I'm experimenting with "norms-pledges" to help reduce forum anxiety. Maybe it could be a good social technology IDK. Click [Show More] to read them all:

🕊 Fresh Slate After Disagreement Pledge: I hereby pledge that if we disagree on the forum, I will not hold it against you. (1) I will try not to allow a disagreement to meaningfully impact how I treat you in further discourse, should we meet in another EA Forum thread, on another website or virtual space, or IRL. I know that if we disagree, it doesn't necessarily mean we will disagree on other topics, nor does it necessarily imply we are on opposing teams. We are most likely on the same team in that we both wish to have the most good done possible and are working in service of finding out what that means. (2) Relatedly, I pledge to not claim to know what you believe in future, I can only confidently claim to know what you wrote or believed at a given time, and I can say what I think you believe given that. I know that people change their minds, and it may be you or me who does so, so I understand that the disagreement may not even still stand and is not necessarily set in stone.

👨‍👩‍👧‍👦 No Gatekeeping Pledge: I hereby pledge that if I am seeking a collaborator, providing an opportunity, or doing hiring or anything akin to hiring, and you would otherwise be a top candidate if not for the following, I will try not to gatekeep: (1) If  an opinion you've shared or broken-norm you've done (on the EA forum or elsewhere) is relevant in a potentially negative way to our collaboration, that I will ask you about it to gain clarity. I will not assume that such an incident means you will not be  suitable for a role. I will especially try hard not to make assumptions about your suitability based on old or isolated incidents, or if your digital footprint is too small to get a good picture of who you are and how you think about things. (2) I will not penalize based on someone being a social or professional newcomer or being otherwise unknown to me or my colleagues. If the person is a top candidate otherwise, I will do my due diligence to determine cultural fit separate from that. 

🤔 Rationalist Discourse Pledge: (1) I hereby pledge to try to uphold rationalist discourse norms as presented here and here, and comedically summed up here.

🦸‍♀️Preferring My Primary Account Pledge: (1) I hereby pledge that this is my main EA Forum account. I will never use multiple accounts to manipulate the system, as by casting multiple votes or stating similar opinions with different accounts. (2) I also pledge that, although I can't be sure what comes, I strongly intend to not use an anonymous or different account (alt or sockpuppet), or any account other than this, my primary account. I pledge that I am willing to take on some reputational risks on this, my primary account, in service of putting truth, transparency, integrity, and a complete narrative over my own anxiety, and to give ideas I think are worth advocating for the best chance at adoption. Therefore I pledge that I will not use an alternate account out of general anxiety around personal or professional retribution or losing clout. CAVEAT 1: I reserve the right to use an alt account in cases where *specific* retribution or other  danger can be expected in my  particular instance. As example: I reserve the right to use an alt account out of concern about riling up a suspected stalker, specific known bad-faith actor, specific known slanderer, etc. CAVEAT 2: I also reserve the right to use an alt account for the benefit of others. Example:  in cases where revealing my own identity would reveal the identity or betray the privacy of some other party I am discussing. 

🙇‍♂️Humility in Pledging Pledge: I hereby pledge that I take these pledges for my own self-improvement and for altruistic reasons. It's okay to disagree that pledges are useful and important for you. (1) I don't expect others should necessarily take a norms pledge. I believe the pledges only work if people take them after deep consideration, and I don't expect I can know all the considerations for others' situations. Therefore I understand there may be situations that it is actually right that a user avoid taking a pledge. Therefore I will not judge others for not having taken a pledge, including that I will not dismiss other's character if I see other accounts without a pledge. (3) Additionally, I don't presume that others not taking a pledge means they would even necessarily act differently than that pledge would imply. I don't assume their intentions are even different from mine. Perhaps a person is new to the idea or just trying to protect their energy by not opening themselves to criticism. (3) I won't automatically dismiss a user's reasoning if I see the user violating norms pledges I've made. I still will give their claims a chance to stand on their own merits. (4) If you see me violating a pledge I've taken, I will always appreciate if you bring it up to me. 


Sorted by New


I've read this comment a few times, and my brain goes "???" whenever I get to your last clause: "I also had quite a strong reaction that nobody seemed to be acting on all of these warning flags"

I just don't get it in a way that connects to my reading of the article. What are "all these warning flags" and what counts as "inaction"? I don't want to say your take is wrong because you are sort of sharing feelings, but like.... according to the article, ex-Alameda employees don't seem to think that those flags were warning flags for the massive fraud and crash-and-burn failure that was to come. And re: inaction, the article says CEA did an internal investigation in 2019 (it drops the info kinda randomly. As you say, the article isn't well-optimized to come away with an understanding of the details). And idk what new warning flags came after 2019, I'm not seeing any in the article.

I mostly like your comment, but I'm also left wondering... Do you know things not in the article? Did I miss something? [Is this just a "vibe" we will disagree on regardless?] I can't quite reconcile your take.

[Edit: I had been thinking about asking this over DM for a couple days, but now that this post is no longer an active topic, I figured, "what the hay, ask it in thread". However you can answer over DM if you prefer, or ignore cuz the post is giving dying breaths, np.]

I vouch for Monica as a kind, intelligent person who has indeed focused much of her career around helping people with autism :) I haven't worked with her myself, but we are in the same local EA community. 

If you have autism, (with or without formal diagnosis), there's nothing to lose by reaching out, and a lot you might gain :)

Big improvement on the left sidebar. Good job all 👍

That makes sense. It is really shocking. I agree on blaming regulators [although I don't give others a pass].

[I think a section on regulations def belongs from the POV of improving world models too. Before I added my long thinking-out-loud footnote, I didn't realize just how much it all points at regulators as the original permissiveness break.]

Yeah my thoughts exactly. Or, it doesn't send a big signal to non-finance people. But like, I think it should send a signal to people in finance, eg the auditors. That FTX should have been able to afford a different service and yet didn't. Or maybe, idk, should have revealed their internals for different certification (better than GAAP cert idk, I know nothing). I just think it should have raised flags for the auditor. If someone is enlisting you for purposes of increasing trust but they are clearly not doing their damnedest, according to their abilities, to ensure that trust is accurate. The US consumers can't be expected to know the difference, but the auditor should. I think.

Yeah, turns out there was not only sloppiness here though. Like Enron, things were labelled to look much better than they really were. Like, stocks of coins held in such volumes and accounted as their current exchange price such that it looked like billions in net worth, IIRC, but of course if they would have sold the coins, the value of the coins would have plummeted far before they got through their sale order, so there is no way they could have expected to gather that much USD for the coins they held. This might be normal for stocks valuation? But my impression from the tone is that it was handled differently/worse. Maybe there were other things too, tbh I didn't look very closely. Plus there was SBF's backdoor (that was real right?) and I'd bet other weird money movements I assume he has noticed by now.

Glad you liked it. Perhaps I wasn't clear in purpose though. My point was not to talk about payouts, but to explain how things like this can happen. Because it "violates model of how the world works". [Edit: I cut stuff from here and expanded on it in footnote above]

I'm just saying there are systemic reasons why the fiasco got as far as it did.

[Edit: Ah this is one of those times I might be being dramatic, but I may as well say] I'm a bit sad to hear "fishing expedition" to this. But [so many EAs didn't feel similarly] about EA leaders taking some of the blame. I didn't want to bring EA into this particular comment, but gee. The actors I've named here had like 1000x the likelihood of knowing what was going on and being negligent or otherwise at fault somehow than non-Alameda EAs did. It's a bit flippant to call it a fishing expedition to go after others, if it is worth doing, and when our community has just spent so much time talking about EA fault. dies a bit inside on behalf of EA  Tbh they seem like reasonable suits to me in general, and I hope disincentivize something like this from happening again :/

Eh, I still agree with Ben here. He said they are the worst financial documents he has ever seen, including Enron. And given he's the guy who steered Enron after bankruptcy, that's a concrete claim. And as the documents will surely all be submitted, he'd be putting his neck on the line to say such things if it weren't clearly like the most egregious thing you could imagine. shrugs :/

Somehow they were doing this while having audited financials, passing due diligence from major investors, etc.? And Sam was supposedly a great fundraiser but was circulating a balance sheet with a $8B line item for “hidden poorly labeled account”? I would find it pretty helpful for someone to explain what actually happened here because this violates my models of how the world works.

Mine too, so I went digging. All in all, one can argue (and lawyers are) that there were a lot of enablers. Certainly, people trying to actively dodge being noticed as they violate standard models, is a predictable (but not necessarily pin-pointable at the time) way our models end up failing us. 

For talk about audits specifically, there's this and this. Essentially, "the firm's auditors weren't tapped to look into internal controls at FTX, and auditing the internal workings of a company isn't a [legal] requisite for private corporations." Of course plenty of people feel that doesn't hold water, and a lawsuit by FTX customers is pending.
[Edit: A suit may be reasonable because audit firms know well they are commissioned to literally prove financial safety, and the audit was used in FTX's self-promotion. That FTX was still very unsafe might prove there is negligence in the firm's business model. If I try to make this fit my model of the world, I get a thought like "corporate greed incentivizes taking on clients that want useless/partial audits that end up being no better than shams, and you look the other way as you assist them in their likely sham". I reflect more on this in the footnote which you don't have to read but>>[1]]

For talk about investors, there is this Feb 23rd piece on a major lawsuit against a few different VC banks who helped out FTX. That piece is frickin nuts just by virtue of the amount of info they fit into a mere 5 paragraphs. Here's one quote to inspire a read:

The suit alleges that some of world’s largest venture capital firms, including Sequoia Capital, SoftBank Group, and Thoma Bravo, learned through due diligence that FTX was a fraudulent scheme, but nevertheless incited the Fraud with billions in necessary capital, provided guidance and other support critical to the Fraud, and publicly promoted Bankman-Fried and FTX to keep the Fraud concealed until FTX could go public or cash out in a private sale

And of course there was internal assistance. I'd bet there were both (1) employees who knew what they were doing and (2) employees just following direction without realizing they were committing crimes or aiding in them. 

When it comes to improving models of how the world works, Zvi's piece had some good discussion of this (and much more!). It's long, but very worth reading even months later. Here's a section on VC:

If you are a VC investor or taking VC money, the optimal amount of fraud, from your perspective, is not zero. You are more excited to invest if you suspect a fraud, those kinds of founders make good and make dreams real for you. So what if sometimes it all blows up? This is a game of hits for you, and you are much more worried about being asked why you didn’t invest in FTX than you should worry about people asking why you were. This goes double for crypto. 

Notice this VC ready to invest in SBF again if he asked. Which, from a pure EV perspective, sure why not, same as they invested in Adam Neumann. 

Therefore, I do not view ‘the VCs didn’t catch this’ as much of a justification. They are not supposed to catch it. It is not their job to catch it. I mean, they are supposed to do some things to catch this level of fraud, demanding voting shares and a company board in which they have a seat and doing proper audits, but they got none of that, because [SBF] needed to not allow it and the pitch was so good otherwise they went along (or, they took this to mean ‘this is a fraud’ and invested anyway thinking they were not the sucker, also plausible). [If] SBF said ‘you can’t check for fraud, invest or don’t’ then they are doing what they do, as much as we might hate that. Better to make that common knowledge. [emphasis mine]

The institutional traders are different. They face a different risk profile. If the exchange blows up in 10% of years that is a real drag on returns, whereas a VC expects 80% or more of their investments to go to zero, fraud or otherwise. Why did they trade?

Some of them Did Not Do the Research, no doubt. Others likely decided it was worth the risk. If FTX is an easy place to make money, because trading is very good against Alameda and generally they treat customers great, and they pay interest on USD and BTC, it is not difficult to imagine a decision that the blowup risk is worth taking, at least for some portion of one’s bankroll. Or one can say that if FTX goes down everything is terrible anyway, it’s already systemic risk, so assume it won’t fail. 

Another play is to hedge the risk with a short position elsewhere, and put your leveraged longs at FTX since if FTX goes broke you would have gotten liquidated at Binance anyway.

[Note: There could also be a section about lack of regulation here. As commented on here. From a "how the world works" POV, it is the mother of all permissiveness that allowed the rest, especially the useless audits. Global heuristic = Pretty nuts how much can go wrong with a regulation gap. But FTX also maybe would have found  away around/already did break regulations so idk] 

Anyway, just thought that all might interest you. Thank you for sharing your insights. Really useful stuff.

  1. ^

    Interestingly, now the company has stopped offering crypto audits due to "changing market conditions". And I was going to say "good of them to notice that the bare legal minimum is not enough for one of the most volatile, least regulated assets the world has ever seen, eyeroll." But actually that auditor firm is starting up crypto audits again under a different name, sigh. It appears that they only stopped offering crypto audits because some non-crypto clients of the audit firm put pressure on them to do so. Those non-crypto clients felt that the auditor being known to audit crypto, might reduce the trust their potential clients/users end up having in their own commissioned audits. So it appears that, to make sure they can catch the most non-crypto and crypto audit contracts without one segment compromising the other, the auditors are just splitting them under two companies. I didn't see any mention of raising the crypto audit standards.

    Additionally, through this process. I found mention a couple times that there are other auditing firms who do a much better job, "The big 4", and it is sort-of business wisdom to downgrade trust in audits that aren't from them. I was very surprised at this! There are auditors that business-people know to trust less? Then why do they exist? I guess because most consumers don't know enough to downgrade trust? One has to wonder if the big 4 would have put their stamp of approval on FTX or even many of this audit firm's non-FTX-clients. Probably not, right? And it makes you wonder why companies would go to these known-worse-auditors, especially if they can afford the best auditing like FTX should have been able to, if they don't have something to hide. If the goal is consumer trust, and a company doesn't have something to hide, that company should go to the best auditors, those that in-the-know-consumers trust the most, right? And surely a "worse" audit firm like this one would realize that is how the selection effect cookie crumbles. So lesser-known auditors should expect they will likely get a slew of clients who are nowhere near as reputable as those who hire the big 4 audit firms. 

    Sooo I think it's easy to argue that the minor audit firm(s) are negligent or intentionally turning a blind eye somewhere in their business model. But it's more a problem of incentives of corporate America and low federal standards for private corporations allowing those incentives to play out in auditing contracts, than it is a problem of a single actor. Even though the single actor/auditor might know damn well that they will end up giving a scammer cover eventually, maybe even frequently among their really wealthy clients like FTX who could have afforded better-known services. It appears to be part of the business model, chronic not acute. And as long as the private corporation and auditor are following the law in reviewing the bare minimum, oh well?

I think that most of your comment is reasonable, so I'm only going to respond to the second-to-last paragraph. Because that is the bit that critiques my comment, my response is going to sound defensive. But I agree with everything else, and I also think what went on with my original comment leads back into what I see as the actual crux, so it's worth me saying what's on my mind:

But I do think that confidently asserting that the only thing Jacy did was "ask some people out over FB messenger" is likely inaccurate, and it is important to track that. It might be accurate to say "the only thing he has been publicly accused of is asking people out" or "the only thing he has admitted to is asking people out" or "No one has provided any proof he did anything beyond ask people out",  but none of those are the same as "the only thing he did is ask people out".  

I have long ago edited the original comment where I wrote that. I didn't change that particular wording because I wrote the original on mobile (which I deeply regretted and am now incredibly averse to) so I didn't have fancy strikethrough edit features, even when I tried on PC (I didn't realize it worked like that). Without strikethrough ability, I thought it would be epistemically dishonest to just edit that sentence. Instead I promptly, right after that sentence, told people to make their conclusions elsewhere in a way that I feel clearly tells readers to take that part with a grain of salt. All in all I edited that comment ~5 times. I don't have the spoons to re-edit again given I think it's fine. 

More importantly, the transparency of info is obviously a problem if someone like me who usually tries to be pretty airtight on EA Forum things had to edit so much going back and forth from "here's a thing" to "maybe he did worse" to "maybe he did less" to "maybe he did worse" again. That's not okay. And now I feel like I'm getting punished for trying to do what no other outsider of the case was willing to try to do (that I saw)... figure out the ground truth [and what it means for EA behavior] publicly.

Honestly trying to figure out what happened regarding Jacy was a heckin nightmare with people coming out of left field about it after each correction I tried to make, including over DM (again not publicly), and giving multiple comments to comb through on multiple other posts and with their own attached threads. It's good people chimed in sharing the existence of different pieces of discussion/info that I'd guess hardly any single person knew every single bit of, but damn, I have to be honest that I'm now really frustrated about what a nightmare it was. I was trying to do a public service and it was a huge waste of time with little to glean for certain. [And some of the more interesting bits are not public and I feel very, very weird about that, even saying that I now know of (know of, not know for certain) stuff others don't and can't find out about (I can't even doublecheck myself).] 

Was that always the expected outcome just lurking underneath the surface? If so then why would people judge SFF? I'm no longer surprised SFF just granted tbh. They saved themselves the time I wasted. I no longer expect any single person to get it right and I see that as a problem worth talking about because that will lead to either (1) actually-abusive people getting involvement sooner which is a safety risk, or (2) appearing-abusive-but-actually-non-abusive people getting involvement sooner which is a PR-risk and comfort risk. 

I apologize for fucking up. I am now frustrated at myself for even trying. But if people other than me care about my messed up original comment they need to look at the systems because other people will fuck up as I did. It just won't be public til after the decision is made, if ever. And you won't get to correct them as they make their fumbles along the way. 

Load more