Tom Williams/CQ Roll Call via AP Images
Sam Bankman-Fried testifies during a House Agriculture Committee hearing last May.
The fallout from the implosion of the cryptocurrency exchange FTX is still playing out, but one thing is already clear. The disgrace of FTX founder Sam Bankman-Fried has done significant reputational damage to the influential philanthropic movement called effective altruism, or EA. In recent years, EA has dramatically expanded its ambitions thanks, in part, to commitments in the $100 million range from Bankman-Fried, commonly known as SBF. Those commitments are worthless now, and as the magnitude of SBF’s frauds come into view, the EA community finds itself under an uncomfortable spotlight.
Many effective altruists have engaged in laudable and public self-reflection. EA is also a varied movement with lots of ideas. Yet the movement has not faced up to some deep contradictions in its guiding philosophy. One of its best-known mantras, “earn to give”—the idea that you should try to make a lot of money so that you can donate as much as possible—overlooks the reality that, the moment you approach any appreciable scale of action, how you earn matters far more than how you give.
But there is a deeper issue, best illustrated by EA’s recent turn toward “longtermism.” The longtermist conceit offers a spectacular gamble on the future of human civilization, one that parallels SBF’s calculated bet that he could bluff crypto finance into the mainstream. While effective altruists have disavowed his frauds, they have not acknowledged that the bet itself was consistent with a common EA approach to the future: a one-sided punter’s rationality laser-focused on how to spend money well, without asking about the structural implications, including the immense risks, of getting the money in the first place.
The effective altruism movement has not faced up to some deep contradictions in its guiding philosophy.
The standard story on SBF is that he became an earn-to-giver after a conversation with prominent EA philosopher Will MacAskill. It now looks like SBF’s commitment to EA may have been a sham. In at least one case, he appears to have used philanthropy as an opening to hook a major venture capital investor, and in a bizarre DM exchange reported by Vox’s Kelsey Piper, he seems to say that it was all a PR stunt. But SBF’s EA pretensions remain entirely plausible. He did exactly what many other EAs have done, just on a massively larger scale. As Dylan Matthews put it in an August write-up of EA’s newfound riches, SBF “‘earned to give’ . . . so hard that he’s now worth about $12.8 billion.” It was all to the good if you didn’t think about it too much, because it vastly expanded the philanthropic potential. The EA community broadly embraced SBF and his money, failing to ask questions about what the crypto boom was really doing.
EA promotes a rigorous ethical life based on philosophy called consequentialist utilitarianism. Actions must be evaluated in light of the effects that they produce (consequentialism) and these effects can be calculated in order to rationally choose among options (utilitarianism). It is, in some ways, an admirable creed with much to recommend it. Many of its adherents show a remarkable personal commitment to doing right. Devoted earn-to-givers, for instance, sometimes donate 50 percent or more of the high incomes they attain by working in finance or tech. Givewell, a clearinghouse for philanthropy based on documented impact, exemplifies a community of serious-minded givers.
To maximize the impact of that giving, EAs try to calculate the probable effects of various possible courses of action. The basic way of doing this is to sum the product of outcomes and probabilities. In other words, you give a numerical value to the good thing you believe could come of a philanthropic action, give another value to the chance this good thing will actually come about, multiply the two, then add together all such possible weighted outcomes. An arithmetic quirk of this technique is that an extremely good but very unlikely outcome might emerge as the rational choice, leading you to spend a great deal of money on a project that will probably come to nothing.
Longtermism lives on these kinds of computations. As outlined with particular stridency in a 2021 report authored by MacAskill and Hilary Greaves, longtermism argues that governments should invest heavily in programs that will do nothing for people in the near term but just might provide important benefits to humanity in the very long term. The calculation here is simple: There will be so many trillions and trillions of people in the far future that any action to ensure their survival (and, by presumption, their happiness) will dwarf near-future benefits. This remains mathematically true even if the action is very likely not to make any difference whatsoever.
The longtermist version of consequentialist utilitarian reasoning therefore justifies big and risky bets. Little surprise, then, that several people in the crypto world have said that SBF was a high-stakes gambler. Jason Choi, a self-described “full time crypto native investor,” tweeted that FTX’s crash came down to SBF’s “philosophy of betting big.” On an episode of the “Hidden Forces” podcast recorded a few days earlier, crypto venture capitalists Vance Spencer and Michael Anderson commented that SBF once said he would pay a lot of money to play the so-called St. Petersburg paradox, a philosophical conundrum in which it appears rational to pay large sums to participate in a game with a small chance of winning a near-infinite prize.
This seems to be what best explains SBF’s strategy. SBF took enormous, high-stakes risks to leverage an initial quantum of wealth into an empire of influence that portended crypto’s full-scale mainstreaming. The strategy had three prongs: (1) aggressive marketing, including celebrity endorsements, to expand the pool of retail crypto buyers; (2) creation of a mystique around SBF himself to establish credibility on Wall Street, generating demand for crypto “exposure” by mainstream financial institutions; and (3) lobbying in Washington for favorable regulatory treatment, which would open the floodgates for big-time institutional investors to flow real money into crypto while simultaneously reassuring the public with a show of regulatory oversight. It was a snow job of epic proportions.
At this point, it is important to understand that cryptocurrencies have no genuine use case, at least for now. The market volatility for Bitcoin and other tokens means that crypto cannot sustain as a stable “store of value.” Instead, people buy and sell crypto as speculative assets. As long as fresh money comes pouring in from a growing pool of investors, the Ponzi-like scheme can continue. What SBF built was a relentless hype machine to keep the wheel spinning. The bet was that network effects could build on themselves, making crypto so big and so inevitable that it would finally begin to act like real money. The odds of this charade succeeding were long indeed, but the bet must have seemed worth it because the potential payoff was so gargantuan. Multiply a small number by a much much larger number and you still get a damn big number.
If successful, it could have brought about a global meltdown on the order of 2008. Observers have been aghast at established venture funds like Sequoia and Temasek going in on FTX, but that was only the tip of a financial iceberg. In comments to Congress, Fed vice chair for financial supervision Michael Barr observed “the potential for systemic risk if interlinkages develop between the crypto system that exists today and the traditional financial system.” The absence of these interconnections is all that prevented “a financial crisis, crash and bailouts,” according to Dennis Kelleher of the financial watchdog Better Markets. Yet interlinkages were precisely what SBF was attempting to build. The potential endgame here was serious risk to the global financial system and therefore to the livelihoods and welfare of billions of people.
For effective altruists who mostly looked the other way, the situation is damning. It might seem odd that a group of people dedicated to thinking rationally about the consequences of their actions would be so blasé about something so obviously begging for reasoned skepticism. But the EA failure here is not really surprising. Earn-to-give presupposes that the earning side of the equation is neutral and unproblematic. It is only the giving that is subjected to rigorous analysis, through careful calculation of expected utilities. Moreover, doing consequentialist utilitarianism involves manipulating highly abstract symbols with formal logic. The operations are similar to what finance and tech workers do in their day jobs, perhaps explaining the affinity between EA and these sectors. SBF’s crypto combined all three.
EA mea culpas have focused on the issue of fraud, deception, and illegality in SBF’s specific dealings. But the wider context here is a financial sector that only a decade and a half ago brought disaster by engaging in widespread speculative overreach and shenanigans involving mortgage securities. Plenty of respected and seemingly aboveboard financial institutions participated. Yet EA sent bright and idealistic young people into the sector without a second thought. The MacAskill co-founded 80,000 Hours project, for instance, glosses over systemic risk by defining financial trading as simply “moving money around.” Questions might also be raised about the effects of the tech sector on workers, elections, psychological well-being, and more.
Some EAs I questioned along these lines in the past said that, since the finance jobs were going to get done by somebody, it was better for the earnings to be channeled into EA than into mansions and Ferraris. This is a fair point as far as it goes, but it doesn’t go very far. It is valid only if earn-to-givers remain a tiny fraction of the finance business: that is, only if they have no structural impact.
It is structure, however, that most accounts for the distribution of goods and harms among the world’s people. No matter how personally committed earn-to-givers are, their efforts can have only marginal effects on global injustice and suffering. Even such truly admirable interventions as direct cash payments to poor people and distribution of antimalarial mosquito nets—unmitigated goods that I support wholeheartedly—can never overcome the deeper structural issues that keep people impoverished or threaten them with catastrophic damage from climate change. With emerging markets leading global crypto adoption and an estimated three-quarters of Bitcoin investors losing money, it seems possible that FTX’s fall alone has done more harm than the sum total good of EA actions.
SBF may or may not have been genuinely aiming at altruism. The rub for EA adherents is that it doesn’t really matter. He adopted an EA-like philosophy of action with unacceptable and ill-considered risks. It turns out that the consistent application of formal rationality to ethical problems—if not disciplined by common sense, basic decency, and, above all, a structural analysis of power—is apt to birth monsters.