschenbrenner sent to his LPs last night.
Moskovitz confirmed that they are “surging budgets in 2026 across all categories”
https://bsky.app/profile/moskov.goodventures.org/post/3mrdb7fhfqs2w
I expect (most of?) their grantmaking in other areas will be less public, but unprecedentedly massive.
E.g. Dustin Moskovitz is personally aiming to donate ~$5M to Lightcone Commons
Oh! I had the idea that new posts start with 3 karma, but then I was wrong.
They start with a self-strong-upvote, which depends on the user karma https://github.com/centre-for-effective-altruism/ea-forum/blob/27d24f53b7dce59b770040e6ac9ff05fcb9c4b16/src/lib/votes/voteHelpers.ts#L71-L91
If you look at the newest posts on any given day, you’ll see a number with 0 or 1 karma. If my understanding is correct, this means they have been downvoted.
I don’t think this is correct, you can mouseover the karma score of a post to see the number of votes. E.g. here it has 1 karma from 1 vote (the automatic self-vote that is given when you post) and no downvotes
I think these posts are either not being shown on the frontpage to most people, or don’t get voted on before disappearing
Thank you for writing this comment. It made me realize that a main reason I feel uncomfortable talking about effective giving with the wealthiest people I know is that I implicitly worry they might decide to buy a mansion or a yacht, instead of saving thousands of children from death or poverty, depending on how exactly I phrase things. This is not particularly respectful towards them[1] and I’m going to think differently about this going forward, and try to respect their judgment and their character more, and have a lower bar for talking to them about effective giving.
But I want to push back as strongly as I can on this:
I’m left wondering whether part of the concern about cold outreach is that, when it does work, it allows organisations to bypass the existing intermediaries. Middlemen tend not to be enthusiastic about that.
I think it’s valuable to flag potential conflicts of interest that might bias someone, and to flag self-promotion, but we should have a very high bar for imputing selfish intent.
Besides being against forum norms, I think these speculations about motives empirically have had significant negative consequences (three examples here), make cooperation harder, and discourage people from doing anything that can be perceived as self-serving (which includes many of the most impactful things someone can work on).
It might be a stretched example, but e.g. you mention in another comment that you had success doing cold outreach. I think it would be much more productive to discuss what makes cold outreach net-positive vs net-negative, and discuss your experience, compared to whether you’re defending cold outreach out of personal interest.[2]
€1M is a reasonable amount for a dedicated person to donate over a few years.
I don’t think this is true for the vast majority of dedicated people, especially someone using € as their currency instead of USD, CHF, or AED. (I would strongly recommend @Andrew Neiman🔸 to focus on moving to a wealthier country with lower taxes before focusing on working hours)
Even when donating €50k a year (the entire yearly net salary of >90% of Germans) it takes 20 years to donate €1M, my sense is that for dedicated solopreneurs even getting €1M in revenue after a few years is considered a huge success, so €1M donated in 10 years would be really impressive.
See also this comment which I also agree with.
One point where I disagree is that I’m more concerned about well-intentioned people genuinely believing their project is extremely valuable, but who lack clear evidence of impact and end up not achieving much. I’m less worried about people deliberately lying/exaggerating to raise money for things they do not themselves believe are impactful.
But of course it’s hard to say from the outside, you could always argue that people are secretly doing whatever they’re doing just for the money & status, and I’ve seen people making such claims even when there’s tons of evidence to the contrary.
Here’s another list from Caroline Ellison during the FTX times on valuable signals: https://forum.effectivealtruism.org/posts/M44i4CiMECP5Xoorz/demandingness-and-time-money-tradeoffs-are-orthogonal (which I’m not sure I agree with, but I found interesting)
I personally think people worry disproportionately about grift on the EA Forum, while implementation challenges and other considerations seem to be bigger risks in practice. (e.g. I don’t think the Wytham project or Dispensers for Safe Water would have been more successful if they had been assigned to more vegan or more frugal people)
PSA to please use https://forum-bots.effectivealtruism.org/ for scraping rather than the main forum
It looks like forum-bots is also soft blocked for AI scrapers?
I’d propose making it as easy as possible for AIs to use this forum. ( it’s likely a very marginal issue, as they can easily work around it by changing the user agent or using ea.greaterwrong.com , but might still be worth it to spend some time to make things easier for them )
You might also be interested in this old post, tangentially related to this topic https://slatestarcodex.com/2019/06/03/repost-epistemic-learned-helplessness/
I used to think the same, but I recently changed my mind, for reasons similar to Zach’s
Some quick thoughts:
Significant additional spare money
I agree money is great, but I think moving to a HCOL area will often increase the amount of money you can earn/save for most people, by giving you higher-earning job opportunties. Even if you’re doing self-funded research, the amount of money you can influence will likely increase enough by moving to a HCOL area to compensate your higher costs. It’s much easier to help people give more and give better if they’re wealthy.
I agree that the above often doesn’t apply if you work in a field with a capped low salary, where living in a HCOL area would cause financial distress
only go to two or fewer EA events per month
This is a lot of events, 10-20 per year. Similarly “often barely talking” to others in a coworking space imho doesn’t mean that there isn’t a lot of valuable talking happening per year in absolute terms.
elitism
Hot take: I think elitism can be as much of an asset as it is an issue image-wise. Lots of fancy enterprises proudly boast their “talent density”
How can they claim to care about cost-effectiveness if they spend all their time in the most expensive parts of the country?
Because that way they can influence more resources to go towards solving the most important issues of our time
Basically:
I think people can earn and donate much more in HCOL areas, compensating the higher costs
I think people can have a ton of impact by influencing the actions of EAs and non-EAs around them, and in HCOL areas those are 1. more likely to find EA useful, 2. have more resources (both money and career-wise) to dedicate to it
I think you’re downplaying the influence of “only <2 events per month” or only a few conversations per week at a coworking space
I’m very far from working in the AI industry, so I don’t think I’m the target audience of this post.
In any case, as someone trying to use part of my resources to help others, I agree with much of it (especially points 1. 2. and 6.) but I personally strongly disagree with other points:
3. The first priority in charity is to mitigate potential harms caused by how you made your
money.For people working in AI the potential harms of their work are immense, and involve most present and future life, so I think it’s reasonable to prioritize mitigating those harms, but I don’t think it’s a good general principle for the vast majority of givers.
Suppose an extreme scenario where I make billions of dollars developing a practical form of nuclear fusion, which quickly devalues fossil fuels. A large investor in fossil fuels loses a lot of money as a consequence, and has to cancel their $2M space tourism trip.
This is a clear harm to them directly caused by how I made my money. Why should my first priority as a philanthropist be to mitigate this harm, instead of preventing thousands of much poorer people losing a child from preventable disease?More practically: most harms caused by most of our careers are local (e.g. if I become a doctor, someone else won’t get to be a doctor) and I think our priority in charity should be to help those that we can help the most, who usually are far removed from any effects of our careers.
5. Give to issues that are close to your heart and/or your field of expertise, so you
understand them better.I think giving to causes close to my heart is a very flawed moral guide which will predictably lead me to help fewer people and help them less. Largely because by default the issues closest to my heart would be problems like Parkinson’s disease that already receive a lot of resources, and with the same amount of resources I can help far more individuals far more, people who are suffering from issues I was born too lucky to ever see (e.g. tuberculosis kills 1.2 million people per year). Their suffering is just as real, and matters just as much. I think many great points raised by Paul Bloom in Against Empathy also apply here.
I do agree that ceteris paribus one should favor issues that are closer to one’s field of expertise, especially when able to identify particularly promising opportunities, but again, my field of expertise is unlikely to involve the world’s most neglected issues (almost by definition) and we should be mindful of that when evaluating funding opportunities.
4. Beyond that, don’t try to optimize charity too much. Is it better to donate to a political
candidate who will change the healthcare system in the long term? Or to donate to sick
people who need expensive surgery right now? Optimize a little, but then focus on the
giving. In order to optimize charity, you need two things that we don’t have: A calculus
of suffering and a crystal ball that tells the future.I agree that there is a risk of trying to optimize charity too much. In practice I think the main risks are analysis paralysis, or feeling paralyzed by extreme cluelessness, and as a consequence donating much less than otherwise, or not at all.
But there is definitely also a risk of optimizing too little, and I worry that optimizing “a little” results in far fewer individuals being helped. It’s true that without crystal balls and a perfect measure of suffering you can’t be certain of whether a choice is optimific. But the same applies to choices people make in their own lives, where we all agree it’s rational to spend more than a little time comparing options.
E.g. If I were diagnosed with a serious medical issue with several treatment options with different risk profiles, nobody would tell me: “There is no perfect measure of health, and no crystal ball about complications, so just compare a little and then focus on getting some surgery.” People would advise me to get a second opinion, study the literature, compare hospitals and surgeons, ask for expert advice, and so on. None of this can guarantee the optimal result, but for things we care about, optimizing ‘a little’ is not enough, and when talking about charitable donations the stakes are extremely high. Unlike hospitals or recommended treatment options, some charities probably do 100x more good per dollar than others, so the returns to comparison are unusually large.
I also imagine that Harari and Sapienship spent quite a bit of time and research before shortlisting those 3 specific recommendations, so it’s likely we don’t actually disagree here.
For anyone curious:
https://impactpro.substack.com/p/the-case-for-having-children-sooner
Thanks!
maybe something on the order of 10% of their income
Not sure how much it matters, but the blog post mentions $10,000 to GiveWell alone, which was likely significantly more than 10% of a graduate student’s income.
$1B is a very low number over 4-5 years, given their wealth and how much it could grow
Yeah I agree, for what it’s worth I would expect (with low confidence) they’ll donate significantly more. But in general, for most planning, it doesn’t matter how much they give in relative terms but in absolute terms.
This is a slight update
A larger update for me was looking at GiveWell’s public board meeting records, in particular Daniela Amodei is the person evaluating GiveWell’s CEO, and seems an active board member here
I would bet that Dario Amodei makes <$1B in donations before June 2027, self-reported
Something interesting/unique about Anthropic’s situation is that all 7 cofounders are now likely worth >$10B, so even if Dario doesn’t donate much there’s still a good chance others do (but obviously they’re correlated.)
Personally, I’m most uncertain about whether they’ll end up donating significantly to improve the welfare of biological beings vs focusing on digital beings. I expect them to take machine welfare quite seriously and increasingly so, and that to be something that markets and other funders won’t care as much about.
I think it’s also possible that they end up donating after 2027, timing donations around a potential critical transition period, and of course they have strong incentives to focus most of their capital on winning the race.
Are you taking bets on whether the Anthropic founders will donate more than 1B before 2030? I think there is a lot of information that should update our “average billionaire” prior (e.g. https://blog.givewell.org/2010/06/03/my-donation-for-2009-guest-post-from-dario-amodei/)
I would interpret all three as signals that orgs find it harder to fill research roles, right?
research [...] which they generally find easier to fill
I’m surprised to read that lots of EA orgs find it easier to hire research roles than ops roles, and it doesn’t match what I heard, or the state of 80k’s job board at the moment, with ~1.8x more research roles than ops roles
Edit to clarify: my sense is that many orgs struggle to hire both for ops and for research
The top six posts by karma of 2026 are all criticisms of EA, two of them of CEA explicitly