What should the EA community learn from the FTX / SBF disaster? An in-depth discussion with Will MacAskill on the Clearer Thinking podcast

Great comment. 

Will says that usually, that most fraudsters aren't just "bad apples" or doing "cost-benefit analysis" on their risk of being punished. Rather, they fail to "conceptualise what they're doing as fraud".

I agree with your analysis but I think Will also sets up a false dichotomy. One's inability to conceptualize or realize that one's actions are wrong is itself a sign of being a bad apple. To simplify a bit, on the one end of the spectrum of the "high integrity to really bad continuum", you have morally scrupulous people who constantly wonder whether their actions are wrong. On the other end of the continuum, you have pathological narcissists whose self-image/internal monologue is so out of whack with reality that they cannot even conceive of themselves doing anything wrong. That doesn't make them great people. If anything, it makes them more scary.

Generally, the internal monologue of the most dangerous types of terrible people (think Hitler, Stalin, Mao, etc.) doesn't go like "I'm so evil and just love to hurt everyone, hahahaha". My best guess is, that in most cases, it goes more like "I'm the messiah, I'm so great and I'm the only one who can save the world. Everyone who disagrees with me is stupid and/or evil and I have every right to get rid of them." ^[1]

Of course, there are people whose internal monologues are more straightforwardly evil/selfish (though even here lots of self-delusion is probably going on) but they usually end up being serial killers or the like, not running countries. 

Also, later when Will talks about bad applies, he mentions that “typical cases of fraud [come] from people who are very successful, actually very well admired”, which again suggests that "bad apples" are not very successful or not very well admired. Well, again, many terrible people were extremely successful and admired. Like, you know, Hitler, Stalin, Mao, etc. 

Nor am I implying that improved governance is not a part of the solution.

Yep, I agree. In fact, the whole character vs. governance thing seems like another false dichotomy to me. You want to have good governance structures but the people in relevant positions of influence should also know a little bit about how to evaluate character. 

1. ^{^}

   In general, bad character is compatible with genuine moral convictions. Hitler, for example, was vegetarian for moral reasons and “used vivid and gruesome descriptions of animal suffering and slaughter at the dinner table to try to dissuade his colleagues from eating meat”. (Fraudster/bad apple vs. person with genuine convictions is another false dichotomy that people keep setting up.)

Quote: (and clearly they calculated incorrectly if they did)

I am less confident that, if an amoral person applied cost-benefit analysis properly here, it would lead to "no fraud" as opposed to "safer amounts of fraud." The risk of getting busted from less extreme or less risky fraud would seem considerably less.

Hypothetically, say SBF misused customer funds to buy stocks and bonds, and limited the amount he misused to 40 percent of customer assets. He'd need a catastrophic stock/bond market crash, plus almost all depositors wanting out, to be unable to honor withdrawals. I guess there is still the risk of a leak.

I don't think we disagree much if any here -- I think pointing out that cost-benefit analysis doesn't necessarily lead to the "no fraud" result underscores the critical importance of side constraints!

(This comment is basically just voicing agreement with points raised in Ryan’s and David’s comments above.) 

One of the things that stood out to me about the episode was the argument^[1] that working on good governance and working on reducing the influence of dangerous actors are mutually exclusive strategies, and that the former is much more tractable and important than the latter. 

Most “good governance” research to date also seems to focus on system-level interventions,^[2] while interventions aimed at reducing the impacts of individuals are very neglected, at least according to this review of nonprofit scandals: 

It is notable that all the preventive tactics that have been studied and championed—audits, governance practices, internal controls—are aimed at the organizational level. It makes sense to focus on this level, as it is the level that managers have most control over. Prevention can also be implemented at the individual and sectoral levels. Training of staff, job-level checks and balances, and staff evaluations could all help prevent violations with individual-level causes. Sector-level regulation and oversight is becoming common in many countries. We, therefore, encourage future research on preventive measures to take a multilevel perspective, or at least consider the neglected sectoral and individual levels.
 

Six years before the review quoted above, this article called for psychopathy screening for public leadership positions (which would have represented one potential approach to interventions at the “individual level,” to adopt the terminology of the review quoted above).^[3]

This leads me to wonder: what are the most compelling reasons for the lack of research (so far) on interventions to reduce the impact of dangerous actors, and which (if any) of these reasons provide strong arguments against doing at least some research in this neglected area? I think there are lots of possible answers here,^[4] but none of them seem strong enough to justify the relative lack of research on this area so far (relative to the scale of the problem).

1. ^{^}

   Here’s a quote from the episode (courtesy of Wei Dai's transcript) demonstrating this claim: 

   [Will MacAskill:] There's really two ways of looking at things: you might ask…is this a bad person - are we focusing on the character? Or you might ask…what oversight, what feedback mechanisms, what incentives does this person face? And yeah, one thing I've really taken away from this is to place even more weight than I did before on just the importance of governance, where that means the, you know, importance of people acting with oversight, with the feedback mechanisms and you know, with incentives to incentivize kind of good rather than bad behavior…

   I agree that all these aspects of governance are important, but disagree that working on these things would entirely protect an organization from the negative impacts of malevolent actors.

2. ^{^}

   To be clear, I am glad people are working on system-level solutions to low integrity and otherwise harmful behaviors, but I think it would be helpful if it wasn’t the *only* class of interventions that had substantial amounts of resources directed towards them.

3. ^{^}

   Interestingly, one of the real-life cases Boddy refers to in support of his argument is the Enron scandal, a case which was also covered in the book Will MacAskill was talking about, Why They Do It.

^{^}

Here are some of the reasons I’ve already thought about (listed roughly in order from most to least convincing to me as a reason to be pessimistic about this approach to risk reduction): potential lack of tractability; lower levels of social and political acceptability/feasibility; lack of existing evidence as to what methods work, to what extent, and in which contexts; and perhaps a perception that the problem (of dangerous actors) is small in scale. I’d be interested to know which (if any) of these reasons are the most important, and if there are other considerations I’m overlooking. Overall, despite these reasons against working on it, I still think this area is worth investigating to a greater extent than it has been to date.

Show all footnotes

Interesting discussion. In the interview, MacAskill mentioned Madoff as an example of the idea that it’s not about "bad apples." ^[1] Giving Madoff as an example in this context doesn’t make sense to me. But maybe MacAskill was meaning to say that it's not about "bad apples that are identified as such before/at the time of their fraud"? That would be the only interpretation that makes sense to me, because Madoff sounds like he really was a "bad apple" based on the info in Why They Do It.

Here's what Soltes says about Madoff in Why They Do It (quoted from the audiobook, with emphasis added):
 

[Madoff] cavalierly remarked, "The reality of it is my son couldn't stand up amongst the pressure anyhow, so he took his own life." In the many hours of conversations I had with Madoff, this statement stood out for its callousness. A father [who] couldn't understand the impact that his actions had on his own son...

...Madoff remains dispassionate even about the circumstances that are of the greatest significance. In September 2014, a colleague emailed me news that Madoff's second son, Andrew, had just died of cancer. As I was beginning to read the article, my office phone rang. I picked it up and was surprised to hear Madoff on the line. 

He had heard the news of his second son's death on the radio, and asked if I could read the obituary to him. Shaken by the fact that a father had called me to convey news of his son's death, I turned my attention to describing the news in the most compassionate way I could. I wasn't a professor or a researcher at that moment, just one person speaking to another. 

I read him a writer's article about his son's death. When I reached the end, I was at a loss to know what to say. Instinctively, as we often do when hearing of a death, I asked him how he was doing. 

Madoff responded, "I'm fine, I'm fine." After a brief pause, he said that he had a question for me. I thought he might want me to send a copy of the obituary to him or deliver a message on his behalf to someone. It wasn't that. Instead, he asked me whether I'd had a chance to look at the LIBOR rates we discussed in our prior conversation. 

This particular phone call with Madoff stuck with me more than any other. Shortly after finding out his son had died, Madoff wanted to discuss interest rates. He didn't lose a beat in the ensuing conversation, continuing to carry on an entirely fluid discussion on the arcane determinants of yields. It didn't seem as though he wanted to switch topics because he was struggling to compose himself, and it didn't seem as though he was avoiding expressing emotion because the news was so overwhelming. In some way, it almost seemed as though I was more personally moved by the death of Andrew in those moments than his father.

To a psychiatrist, Madoff displays many symptoms associated with psychopathy...while labels themselves are of little use, viewing Madoff through this lens helps place his prior actions and current rationalization of that behavior into context. 

Madoff interprets and responds to emotion differently from most people. Regardless of how close he got to his investors, his personal limitations enabled him to continue his fraud without remorse or guilt…Madoff has an inability to empathize with his investors…he never experienced the gut feeling that he needed to stop…he managed to create extraordinary suffering for his investors, his friends, even his family, while experiencing little emotional turmoil himself.

1. ^{^}

   Here’s a quote from MacAskill (emphasis added): 

   So the lesson that Eugene Soltes takes in his study of white collar crime that actually like the normal case of fraud like typical cases of fraud comes from people who are very successful, actually very well admired, really not the sort of people where it's like, Oh yeah, they were, everyone was talking all along about how this person's, you know, bad apple or not up to no good. Instead, you know Bernie Madoff even was the chair of NASDAQ...And so you know what he really emphasizes instead is importance of kind of good feedback mechanisms, again, because, you know, people are not often making this these decisions in this careful calculated way. Instead, is this like mindless, incredibly irrational decision...

The 3% figure for utilitarianism strikes me as a bit misleading on it's own, given what else Will said. (I'm not accusing Will of intent to mislead here, he said something very precise that I, as a philosopher, entirely followed, it was just a bit complicated for lay people.) Firstly, he said a lot of the probability space was taken up by error theory, the view that there is no true morality. So to get what Will himself endorses, whether or not there is a true morality, you have to basically subtract an unknown but large amount for his credence in error theory from 1, and then renormalize his other credence so that they add up to 1 on their own. Secondly, there's the difference between utilitarianism where only the consequences of your actions matter morally, and only consequences for (total or average) pain and pleasure and/or fulfilled preferences matter as consequence, and consequentialism where only the consequences of your actions matter morally, but it's left open what those consequences are. My memory of the podcast (could be wrong, only listened once!) is that Will said that, conditional on error theory being false, his credence in consequentialism, is about 0.5. This really matters in the current context, because many non-utilitarian forms of consequentialism can also promote maximizing in a dangerous way, they just disagree with utilitarianism about exactly what you are maximizing. So really, Will's credence in a view that, interpreted naively recommends dangerous maximizing is functionally (i.e. ignoring error theory in practice) more like 0.5 than 0.03, as I understood him in the podcast. Of course, he isn't actually recommending dangerous max-ing regardless of his credence in consequentialism (at least in most contexts*), because he warns against naivety.  

(Actually, my personal suspicion is that 'consequentialism' on its own is basically vacuous, because any view gives a moral preferability ordering over choices in situations, and really all that the numbers in consequentialism do is help us represent such orderings in a quick and easily manipulable manner, but that's a separate debate.)

*Presumably sometimes dangerous, unethical-looking maximizing actually is best from a consequentialist point of view, because the dangers of not doing so, or the upside of doing so if you are right about the consequences of your options outweigh the risk that you are wrong about the consequences of different options, even when you take into account higher-order evidence that people who think intuitively bad actions maximize utility are nearly always wrong. 

I feel like it's more relevant what a person actually believes than whether they think of themselves as uncertain. Moral certainty seems directly problematic (in terms of risks of recklessness and unilateral action) only when it comes together with moral realism: If you think you know the single correct moral theory, you'll consider yourself justified to override other people's moral beliefs and thwart the goals they've been working towards.

By contrast, there seems to me to be no clear link from "anti-realist moral certainty in some subjectivist axiology" to "considers themselves justified to override other people's life goals." On the contrary, unless someone has an anti-social personality to begin with, it seems only intuitive/natural to me to go from "anti-realism about morality is true" to "we should probably treat moral disagreements between morally certain individuals more like we'd ideally treat political disagreements." How would we want to ideally treat political disagreements? I'd say we want to keep political polarization at a low, accept that there'll be view differences, and we'll agree to play fair and find positive-sum compromises. If some political faction goes around thinking it's okay to sabotage others or use their power unfairly (e.g., restricting free expression of everyone who opposes their talking points), the problem is not that they're "too politically certain in what they believe." The problem is that they're too politically certain that what they believe is what everyone ought to believe. This seems like an important difference! 

There's also something else that I find weird about highlighting uncertainty as a solution to recklessness/fanaticism. Uncertainty can transition to increased certainty later on, as people do more thinking. So, it doesn't feel like a stable solution. (Not to mention that, as EAs tell themselves it's virtuous to remain uncertain, this impedes philosophical progress at the level of individuals.) 

So, while I'm on board with cautioning against overconfidence and would probably concede that there's often a link between overconfidence and unjustified moral or metaehtical confidence, I feel like it's misguided in more than one way to highlight "moral certainty" as the thing that's directly bad here.

(You're of course free to disagree.) 

I don't think the "3% credence in utilitarianism" is particularly meaningful; doubting the merits of a particular philosophical framework someone uses isn't an obvious reason to be suspicious of them. Particularly not when Sam ostensibly reached similar conclusions to Will about global priorities, and MacAskill himself has obviously been profoundly influenced by utilitarian philosophers in his goals too.

But I do think there's one specific area where SBF's public philosophical statements were extremely alarming even at the time, and he was doing so whilst in "explain EA" mode. That's when Sam made it quite clear that if he had a 51% chance of doubling world happiness vs a 49% of ending it, he'd accept the bet....  a train to crazytown not many utilitarians would jump on and also one which sounds a lot like how he actually approached everything. 

Then again, SBF isn't a professional philosopher and never claimed to be, other people have said equally dumb stuff and not gambled away billions of other people's money, and I'm not sure MacAskill himself would even have read or heard Sam utter those words.

fwiw, I wouldn't generally expect "high confidence in utilitarianism" per se to be any cause for concern. (I have high confidence in something close to utilitarianism -- in particular, I have near-zero credence in deontology -- but I can't imagine that anyone who really knows how I think about ethics would find this the least bit practically concerning.)

Note that Will does say a bit in the interview about why he doesn't view SBF's utilitarian beliefs as a major explanatory factor here (the fraud was so obviously negative EV, and the big lesson he took from the Soltes book on white-collar crime was that such crime tends to be more the result of negligence and self-deception than deliberate, explicit planning to that end).

I basically agree with the lessons Will suggests in the interview, about the importance of better "governance" and institutional guard-rails to disincentivize bad behavior, along with warning against both "EA exceptionalism" and SBF-style empirical overconfidence (in his ability to navigate risk, secure lasting business success without professional accounting support or governance, etc.).

I think it would be a big mistake to conflate that sort of "overconfidence in general" with specifically moral confidence (e.g. in the idea that we should fundamentally always prefer better outcomes over worse ones). It's just very obvious that you can have the latter without the former, and it's the former that's the real problem here.

[See also: 'The Abusability Objection' at utilitarianism.net]

Made a transcript with Microsoft Word.

Pro-tip, if you have an iPhone, Apple's podcast app now has transcriptions for most podcasts, including this one :) 

Yes, but not at great length. 

From my memory, which definitely could be faulty since I only listened once: 

He admits people did tell him Sam was untrustworthy. He says that his impression was something like "there was a big fight and I can't really tell what happened or who is right" (not a direct quote!). Stresses that many of the people who warned him about Sam continued to have large amounts of money on FTX later, so they didn't expect the scale of fraud we actually saw either. (They all seem to have told TIME that originally also.) Says Sam wrote a lot of reflections (10k words) on what had gone wrong at early Alameda and how to avoid similar mistakes again, and that he (Will) now understands that Sam was actually omitting stuff that made him look bad, but at the time, his desire to learn from his mistakes seemed convincing. 

He denies threatening Tara, and says he spoke to Tara and she agreed that while their conversation got heated, he did not threaten her.