Yarrow Bouchard 🔸

• Yarrow Bouchard 🔸 10 Oct 2026 21:36 UTC

  2 points

  0 ∶ 0

  in reply to: alesziegler’s comment on: Yar­row Bouchard’s Quick takes

  Huh? Not sure I follow this comment. I have definitely stated outright what my views are about fears of near-term superhuman artificial general intelligence (AGI). I have written a lot about this here on the Effective Altruism and also on my Substack blog. See, for example, this recent post: https://​​strangecosmos.substack.com/​​p/​​the-sloppy-research-fuelling-fears

  Maybe the idea of making a bet about a future prediction seems unusual to you. But it’s not out of place on the Effective Altruism Forum, since the effective altruism community normalizes and encourages bets like this. Personally, I’m not sure prediction-based bets are all they’re cracked up to be, but at least they could be a bit of fun.

  You shouldn’t take a bet you think you’re likely to lose. That defeats the whole purpose. Personally, I also think even if you’re only, say, 60% sure you’ll win a bet, it sort of undermines the purpose as well. It’s way more interesting if two people take opposite sides of a bet and they’re both highly confident that they’ll win.

• Yarrow Bouchard 🔸 7 Oct 2026 19:32 UTC

  2 points

  0 ∶ 0

  in reply to: Yarrow Bouchard 🔸’s comment on: Yar­row Bouchard’s Quick takes

  It occurred to me what a drag it would be for the other person in this bet to wait until January 1, 2034 to win. I thought of a few win conditions that could call the bet in the other person’s favour before the clock runs out. If any one of these conditions are satisfied, the other person would win the bet:

  • Either Anthropic or OpenAI (post-IPO) reports $100 billion in annual GAAP profit (net income) or $25 billion in quarterly GAAP profit (these would have to be audited financials reported to the SEC, like public companies typically do)

  • Ray Kurzweil wins his Long Bet against Mitchell Kapor because the Long Now Foundation determines that an AI system can pass the Turing test by the end of 2029

  • A rigorous survey of AI experts by a reliable organization (e.g., AAAI) finds that a majority (50.1%+) think artificial general intelligence has been created (not something like an informal poll on a blog or forum, obviously)

  • A household humanoid robot is sold to regular consumers that can autonomously, without human assistance, scramble eggs successfully 95%+ of the time

  • Fully autonomous SAE Level 5 robotaxis are widely deployed in the U.S. or any country, with at least 100,000 Level 5 vehicles on the roads, without using any form of human assistance or remote operation, and without being geofenced to an area smaller than 95% of the country’s public roads

  These are rattled off just off the top of my head and we could probably come up with some better ones to add to the list.

• Yarrow Bouchard 🔸 7 Oct 2026 19:13 UTC

  2 points

  0 ∶ 0

  in reply to: fergusq’s comment on: The sloppy re­search fuel­ling fears of su­per­hu­man AI

  I agree. That’s an important point. Thank you. Peer review is a floor, not a ceiling. Replication is a higher standard.

  There’s also something complicated and idiosyncratic about the machine learning field. Machine learning papers, not just METR’s, sometimes want to measure both humans’ performance and AI systems’ performance on the same sort of tasks, in order to compare them. There is a computer science aspect of this research and a social science aspect.

  To say that an AI system achieved such-and-such a score on a benchmark is typically uncontroversial. But can ML researchers properly conduct scientific studies on humans? And what of the validity of the benchmark? Can the researchers accurately, properly score human performance? And, more importantly, does the benchmark actually constitute a meaningful, truthful, apples-to-apples comparison between human performance and AI performance on the same sort of tasks?

  When ML researchers try to do scientific studies involving humans, they might not be competent to do so. They might have no experience, training, or education in that kind of research. Peer review from other ML researchers is not helpful if the reviewers are also just techy laypeople with regard to social science.

• Yarrow Bouchard 🔸 6 Oct 2026 22:53 UTC

  −1 points

  0 ∶ 0

  on: Yar­row Bouchard’s Quick takes

  Open invitation: bet that the AI bubble won’t pop

  I’m taking my second stab at formulating this AI bubble bet in a way that is attractive for people who might want to take the other side of it. I’m betting the AI bubble will pop within about 7 years. I’m looking for people who are willing to bet it won’t pop, or that there is no AI bubble.

  I’m open to proposed modifications.
  
  Stakes:
  
  $20 to the charity of the winner’s choice.
  
  Resolution criteria:
  
  By January 1, 2034, all of the following news outlets will report — as fact, not opinion or editorial — that the AI bubble has popped:

  • The New York Times

  • The Washington Post

  • The Wall Street Journal

  • Bloomberg

  • The Financial Times

  • The Economist

  • The BBC

  • The CBC

  • Le Monde

  To repeat, all of these news outlets have to report it. And they have to report it as news. Opinion columns and editorials don’t count.

  Moreover, no major news publication classified as “generally reliable” by Wikipedia’s editors — either in general factual news information or in the area of finance, technology, economics, or markets — will report that the AI bubble has not actually popped. (Let’s say a “generally reliable” major news publication needs to disagree with the above listed news outlets within a month of all the listed outlets reporting that the bubble has popped.)

  Moreover, if there is any survey of experts (e.g., economists, fund managers, institutional investors, financial analysts) on the matter by an organization generally considered reliable (e.g., Bank of America, which has conducted similar surveys in the past), at least 80% of respondents will agree the AI bubble as popped. If the survey finds that 79.9% or less of respondents agree, then I lose the bet. (Let’s say this survey can happen within a month of when all the listed outlets report that the AI bubble has popped.)

  Finally, all the following individual experts in finance or economics will agree that the AI bubble has popped:

  • Noah Smith (economics writer)

  • Tyler Cowen (economist)

  • Jared Bernstein (economist)

  • Ryan Cummings (economist)

  • Jim Covello (head of global equity research at Goldman Sachs)

  • Peter Oppenheimer (chief global equity strategist at Goldman Sachs)

  Let’s say after all the above listed news outlets report that the AI bubble has popped, these individual experts have a month to publicly weigh in. If they express no opinion within a month, they’ll be removed from the list.

  Explanation:

  Whether the AI bubble has popped is difficult to operationalize and could conceivably involve tracking many different metrics, such as valuations, AI investment, investor sentiment, and so on. This is a complex job even for an expert. Deferring to qualified “judges” is much easier, and much easier to get right than attempting to operationalize it oneself.

  The resolution criteria are designed to remove all doubt that, by some random fluke, a bunch of news outlets might report that the AI bubble has popped when it really hasn’t. If the AI bubble actually pops, it will probably be so clear and uncontroversial that someone will have no reservations about conceding it happened. (Conversely, if the AI bubble doesn’t pop, it will be equally obvious.)

  In fact, the resolution criteria are so slanted toward the person taking the other side of that bet that if literally everyone in the world except Tyler Cowen agreed that the AI bubble had popped, I would still lose the bet.
  
  The stakes are low because this bet is mainly just for fun and also to make a point.

  The bet’s timeframe is about 7 years because I read that many AI investments are amortized over 5 to 7 years. It’s hard to know for sure what exactly AI investors’ time horizons might be, but 7 years is probably long enough. Even if the AI bubble were to pop in 8 years rather than 6 or 7, that would still be a different enough outcome than what I’m expecting that I would say I’ve lost the bet fair and square.

  Again, I’m open to proposed modifications.

• Yarrow Bouchard 🔸 6 Oct 2026 20:12 UTC

  −5 points

  0 ∶ 0

  in reply to: TFD’s comment on: The sloppy re­search fuel­ling fears of su­per­hu­man AI

  I’m already aware of what you’re getting at, and this is already answered in the comment above. Asked and answered.

  Let me repeat for clarity.

  Are ML researchers competent at social science research involving human subjects? Not typically.

  METR’s paper is largely about measuring human performance in the form of human baselines. Could METR’s work on human baselines, including the guesstimation, pass peer review at a credible social science journal? Not likely.

  For one, you can’t just make up data (or “data”).

• Yarrow Bouchard 🔸 6 Oct 2026 19:21 UTC

  2 points

  0 ∶ 1

  in reply to: TFD’s comment on: The sloppy re­search fuel­ling fears of su­per­hu­man AI

  Something complicated about METR’s paper is that it combines social science research involving human subjects and machine learning research. I’m not sure whether anyone who worked on the METR paper has experience conducting social science research involving human subjects. There is no guarantee that typical ML paper reviewers would, either. (We can leave aside, for now, the question of whether machine learning is a science and whether ML journals are scientific journals. ML is a bit of an idiosyncratic field and that’s a whole other topic.)

  The METR graph is about comparing AI to humans quantitatively, so properly measuring human performance is key to whether the graph makes any sense. There are severe problems with how METR measured human performance, including that they largely didn’t bother to measure it at all. They just made up a lot of the numbers for human performance, as I discussed in this post. That’s not science, definitionally.

  I don’t think METR’s work with human subjects would likely pass peer review at a credible journal focused on social science research with human subjects, particularly if the version of the paper submitted disclosed (as Thomas Kwa’s January 2026 forum post that later turned into a blog post did) that the length of most of the longer tasks was just guessed rather than measured. I haven’t looked into what journals out there, if any, publish papers at the intersection of social science and machine learning.

  Something I didn’t discuss in the post is that it might be hard to find a journal or reviewers that have expert-level knowledge of both social scientific research with human subjects and machine learning research. I’m not sure what would be the best way for METR to address this difficulty in practice.

  (I also have no idea how in the ML field the review process differs, if at all, for conference papers versus journal papers. But that’s a side question and not the crux of the matter.)

• Yarrow Bouchard 🔸 6 Oct 2026 15:53 UTC

  −1 points

  0 ∶ 1

  in reply to: Jenna Parr’s comment on: The sloppy re­search fuel­ling fears of su­per­hu­man AI

  Ask Claude to refute its own fact check. And then refute that refutation. And then refute that refutation.

  Better use of time: read Nathan Witkin’s review.

  (But seriously, if an AI chatbot told you something you think might be true and important, take the time to manually fact check it and think about it yourself. If it still seems to check out, then comment about that thing using your own words.)

The sloppy re­search fuel­ling fears of su­per­hu­man AI

• Yarrow Bouchard 🔸 1 Oct 2026 21:59 UTC

  0 points

  1 ∶ 5

  on: Yar­row Bouchard’s Quick takes

  Three big unaddressed problems in EA 2.0:

  • corruption (i.e., the quid pro quo between FTX and Manifund, in which Austin Chen traded Caroline Ellison a job out of prison for cash obtained in the course of stealing customer funds)

  • cults (e.g., Leverage Research and Dragon Army Barracks, whose former members/​leaders such as Tyler Alterman and Duncan Sabien still post on the EA Forum)

  • misinformation (e.g., METR’s use of fabricated data in the famous time horizons graph)

  EA sadly seems sunk to me because dire problems keep mounting and most people seem to think doing nothing is the best solution. There is also a disturbingly loud chorus of voices saying no, no, corruption is actually good (it’s what having principles looks like), cults are actually good (it’s part of what makes the community unique), misinformation is actually good (we’re smarter than everyone else and need to save the world), which makes EA seem doomed. The force of immorality in EA is too strong and there is no equally strong force to stand against it.

  EA today is at best morally ambiguous, if not more evil than good. Things have gotten really rotten. No one is doing enough to clean it up. And there’s apparently no floor to how much worse it can get.

  One of my fears is that there will be an EA mass shooting or terrorist attack because this almost already happened with Stop AI, an EA-funded and EA-supported organization. Sam Kirchner, the Stop AI activist who threatened to do a mass shooting at OpenAI, is still missing. There is a warrant out for his arrest.

  Besides that, a disturbing number of people in the EA community seem sympathetic to far-right, illiberal, racist views (e.g., the Manifest afterparties hosted by Curtis Yarvin, who holds authoritarian and white nationalist views).

  So, really, that’s five big unaddressed problems in EA 2.0, including violent extremism and authoritarian, racist, far-right political views.

  People who want to focus on global poverty and global health and development should probably pour their energy into things not called “EA”, like effective giving. Or more mainstream humanitarian movements like Rotary International, or Hank and John Green’s fundraising through things like Project for Awesome.

  I still like the term “EA 1.0”, but it’s probably too cryptic and confusing for anyone not deeply familiar with effective altruism. “EA 1.0″ could possibly be used to describe something (e.g., in an FAQ), such as an effective giving org, but it would probably be too inaccessible to use in the name.

  It can be extremely painful to think about cutting your losses with EA, but we are called on in this life to make difficult decisions, to sacrifice, and to suffer when it’s the right thing to do. And, in the long run, that ultimately makes you feel more whole and right inside than ethically compromising and ethically compromising… There is a kind of joy in morality.

• Yarrow Bouchard 🔸 30 Sep 2026 22:53 UTC

  1 point

  0 ∶ 0

  in reply to: Maxwell Love’s comment on: Maxwell Love’s Quick Takes

  Pangram says this quick take is 51% AI: https://​​www.pangram.com/​​history/​​01232a86-af15-43a2-8e6a-047896c47336?ucc=2RxEGYyJeZr

  Is it?

• Yarrow Bouchard 🔸 30 Sep 2026 16:07 UTC

  2 points

  0 ∶ 1

  in reply to: harfe’s comment on: Yar­row Bouchard’s Quick takes

  Explained at length here: https://​​forum.effectivealtruism.org/​​posts/​​im59y2Au7zC5ofuKe/​​is-the-ai-industry-in-a-bubble

  GPT-6 Astra seems to fit a trend of diminishing marginal returns for R&D investment in LLMs. I don’t know about coding, which is the most economically important domain for LLMs — and I also don’t know about math research, which LLMs are helpful for but it’s hard to imagine is economically important — but for the regular sort of prose-based web search tasks I use LLMs for, GPT-6 Astra is not (to me) noticeably better than the GPT-5.x Chain of Thought models, which are themselves not noticeably better than o3.

• Yarrow Bouchard 🔸 29 Sep 2026 20:43 UTC

  2 points

  0 ∶ 0

  in reply to: harfe’s comment on: Yar­row Bouchard’s Quick takes

  What are the odds the NYT, WSJ, FT, and Bloomberg all report — as part of their factual news reporting, not simply an opinion column or editorial — that the AI bubble has popped and the AI bubble has not actually popped? Surely those odds are minimal.

  You could add in other news outlets like the Economist, the Washington Post, the BBC, and the CBC, or possibly other kinds of external judges such as economists or financial analysts (either surveys thereof or reputable organizations or individuals) who would all have to agree.

  In terms of the definition of a financial bubble, the Wikipedia definition works perfectly well: https://​​en.wikipedia.org/​​wiki/​​Economic_bubble

  Exactly operationalizing what counts as the AI bubble popping is a task that can be deferred to judges. Requiring unanimous agreement among the selected judges means their verdict is unlikely to be wrong. (You can always spend the time trying to operationalize something like this yourself, but it’s time-consuming and there’s a meaningful risk you look back on your operationalization and realize you screwed up. I prefer the simplicity and elegance of just deferring it to judges. This is just a symbolic $20 bet for charity and honourable people can acknowledge if they won a bet unfairly.)

  If you think a unanimous verdict among all the aforementioned judges and even more besides is likely to be wrong, then, yeah, this bet isn’t for you. I’m always open to counteroffers, though.

• Yarrow Bouchard 🔸 27 Sep 2026 2:41 UTC

  3 points

  0 ∶ 2

  on: Yar­row Bouchard’s Quick takes

  Will anyone take the other side of this bet? By January 1, 2034, the New York Times, Wall Street Journal, Financial Times, and Bloomberg will all agree the AI bubble has popped. (As in, they will all report it as news, not just publish an opinion column saying so.)

  The stakes: $20 to the charity of the winner’s choice.

  Why this bet? Mainly just for fun. But also to make a point.

  I chose 7 years as the time horizon for the bet because I read that 5 to 7 years is the time horizon for many AI investments.

  My hunch is that the AI bubble will pop within 3 years, but who knows. Markets can keep lending and investing in unprofitable ventures for a very long time.

• Yarrow Bouchard 🔸 27 Sep 2026 2:16 UTC

  3 points

  0 ∶ 0

  on: Ex­is­ten­tial Risk Is An Ex­traor­di­nary Claim

  Some may take the claims of a few experts or their own “vibes” as evidence that we should be very worried about existential risk, but it’s important to remember that humans have a natural proclivity towards believing that their generation could be the last. (According to Pew, “39% of adults believe that humanity is ‘living in the end times.’”)

  On this topic, I can’t recommend enough a talk given by sociologist and transhumanist James J. Hughes called “Avoiding Millenialist Cognitive Biases”: https://​​podcasts.ox.ac.uk/​​avoiding-millennialist-cognitive-biases?video=1

  Hughes also wrote an essay on the same topic in the anthology Global Catastrophic Risks.

• Yarrow Bouchard 🔸 27 Sep 2026 2:07 UTC

  6 points

  1 ∶ 0

  in reply to: ASB’s comment on: Are Civ­i­liza­tional Re­fuges Dead?

  E2H = environment-to-human

  https://​​defensesindepth.bio/​​prioritizing-environment-to-human-biological-threats/​​

• Yarrow Bouchard 🔸 27 Sep 2026 1:45 UTC

  4 points

  0 ∶ 0

  in reply to: Vasco Grilo🔸’s comment on: Com­mu­nity (Part 1: From FTX to Man­i­fund) | Reflec­tive Altruism

  Hi Vasco. I think we know for sure the probability that people who’ve been convicted of crimes will reoffend is far greater than for people who have never committed a crime. We don’t have to ask what if.

  Beyond the risk of Caroline Ellison committing illegal or unethical acts at Manifund, and beyond the risk of Austin Chen encouraging such acts or covering them up — which are both serious, rational concerns — there is also the corruption aspect. This hire is explicitly a reward, a show of thanks, an act of loyalty on Austin Chen’s part to reciprocate the crimes committed by Caroline Ellison.

  The corruption aspect would still exist if the hire were for a non-financial role or if the reward came in some other form, such as cash given by Austin to Caroline. Usually people don’t admit corrupt intent out loud but Austin took the unusual step of writing a public confession.

  In my view, the effective altruist community is still underreacting to this even though the reaction has been strong and negative. This is brazen financial corruption and outright encouragement of financial crime. The underreaction of the EA community makes it seem like everything said in defense of EA about the crimes at FTX was a lie. The corruption is now happening in plain sight. No one can pretend they don’t know what’s happening. Where is the accountability? Where are the consequences? The Centre for Effective Altruism’s sanction of Manifund is a good first step, but it’s not enough.

  I don’t know what the wider world’s reaction will be if or when the news spreads of just how tolerant the EA community is of financial corruption. But I can say for me, my trust in the EA community has been decimated by this. I personally no longer trust the EA community to not engage in unethical (or even illegal) financial conduct, such as self-dealing with money from EA charities, or covering up improper acts committed by others at EA charities. When a community is so tolerant of such brazen corruption, how can I trust them?

  As Maya Angelou is reported to have said, “When someone shows you who they are, believe them.”

• Yarrow Bouchard 🔸 19 Sep 2026 15:07 UTC

  9 points

  2 ∶ 0

  in reply to: Tobias Häberli’s comment on: Caroline Elli­son has joined Manifund

  As recently as April 2025, Austin defended Sam Bankman-Fried on a podcast. In response to a host’s assertion that Sam Bankman-Fried “did [things] badly and hurt a lot of people”, Austin said:

  I don’t know, still debatable. I don’t know how much we should go into the object level here, but everyone made their money back. Yes, some things broke for that to be able to happen, but all the people got 120% of the value of the crypto holdings at the time...

  He followed-up by saying:

  ...I think we should free SBF and put him to work.

  Continuing on this thread a bit later on, Austin said:

  ...yes, [Sam Bankman-Fried] made like big mistakes, but there were things that he did really right. Oh, I think this is a piece I talked about a lot in my post, and is important to me.

  Like in some sense Sam made a lot of really, good decisions and maybe a few really bad ones. And I think it’s easy to focus on oh, those bad decisions, I would never have made those bad decisions. But I think I have a lot more sympathy for somebody like Sam because I am sometimes in a position making a lot of decisions too.

  And I think oh, the hard thing is not to avoid the couple of bad decisions. The hard thing is to actually have figured out what all the right decisions you could have made, were such that you are in the position to then have to think hard about the one or two bad decisions too. I’m like, many people never get to that place for the goals that matter to them. And I think it is like something somewhat instructive to learn from the perspective of, Sam Bankman-Fried having won a lot at something. Of course.

  Austin goes on further in that vein. You can listen or read the full transcript.

  Austin’s post about hiring Caroline reads, to me, as if he’s still downplaying the extent of crime and wrongdoing that FTX and Alameda executives committed, similar to what he said on this podcast in April 2025.

  My interpretation is that Austin’s views on this topic have not fundamentally changed since April 2025.

  For context, Austin also defended Sam Bankman-Fried in November 2022, lambasting people in EA who spoke out against him:

  EA leaders have posted denunciations of fraud, and distanced themselves from Sam; the most upvoted all-time EA Forum post is a community condemnation; the entire Future Fund team up and resigned.

  Austin wrote that this showed a lack of “loyalty”:

  ...from EA folks, this behavior strikes me as cowardly, coldhearted, opportunistic, bandwagon-y, two-faced and distasteful. … Whatever happened to loyalty? To supporting those who have helped you in the past?

  To my knowledge, he has never retracted these comments. On the contrary. On the April 2025 podcast, he reiterated this same point, saying:

  I think if you abandon people who have what I viewed as earnestly try to support you and your movement. If you just turn tail on them, be like, “oh yeah, no, actually SBF he was not acting according to our principles and we do not endorse this” and we like give up on all of the stuff that he did. Then when some future billionaire is looking at your movement, what will they think?

  And:

  Maybe I am a person who someday hopes to be a billionaire and I was looking at the actions of people l in this space and be like, do I want to be their friends, their allies? If I feel like in like cases when things go badly, they will just not stand by me.

  My interpretation is that Austin’s view is some combination of 1) crime and wrongdoing did not occur at FTX or Alameda to the extent that most people believe and 2) some amount of crime and wrongdoing is good or acceptable in the service of Austin’s goals and EA’s goals. He both downplays the extent of what happened at FTX and also makes an argument along the lines of you can’t make an omelette without breaking a few eggs.

• Yarrow Bouchard 🔸 19 Sep 2026 14:48 UTC

  −3 points

  2 ∶ 2

  in reply to: Aaron Gertler 🔸’s comment on: Caroline Elli­son has joined Manifund

  ...I generally don’t think it’s unethical to make private hiring decisions based on whatever criteria you want.

  It’s unethical for Austin to give someone a job out of prison as a reward for her financial crimes, as a way of saying thanks or repaying a perceived debt (as he put it) because her crimes personally benefited him and his organizations. That’s just straight-up corruption. That rewards and incentivizes, not to mention endorses, financial crimes.

  This is a quid pro quo in some sense. It may not be a quid pro quo in the sense that this was the explicit arrangement from the outset, but it’s a quid pro quo in the sense that the hire was made at least in part in exchange for the $1.5 million that Manifold and Manifund received from FTX.

  The precedent this sets up is that if you commit crimes that benefit Austin, he’ll reward you when you get out of prison (if you get caught).

  In terms of what’s illegal or not, I don’t know (ask a lawyer, not me), but this is absolutely unethical.

  ---

  Edited to add:

  Given this precedent, Caroline has reason to believe that, while working in a financial capacity at Manifund, if she commits crimes or unethical acts that benefit Manifund, Manifold, or Austin, then Austin may cover her, cover up for her, defend her, not report her, etc. Austin has again and again emphasized — publicly! — that “loyalty” and “winning” are more important than the law or business ethics. At this point, Caroline has reason to believe Austin has her back even if what she does is illegal or unethical, as long as it directly benefits Austin and his organizations.

• Yarrow Bouchard 🔸 18 Sep 2026 18:18 UTC

  2 points

  2 ∶ 2

  in reply to: Rachel N. ’s comment on: Caroline Elli­son has joined Manifund

  Bringing it back to EA: I understand the argument that Manifund and Austin do not represent most of the EA community. But it’s reasonable that outsiders might think they do if EA orgs don’t take steps to distance themselves. I’m glad to see CEA doing so.

  If the EA community endorses or even just tolerates the decision and allows Manifund, Austin, and Caroline to participate in the community, e.g., through events hosted by EA orgs like the CEA (which as far as I know has not yet been ruled out), then in that sense Manifund, Austin, and Caroline will represent the EA community.

  The sanction against Manifund that Zachary Robinson announced yesterday is a great start. But there are additional straightforward steps that can and should be taken to dissociate effective altruism from the financial corruption Austin just committed and the financial crime Caroline committed and served prison time for.

  If the EA community can’t agree that financial corruption or financial crime is an ethical red line and won’t take decisive action to discourage it, the EA community should be correctly understood — by everyone, including both community insiders and outsiders — as ethically compromised.

  It’s not really about how things look. It’s about how things are. Caroline really did steal money. Austin really did reward her for it.

  The EA community should not tolerate this:

  

• Yarrow Bouchard 🔸 18 Sep 2026 18:10 UTC

  −8 points

  1 ∶ 4

  in reply to: Lorenzo Buonanno🔸’s comment on: Caroline Elli­son has joined Manifund

  Answers:

  1. There are no controls because Manifund is only Austin Chen and Caroline Ellison, so this is 100% trust-based. I am concerned about the risk of Caroline Ellison committing financial crimes again. I take it many others are concerned about this risk as well. Between here and Austin’s Substack post, some comments have explicitly expressed concern that this is an unnecessary risk, e.g., noting how high recidivism rates are even when recidivism seems completely irrational. Beyond that, there is also the corruption aspect, which is not just a risk or a possibility — the corrupt quid pro quo has already occurred. That horse is out of the barn.

  2. What an off-base thing to say. Caroline just got out of federal prison for one of the largest financial crimes in recent American history. Austin said one of the motivations for the hire is to reward her for the money FTX gave to Austin’s organizations in the course of committing these crimes. It looks completely insane and corrupt because it is completely insane and corrupt. What are you talking about?

  3. What?? The only way to get top talent for financial roles is to hire people who have been convicted of financial crimes?? What are you talking about?!

  4. I don’t know what specific people you’re referring to or the mistakes (as you put it) they made. So, I can’t comment.

  5. People are being quite honest, open, and direct about the fact that trust and reputation are a big concern here, among other concerns such as corruption and increasing the risk of future financial crimes involving EA money. Who is pretending?

  The philosopher David Thorstad’s recent blog post methodically goes through the multiple different concerns with this hire, point by point. David has a commendable amount of patience in explaining things like this.