“Earn to Give” and the Ethics of Mediocrity (The Future as an Emergency, Pt. 2)

We’re very close to the end, folks. This the second of three planned excerpts from the last chapter I’m wrapping up in the book - one that will actually become Chapter 2, in which I aim to create a broad preview of the book’s core arguments.

Your support is particularly meaningful at this moment: it will directly help me stay focused on this project over the next few weeks as it nears its final form. If you’re looking forward to the book, have the means, and want to see it reach its maximum potential, please consider becoming a supporter.

Effective Altruism’s focus on measurability and quantification led inexorably from Peter Singer’s subtly financialized allegory to the “Earn to Give” concept Will MacAskill sold Bankman-Fried in their first meeting.

“Earn to Give,” as an approach to making the world a better place, is intuitively perverse. MacAskill convinced Bankman-Fried that, especially for certain people with altruistic intentions and very high earning potential, it is not maximally “effective” to simply go into professions that directly benefited people, such as becoming a doctor or teacher. Instead, all across the EA and Rationalist ecosystems, those with a knack for finance, banking, or entrepreneurship were encouraged to pursue them, then donate their earnings generously.

This pitch hit Bankman-Fried like a bolt of lightning - it comported with his background, and his shortcomings, in an almost uncanny way. Bankman-Fried had been steeped in the logic of “effectiveness” his entire life, most of all by extended conversations with his mother Barbara about utilitarianism, an ethos that pioneered the idea of maximizing fungible units of human benefit. But meeting MacAskill gave him a novel way to apply those deep principles, within a then-growing movement that offered a simultaneous path to material power and moral authority through the accumulation of wealth.

The ethos also helps fuel the EA movement’s own growth and power. In the exceptional case of Bankman-Fried, the earn-to-give injunction resulted by mid-2022 in $36.5 million in donations to Effective Altruism organizations, dedicated to developing EA itself rather than directly helping anyone else, and in several cases led by MacAskill himself (and unfortunately, effectively consisting of a high proportion of stolen FTX user money). This included millions for the Center for Effective Altruism; for a group called Longview Philanthropy, a kind of philantrhopy-investment advisor for wealthy individuals; for scholarships, writing contests, and fellowships aimed at developing EA ideas and spreading them among young people. Investing in EA itself, by its own logic of leverage and maximization, would produce more net benefit for humanity over time than trying to directly “do good.”

Earn to Give also suited Bankman-Fried’s combination of vast ambition, and a near-total lack of direction - a common psychological profile of EAs. The absence of any native sense of purpose was combined, in Sam’s case, with a surprisingly limited skill set for such a supposed genius.

“I grew up in in Stanford, California, went to MIT after that, and really had no clue what i was gonna do with my life,” he told New Atheist figurehead Sam Harris in a December, 2021 interview.

“I sort of like half-heartedly thought maybe I'd be a physics professor, for kind of no good reason, and and quickly at MIT learned that I didn't really like research, and I probably wasn't really built for it, and that that was not gonna happen.”

“Around the same time [I],” Bankman-Fried continued, “I started thinking for the first time about what I should do with my life … coming from a utilitarian standpoint of, what would maximize ultimate well-being of the world? I hadn't thought about it very carefully, but when I like finally confronted this … it quickly became clear that at least there can be some things I could do that would have real impact, and one of those was going to be earning to give… trying to make what I can so I can donate.”

“At the time I was most involved with animal welfare organizations and went to them and said, Hey, would you prefer my time or my money? And they said definitely your money. You're not very good at leaflets.”

While offered as sardonic self-deprecation, it’s not hard to detect pain as Bankman-Fried recounts this story: not good at physics research or making xeroxed leaflets, he is shunted into the role of breadwinning workhorse, where he could make the most of his one true set of gifts - not even mathematics per se, but oddsmaking, like some shady Vegas bookie. Bankman-Fried was a born gambler, and not much else, preoccupied with winning, and with scale.

In a vacuum, this interest in money is gauche - but Effective Altruism and Earning to Give imbued it with virtue.

“There’s just this classic decreasing marginal utility of money to [individual] people,” Bankman-Fried said on the FTX Podcast in 2020. “The more money you have, the less the marginal [extra] dollar matters.” In other words, there’s no actual reason for any human to have a billion dollars. Unless …

“But that’s a lot less true of donating. If you think about the scale of problems in the world, you kind of get to multiply everything by $7 billion. So instead of asking questions like, how much money do I need to eat, you’re asking questions like, how much money does the world need to eat?[ FTX Podcast Episode #3, Sam Bankman-Fried origins]”

“More is always better. Better is always better. Whenever you’re doing something that can increase the amount that you can give, that’s sort of always good.”

What remained studiously unasked in EA was why Sam Bankman-Fried or any other individual should be the one gathering and then re-distributing this ever-expanding trove of resources. The movement implicitly positions the Altruist as the one who knows how to do the right thing, against other kinds of power - including democratic governance.

But the argument for earning-to-give is also quietly premised on a deep anxiety of failure. MacAskill warned Bankman-Fried of the risk of making a big bet on your own ability to directly change the world - even, perhaps especially, for Harvard and MIT students who would in fact be best positioned to actually shape reality. While framed as something akin to heroism, earn-to-give is in fact a hedge against the terror of mediocrity, and a radical submission to a system EA has no ambition to change. Sam and other children of Palo Alto, as we’ll see, are rigorously trained to fear failure, or even averageness - and he wasn’t even good at making flyers.

The Effective Altruist need not face that anxiety. Like the venture capitalist or asset trader, they are remote from the success or failure of any single intervention. This insulation from responsibility is mirrored in the expanding abstraction of the financial system itself, and would be reiterated in Bankman-Fried’s delusions of his own innocence. As with a diversified financial portfolio, no deep commitments or beliefs are required - only models and projections.

Effective Altruism is ideal for those to whom, as Kierkegaard put it, “the lack of possibility is like being dumb.”

When Effectiveness Meets Rationalism: FTX and CFAR

Sam Bankman-Fried took the maximum-leverage logic of Earning to Give several steps further - he Gave money he hadn’t Earned.

As detailed during Bankman-Fried’s criminal trial by Notre Dame accounting professor Peter Easton, the FTX Group’s extensive philanthropic donations were consistently funded in part or in full from money that it had agreed to custody on behalf of customers. $5 million of those commingled funds - a tiny drop in the flood - went to the Center for Applied Rationality, or CFAR. CFAR was the primary outpost of a movement known as “Rationalism,” which substantially melded with Effective Altruism by the mid-2010s.

“FTX Group,” here and throughout this book, is not a legal entity, but a functional term adopted in related litigation to encompass FTX’s Bahamian entity, its many branches in other countries, Alameda Research, investment vehicles largely owned by Bankman-Fried personally, shell companies such as North Dimension (used to fraudulently seek U.S. banking) and any number of other legal entities. Lawyers and liquidators found it easier to refer to it all under one name, because that’s how Bankman-Fried treated the totality, moving money without any financial rationale between nodes of a sprawling network he treated as a personal empire.

The Center for Applied Rationality is a California non-profit organization, founded in 2012 to provide workshops on combating cognitive bias, a mission which has overlapped substantially with promoting Effective Altruism. CFAR is deeply tied to the broader Rationalist movement, and specifically to Eleizer Yudkowsky’s Machine Intelligence Research Institute (MIRI). Yudkowsky and MIRI received early funding and support from far-right Silicon Valley vanguardist Peter Thiel.

While Yudkowsky has disavowed any personal tolerance for the far-right neoreactionary movement with which Thiel is increasingly synonymous, the projects share an understanding of the human mind as little more than a thinking machine, the body as an appendage, technology as the engine of progress, and immortality as a goal of human thriving. Thiel acolyte and early LessWrong reader Blake Masters, who proved far too weird to win elective office, once declared his personal motto to be: “Your mind is software. Program it. Your body is a shell. Change it. Death is a disease. Cure it. Extinction is approaching. Fight it.”

Other tech billionaires form the foundation of the Rationalist and EA universe’s sprawling network of think-tanks and institutions. Dustin Moskovitz, a lesser-known co-founder of Facebook, is the biggest single funder of Effective Altruism - only briefly displaced by Bankman-Fried. Jaan Tallin, a founder of both Skype and Kazaa, is a cofounder of both the Cambridge Centre for the Study of Existential Risk (CSER) and the Future of Life Institute, whose concern with predicting extinction-level threats to the human species as a whole became the main concern of Effective Altruism, as well. Tallin loaned at least $110 million for the founding of Alameda Research.

As unsurprising as the approval of billionaires, then, are EA and Rationalism’s shared profound faith in “technology” as the key engine of human progress. While AI is a nightmare for Yudkowsky, it is feared because it is imbued with such god-like powers. Sam Altman, the EA-linked CEO of OpenAI, has promised a future where AI solves major problems in human society. Among the greatest risks to the human race, MacAskill includes “technological stagnation” (though not climate change, which EAs generally regard as bad, but not an extinction-level threat to humanity).

CFAR also has ties to Peter Thiel through its support for the Thiel Fellowship, a program that provides young people with incentives to forego a college education. While often celebrated for supporting youthful genius outliers, the Thiel Fellowship also echoes Rationalism’s disdain for expertise and experience in favor of raw instrumental intellect: a bias that became wildly destructive for Bankman-Fried and his customers.

Also like Effective Altruism and the Rationalist movement more broadly, CFAR has cultish elements. Its Rationalism workshops are four-day, on-site intensives in reportedly spartan conditions, despite a hefty $3,900 price tag. This evokes the tactics of “totalism” used in seminars of the sort once run by est and the Landmark Forum - other Bay Area groups who claimed to offer better living through rationality, and which conceptualized the human mind as a piece of upgradeable technology.

The Rationalist movement does not pursue the elimination of bias for its own sake. Instead, Yudkowsky believed that more logical thinking would help people see the future that he already grasped: the threat that superintelligent AI would annihilate the human race entirely, unless it was built in ‘alignment’ with human values. It is a representative irony, and irrecoverable original sin, that Yudkowsky’s “Rationalism” is not a neutral set of tools, but a tendentious mission to convince the world of the absolute certainty of a longshot science-fictional hypothesis - with the goal of motivating confident, radical, and immediate action.

Leverage and Paperclips

“It’s tough to make predictions, especially about the future.” - Yogi Berra (apocryphal)

A saying is attributed to the Greek mathematician Archimedes: “Give me a lever long enough and a fulcrum on which to place it, and I shall move the world.”

Nick Cooney’s Lever VC is just one of the innumerable uses of “leverage” in EA and Rationalist discourse and naming conventions. Broadly, leverage refers to a relatively small amount of force, used with foresight and strategy to enact large changes. In philanthropic terms, “leverage” might imply finding a very targeted, low-cost solution to a dire problem - the most repeated example is using cheap bed nets to fight malaria. In the context of Rationalism, the implied “lever” is the elimination of human bias and the crafting of perfectly rational decision-makers. The term also implies the way that a small investment or intervention in the present can have much larger effects in the future.

But in the case of EAs, given their emphasis on the power of wealth, leverage must be interpreted substantially in its specific financial sense: as a method of investing more wealth than one actually possesses, using borrowed funds, to create larger future impacts. A “leveraged trade” in a financial market is a speculative investment made with borrowed money, allowing the trader to make more money than if they had only been able to deploy their own money. The downside is that leveraged bets are at risk of losing far more money if things don’t go according to plan. In fact, a leveraged trade is far more likely to lose all of the investors funds: if they bet wrong, their initial principle gets liquidated to cover the outstanding loan. This side of leverage is not foregrounded in the EA conception, any more than is the losing side of a world-destroying coin flip.

This is essentially what happened to Sam Bankman-Fried: his customers’ decision to withdraw their funds from FTX wasn’t a bank run, but the world’s biggest margin call. He was caught without enough free cash to cover it, and got liquidated, via bankruptcy.

The thing about financial leverage is that you have to borrow all that money from someone, and you have to convince them that you can be trusted with the loan - whether to pay it back, or in the case of world-changing philanthropy, to have a real impact on the world. In conventional markets, things like credit ratings determine loan eligibility - whether you’ve treated a loan responsibly in the past is a good indicator of what you’ll do with it in the future. In its early form, EA’s focus on experimental evidence of the effectiveness of particular efforts served something of the same role.

More confident knowledge is sought as a means of maximizing leverage, and impact. This is obviously not a goal unique to the techno-utopian mind, but they are notable for the strength and confidence of their claims. “Instrumental rationality,” Yudkowsky wrote in a particularly influential description of his agenda, “is about steering reality—sending the future where you want it to go. It’s the art of choosing actions that lead to outcomes ranked higher in your preferences. I sometimes call this ‘winning.’” Winning, for the computational mind, is the point of knowledge and reason.

But as the EAs increasingly came to think about far future extinction risk via asteroid, or share Rationalist’s emphasis on the speculative possibility of a genocidal artificial intelligence, there was less and less actual evidence available for proving just how effective their efforts would be ahead of time. Justifying the investments needed to exercise leverage came to rely less on evidence than on argumentation. That has had significant distorting effects on how the techno-utopians make their arguments - above all, increasing the confidence and stridency of their claims.

The discussion of “AI Doom” is a key illustration of this effect. The hypothesis of “AI Doom” was first laid out by an Oxford philosopher named Nick Bostrom in 2003.[ Bostrom in 2005 founded the Future of Humanity Institute, largely to study the supposed problem of AI Doom. FOHI was shut down by Oxford in 2024, seemingly in part after the unearthing of Bostrom’s seemingly enduring belief in the genetic inferiority of Black people.] Bostrom posited that an artificial intelligence capable of self-improvement might quickly become orders of magnitude smarter than humans, and capable of pursuing its own ends against any human resistance. This event is known in techno-utopian circles as “the Singularity,” and believers in this “God AI,” including Singularity pioneer Ray Kurzweil, have generally imagined it as a purely benevolent force, one capable of solving humanity’s biggest problems - including granting humans eternal life as simulated digital minds.

But Bostrom saw the downside risk. He specifically imagined that a “superintelligence whose top goal is the manufacturing of paperclips” would use nanobots to turn every atom on Earth into paperclips - including human bodies. Elon Musk saw in Bostrom’s predictions the likelihood that human beings are merely “the biological boot loader for digital superintelligence” - for better or worse.

The concept was clearly inspired less by any rigorous analysis of the likely evolution of artificial intelligence - which didn’t exist in any meaningful sense at the time, and still doesn’t - but by a huge swathe of science fiction, from Harlan Ellison’s “I Have no Mouth & I Must Scream” to Frank Herbert’s Dune to James Cameron’s Terminator. While the techno-utopians overtly reject the value of the humanities, they are nonetheless in the thrall of myths and stories - but these and other stories are digested as straightforward claims about reality, rather than nuanced lessons about human nature.