A Lever Long Enough (Draft Excerpt, Stealing the Future)

Forensic details of Sam Bankman-Fried's comically rampant thievery.

Welcome to another draft excerpt from “Stealing the Future,” my forthcoming book on Sam Bankman-Fried, the FTX fraud, and its roots in Effective Altruism, Rationalist, and Longtermist philosophies.

The following excerpt is a portion of a chapter that dives into the gritty details of various transfers made from the FTX Group to an array of recipients. It draws heavily on Duke professor Peter Easton’s forensic analysis of dollar flows, presented under oath at Bankman-Fried’s criminal trial, which specifically identified the movement of customer funds out of either cash depository accounts or crypto wallets, through Alameda Research, and onward to entities including Anthropic AI and Michael Kives’ K5 group. It also incorporates civil lawsuits filed by the FTX estate and seeking clawbacks from recipients including the Center for Applied Rationality and its subsidiary, Lightcone. Future additions will grapple with political donations, including those laundered through Barbara Fried’s Mind the Gap.

Like all excerpts, this one is explicitly a draft, which means it’s both messy and incomplete. For this reason, most of it is behind a paywall. I appreciate any suggestions and comments from supporters as I continue to refine it.

If you are a new or recent paid subscriber, you can also access these previous drafts and excerpts:

The Racehorse Who Gambled: On Expected Value and Discounted Future Cash Flows

Effective Altruism and Jean Baudrillard’s Logic of the Code

Stealing the Future: From the Horse’s Mouth

Prompt: a famous technology and cryptocurrency founder with a huge sphere of curly black hair uses a long lever to roll the earth over a cliff

It is clear from Sam Bankman-Fried’s public and private statements, and from his understanding of leverage itself in the context of FTX, that from the beginning the goal of his massive embezzlement was to grow his empire as fast as possible. Caroline Ellison testified that Bankman-Fried told her customer funds were “a good source of capital.” Even as the depths of the crisis became clear in November of 2022, when a panicked Nishad Singh confronted Bankman-Fried about the growing hole caused by Alameda’s “allow negative” exemption, Bankman-Fried responded that the “main plan remains, making FTX successful and growing it.[ Trial transcript: Id. 1411:3–1412:10]” Former FTX.com General Counsel Can Sun testified to hearing Bankman-Fried tell Singh that the hole “is what it is and there is nothing we can do about it. The only thing we can do is grow the company and fill the hole.[ Id. 1964:12-1965:5. Trial transcripts]”

That the hole he believed he could fill was greater than all the money Alameda had ever earned trading, and more than FTX had ever earned in legitimate fees, highlights just how far in the future Sam Bankman-Fried’s mind was. But this was no surprise to Sam - the panoply of expenditures funded with customer property amounted to long-term bets on the growth of FTX, and they were already unbelievably risky. Most obviously, the deployment of billions of dollars on venture capital was a wild gamble on absolutely nothing going wrong: even successful venture-funded companies are generally “illiquid” for years before an IPO or acquisition turns their notional equity value into real cash. This is why most formal venture funds lock investors into their positions for years, and why there was no cash left when withdrawals began. The implicit belief backing those bets was that FTX would see no major obstacles, and that the market itself would experience no major downturns, for at least half a decade.

Bankman-Fried’s other bets, though, were even more deranged. Rather than any financial position at all, they often amounted to leveraged bets on Bankman-Fried’s own stature and influence, such as the tens of millions spent on electoral campaigns, from professional sports endorsement deals, for networking with celebrities, or in donations to nonprofits. Notionally, these were also part of the strategy to grow FTX, build up its revenue, and pay back the borrowed customer funds. But the value of this sort of influence-investment is even harder than a venture investment to reliably turn into cash.

To be quite clear, this was never a truly viable strategy. The idea that cryptocurrency markets would continue rising in both value and trading volume had no historical precedent whatsoever - every peak like the one that fueled FTX to such heights circa 2021 had eventually retrenched. Moreover, this boom-bust cycle has been true of every innovative technology in the modern era. Bankman-Fried’s behavior despite this reflected his ignorance and naivete, but also other factors: a cultural obsession with youth that convinced him he had only a short window for success, and most of all, a worldview premised on mathematical certainty and deterministic human behavior that gave him false confidence in his own predictive and analytical powers.

Bankman-Fried had practically no checks on his authority to spend any of the money across the FTX and Alameda. FTX had no board of directors to check his total authority, and no Chief Risk Officer with the official duty to supervise his choices. In part through a purge early in the life of Alameda Research, he had surrounded himself with compliant yes-men and -women, most notably within the inner circle of his confessed co-conspirators. Even on the rare occasions when a member of his inner circle objected, as Caroline Ellison did when Bankman-Fried decided to allocate $3 billion to illiquid venture capital investments, he routinely, casually overrode them.

So his theft-fueled expenditures reflected his sole vision of the future, for himself, and for the world. Both in their direction and their fundamental fraudulence, they are a map of Bankman-Fried’s mind and thinking. What is most clear is that he saw himself not merely as a financial titan, but as something far grander: as a savior of society itself, ready to take the reins of wealth and power and do things better. While he stated this outright in the grade-school fantasy of becoming President of the United States, [He gave himself, with typically improvised precision, a 5% chance of becoming President someday: https://www.businessinsider.com/sam-bankman-fried-wanted-president-caroline-ellison-testimony-2023-10], in practice he knew that the power of money trumped any democratic institution.

As we will see in more detail, this authoritarian impulse is deeply embedded in Effective Altruism and its affiliated movements. It is overt in EA cofounder Toby Ord’s call for a ruling council of experts who would have veto power over the actions of world governments. Only slightly more subtle are para-academic institutions, such as Eleizer Yudkowsky’s Machine Intelligence Research Institute, structured to assert the inherent authority of a “rational” approach to the management of society.

The various expenditures, investments, and donations made with stolen and commingled FTX customer funds broadly fell into five categories:

• Illiquid Venture capital investments, including into firms closely associated with Effective Altruism via work on artificial intelligence;

• “Philanthropic” donations to organizations that either claimed to operate according to Effective Altruist principles, or directly worked on developing Effective Altruism and affiliated movements including rationalism, longtermism - and, in at least one case, eugenics;

• Marketing and networking efforts designed to build the public profile of FTX and Sam Bankman-Fried personally;

• Political donations and campaigns nominally intended to further Effective Altruist goals in the political sphere - but with the added, and eventually superseding goal of shaping cryptocurrency regulation.

• Funds lost to bad trades, hacks, and sheer disorganization.

In addition to the lack of any internal voices to countermand his chaotic whims, the total internal financial chaos described by FTX liquidation and recovery CEO John Ray III enabled Bankman-Fried’s unconstrained spending. Movements of customer funds were authorized via Slack messages and tracked in Quickbooks, if at all. Unravelling this chaos in any final way even now seems like an impossible task - Bankman-Fried was seemingly correct when he once commented that Alameda Research was “unauditable.”

Bankman-Fried mistook this lack of constraints for something else: He told Michael Lewis that at some point it began to feel like FTX had “infinite money.” This confession of a mind untethered from reality is the source of the strangely laudatory title of Lewis’ book, “Going Infinite.” That infinitude itself was implicitly based on an assumption of rapid future growth. As with much of his behavior, Bankman-Fried spent other people’s money as if that infinite future had already arrived.

a famed white technology and cryptocurrency founder with a large jewfro of curly black hair uses a long lever to push a detailed globe of the Earth showing continents and oceans off a tall cliff while wearing an ftx t-shirt

***

While the picture may never be truly complete, the best and clearest effort so far to trace FTX/Alameda’s spending was made by Peter Easton, a professor of accounting at the University of Notre Dame. Easton performed a forensic analysis of FTX and Alameda’s spending on behalf of Federal prosecutors in Bankman-Fried’s criminal trial, and testified on October 18 of 2023.

Easton, who resembles a skinnier, better groomed Santa Claus, was questioned by an affable, hulking, and even more thoroughly bald prosecutor named Nicholas Roos. Under oath, Easton described how he had reconciled financial records to create charts showing the assets that Alameda and FTX had on hand, and the amount that each owed to various parties, over time. Easton also traced numerous specific transactions, using public blockchain records, FTX’s internal trading database, third-party bank statements, and the records of third party lenders.

Easton described the most important conclusion of his analysis: FTX and Alameda were in various senses insolvent for nearly their entire existences, with liabilities that vastly outstripped their obligations.

“The amount of customer fiat deposits … held in Alameda Research and FTX.com accounts was way less than was owed to customers on FTX,” Easton found. “And the question is, of course, is what happened to that money. Well, Alameda Research used it for their own expenditures.” Easton found that FTX’s liability to customers in fiat money peaked at $11.3 billion in June of 2022. At that time, the actual cash in relevant accounts was just $2.3 billion.

“Similarly,” Easton told Roos, “the amount of [crypto held in] crypto wallets was far less … than the amount that was owed to FTX customers, and, again, Alameda Research used customer crypto funds to pay for expenditures.”

Finally, Easton conducted an analysis of all Alameda accounts on FTX, measuring the hedge fund’s indebtedness to FTX customers. Taken together, Alameda accounts on FTX were “almost always in the red,” Easton found. On May 12, 2022, they were negative by a total of $12.6 billion. On June 14, 2022, they were negative $10.9 billion. November 1, 2022 they were negative $9.2 billion.

Easton also flatly rebutted the idea that Alameda’s debt to FTX might have been explained under the so-called spot margin program: there simply were not remotely enough FTX customers who had opted to make their funds available for borrowing. “The actual Alameda borrowing on the spot-margin program,” he told Roos, “Was far less than the amount needed to cover [Alameda’s] negative balance … For example, on May 12, the … difference between the [spot margin borrowing] needed and what we did have, is $10.8 billion.”

Easton also delved further, closely tracing the movement of fiat and crypto funds into, through, between, and (inevitably) out of FTX and Alameda Research. Even a professional like Easton at times discussed FTX and Alameda as a single entity, for the very good reason that they effectively were. The graphs Easton presented resembled nothing so much as rat’s warrens, with trails leading between and among a dozen different bank accounts, forming one indistinguishable blob.

The most notorious example of this rampant commingling were bank accounts nominally owned by a U.S. entity known as North Dimension. North Dimension was set up by top FTX Lawyer Dan Friedberg to help FTX set up bank accounts that U.S. and other customers could deposit funds into. Banking was difficult for FTX thanks to skepticism of an offshore exchange in poorly-understood assets, so North Dimension was represented as an electronics retailer - including with a website made to look like it actually sold headphones, computer monitors, and tablets.[ https://cointelegraph.com/news/here-s-how-sbf-s-fake-electronics-outlet-north-dimension-looks-like] The North Dimension banking accounts were in fact controlled by Alameda Research, though, and Bankman-Fried at various times tried to characterize this strange arrangement as the reason he didn’t notice billions of dollars had gone missing.

But the truth is that no one account or arrangement can be blamed - Easton identified 47 different bank accounts, across various entities, that accepted customer fiat deposits. There are some elements of this chaos that showed the players trying to justify the movement of funds, such as the use of FTT as collateral for loans between and beyond FTX and Alameda. But more generally, no single confusing account structure or ill-advised loan explains the broader truth of the total chaos of the firms’ accounting. What really mattered was Bankman-Fried’s personal power to move money wherever he wanted between FTX, Alameda, and his own pockets, or into any investment or expenditure at all, using only a Slack message - and with no meaningful checks on his authority at all, and regardless of which entity (or customer) the money technically belonged to.

Easton’s analysis of individual transactions are more informative. Easton found that FTX customer deposits, which according to the Terms of Service signed by those customers were meant to be sacrosanct, were being spent with abandon. Many major transactions were funded 50% or more by customer funds.

The specific transactions he reviewed represented only a portion of what was spent in Bankman-Fried’s leveraged play for power, influence, and wealth. Other expenditures were detailed in later lawsuits by the bankruptcy estate attempting to “claw back” payments for creditors. Their details help show the end goals of Bankman-Fried’s attempts to turn money into influence. They show, in short, exactly what effect his altruism was intended to have.

Venture Capital and Other Investments

The Binance Buyout: $2.2 billion. Customer Funds: $1.2 billion

Appropriately enough, Bankman-Fried’s first major direct misappropriation of FTX customer funds was used to make an investment in … himself.

For many years before FTX arrived on the scene (and for many years after), the largest global cryptocurrency exchange was Binance, founded by Canadian entrepreneur Changpeng Zhao. Zhao, who was popularly known as “CZ,” was not only successful, but one of the most widely respected cryptocurrency entrepreneurs in the world.[ CZ would step down from leadership of Binance as part of a settlement with the U.S. government on financial regulatory and compliance violations. The settlement was largely viewed as a vindication and victory for Binance, which has never faced serious accusations of mishandling customer funds.] Zhao had been an early supporter of Bankman-Fried, purchasing a reported 20% stake in FTX for $80 million[ https://www.dlnews.com/articles/people-culture/inside-the-sbf-feud-with-binance-ceo-changpeng-zhao/] in late 2019 - just months after it was founded.[ https://www.binance.com/en/blog/all/binance-announces-strategic-investment-in-cryptocurrency-derivatives-exchange-ftx-414610870200725504]

But as the years went on, Bankman-Fried increasingly regarded CZ not as a supporter, but as a rival for the title of crypto’s top dog. By mid-2021, Bankman-Fried decided to buy back Binance’s stake in FTX: a stake whose notional valuation had skyrocketed from $80 million to $2.2 billion.

A venture-funded startup trying to buy out investors before going public is itself a truly bizarre, in fact nearly unheard-of tactical decision. As a general rule, startups need outside funding because they don’t have their own capital lying around, and early investors are only able to convert their stakes back into cash when a startup goes public in an IPO, or is acquired.

FTX was no exception, as Caroline Ellison quickly discovered when Bankman-Fried asked her to explore the possibility of buying back Binance’s stake. Up to that point in mid-2021, FTX’s total lifetime revenue was estimated to have been $1 billion dollars - less than half of what was needed to buy out Binance. Ellison testified to telling Bankman-Fried that “We don’t really have the money for this, we’ll have to borrow from FTX to do it.” Bankman-Fried responded, “That’s okay. I think this is really important, we have to get it done.[ Trial (Tr. 668)]”

According to Peter Easton’s analysis, they “got it done” with $1.2 million of customer funds, taken from customers using the “allow negative” designation on Alameda’s accounts on FTX. Ellison testified that this was the first time such a large single block of money was drawn from customer funds. A portion of this payout was in the form of FTT tokens, the equity-like instrument, created by FTX itself, which would play a key role in the fraud’s collapse.

Despite buying out Binance, Changpeng Zhao remained an antagonistic obsession for Bankman-Fried. Many allege that Bankman-Fried’s political maneuvering in Washington was partly intended to hamstring Binance and give FTX an upper hand, which Bankman-Fried seemed to allude to in an antagonistic October 2022 tweet: “excited to see [Zhao] repping the [crypto] industry in DC going forward! uh,[sic] he is still allowed to go to DC, right?[ https://www.dlnews.com/articles/people-culture/inside-the-sbf-feud-with-binance-ceo-changpeng-zhao/]” The suggestion here being, ironically, that CZ was or should be under investigation.

Even more pointedly, immediately after FTX’s collapse,Bankman-Fried would allude to CZ as “a particular sparring partner, so to speak … For now, all I’ll say is: well played; you won.[ “At some point I might have more to say about a particular sparring partner, so to speak.”

“For now, all I’ll say is: well played; you won.”]” The allusion was an attempt to lay blame for FTX’s collapse on CZ’s liquidation of FTT tokens, which helped topple Bankman-Fried’s house of cards. Given what he now know about the risk that Bankman-Fried himself secretly imposed on his own users, this is yet another display of a profound lack of self-awareness.

Or, as CZ himself would later put it with perhaps more accuracy than he knew, “only a psychopath can write that Tweet.[ https://decrypt.co/114923/only-a-psychopath-can-write-that-tweet-binance-ceo-cz-sbf]”

FTX Ventures: $ 3 billion

The courtroom was buzzing when former Alameda Research CEO Caroline Ellison took the stand on October 10 of 2023. In the wake of FTX’s collapse, it had become widely known that Ellison and Bankman-Fried had been on-and-off romantic partners, including during her time as CEO. Ellison’s private diary, leaked to the New York Times by Bankman-Fried’s in the attempt to intimidate Ellison that led to revocation of his bail, had painted a picture of her as strangely enthralled by her boss. She described him, in turn, as aloof and manipulative, studiously keeping her at a distance while leveraging her emotions to control her.

On the stand, Ellison detailed an instance that highlighted both Bankman-Fried’s extreme approach to financial risk, and his effective control of Alameda Research, in part through his romantic hold on her. In the fall of 2021, at a time when both crypto and FTX were riding very high, Bankman-Fried asked Ellison to assess the viability of creating a $3 billion venture capital fund through which FTX could invest in other companies. In calculating the viability of this outlay, Ellison said FTX customer balances were regarded as a source of capital.

The idea of FTX creating this huge venture fund is strange on its face - FTX and Alameda combined had not generated $3 billion in profit during their existence to that point, and were not yet themselves profitable on an ongoing basis. As with most other major FTX spending, that made customer funds a necessary part of the equation. Those funds would ultimately be borrowed from FTX by Alameda under Bankman-Fried’s unconventional conception of margin lending.

Ellison testified that she warned Bankman-Fried against the creation of the fund, specifically because in the event of a downturn in cryptocurrency prices, it would leave FTX unable to return customer deposits. Bankman-Fried not only overrode her, but insisted that the fund - despite being directly funded by obligations that Alameda had shouldered from FTX - be publicly referred to as FTX Ventures. The entity was still referred to internally as “Alameda Ventures,” and its bank accounts played a role under that name in further extractions, such as to Michael Kives’ K5 entities (see below). But Ellison testified that Bankman-Fried insisted on using the FTX Ventures name, because he felt the FTX brand was stronger than Alameda’s - yet again demonstrating his disregard for, or genuine inability to understand, separations between the two entities.

Anthropic AI: $500 million. Customer Funds: $500 million

“We have to wire 500 million to Anthropic,” the 2021 Slack message reads. “This should come from an Alameda Research Ventures bank account.”

The message was sent by Sam Bankman-Fried to an administrative subordinate, initiating perhaps the most singularly revealing of his many depositor-funded outlays. Most immediately, of course, we notice that funds controlled by Alameda are being directed by Bankman-Fried, who claimed by this time claimed to barely have influence, much less any leadership role, at the hedge fund. “Alameda Research Ventures” here seems interchangeable with FTX Ventures - as we’ve seen, Bankman-Fried treated all accounting and corporate boundaries as beneath his concern.

Sam’s message initiated a series of transfers described by Peter Easton. Money passed “through a series of Alameda Research customer depository accounts, through to an Alameda Research external account – in other words, this is an account that does not hold customer funds — of $500 million.” Though controlled by Alameda Research, remember, these “customer depository accounts” collected FTX customers’ fiat deposits. Easton concluded that the Anthropic investment was funded entirely by money taken from FTX customer deposits.

The Anthropic stake is also representative of the strong ties between FTX, Effective Altruism, and a broader coalition of Silicon Valley-centered philosophical movements including so-called “rationalists” and Longtermists. Anthropic describes itself as an “AI Safety and Research” startup, and its goals are heavily influenced by the thinking of the Rationalist movement. The poster boy for rationalism is a man named Eleizer Yudkowsky, head of the Machine Intelligence Research Institute (MIRI).

Supported by funders including pioneering Silicon Valley authoritarian Peter Thiel, MIRI’s overriding concern is the possibility of “AI Doom” - that a powerful enough artificial intelligence might run amok and destroy human civilization. That argument eventually came to dominate discussions of “existential risk” more generally, edging out somewhat more grounded risks like climate change and human-led genocide. As we will see, framing artificial intelligence as a risk to human survival made both for-profit and nonprofit entities focused on it very appealing recipients of funds for Effective Altruists.

More specifically, the Anthropic stake highlights a clear pattern by which Bankman-Fried’s spending disproportionately benefited people close to him. As described by Michael Lewis, Sam’s journey to becoming an Effective Altruist began with a meeting with Will MacAskill, the young Oxford philosopher perhaps most closely associated with Effective Altruism as a living movement. MacAskill was born William David Crouch, but he and Amanda Askell both adopted her grandmother’s maiden name, MacAskill, when the two wed.[ https://www.theatlantic.com/sexes/archive/2013/03/men-should-consider-changing-their-last-names-when-they-get-married/273718/] The pair separated in 2015, but Will kept the name.

Amanda Askell, at this writing, is still employed at Enthropic as an “AI Alignment Researcher.” This means that, in the same sense that FTX’s early theft of customer funds underwrote all of its future growth and appearance of success, money stolen from FTX customers continues to partly underwrite the employment of the ex-wife of the man who converted Sam Bankman-Fried to Effective Altruism.

The Anthropic stake also once again highlights how Sam Bankman-Fried was thinking about his use of customer funds. By the time his criminal trial arrived, the “AI Boom” triggered by the release of OpenAI’s ChatGPT was in full swing, and Amazon had, just the month before the trial, announced plans to invest $4 billion in Anthropic[ https://blockworks.co/news/amazon-anthropic-ftx-stake-creditors]. That meant the value of the investment made by Bankman-Fried had increased, and his legal team (perhaps at Sam’s behest) wanted to introduce that rise in value as exculpatory evidence in his defense.

Bankman-Fried’s desire to have the increased value of the Anthropic investment considered as evidence for his innocence was problematic for two reasons. Most obviously, as Judge Kaplan phrased the legal nuance when rejecting the request, “If you rob the Federal Reserve, then go to Vegas and win it all back, you still robbed the Federal Reserve.”[[ Find Exact Quote]] While the request may have represented the defense team grasping at any available straw to make their client look better, the idea that good returns on an investment funded by a gargantuan theft also seems to represent Bankman-Fried’s true beliefs about ethics: that results matter more than decisive moral rules like, as Caroline Ellison put it, “don’t lie and don’t steal.”

Though well beyond the scope of Judge Kaplan’s concerns, the idea that the Anthropic stake had “paid off” was problematic for a second reason: As an equity stake in a startup, it was no inherently “liquid,” or convertible to cash. Venture capital stakes are illiquid because the valuations placed on startups represent expectations of future earnings

, not present revenue - and Anthropic, to date, remains unprofitable. By the end of 2024, many of FTX’s theft-funded venture investments would still be illiquid, complicating attempts to repay depositors.

Modulo Capital: $450 Million. Customer Funds: $450 Million.

If his transfers to Anthropic illustrated Sam Bankman-Fried’s ideological commitments, the huge allocation of funds to a virtually unknown hedge fund called Modulo Capital highlighted his use of money to manipulate people in his personal life. Modulo was a crypto trading firm, or hedge fund, not notably different from Alameda Research. Its two young directors were Duncan Rheingans-Yoo, then newly graduated from college; and a woman Xiaoyun Zhang, known as Lily.

Lily Zhang had worked at Jane Street at the same time as Sam Bankman-Fried, and the two had dated.[ https://www.nytimes.com/2023/01/24/business/ftx-sbf-modulo-capital.html] In later testimony detailing internal discussions about the unfolding crisis, Bankman-Fried would suggest closing down Alameda Research and letting Modulo take over its role as a market-maker on FTX. Though Bankman-Fried was by this period not actively dating either Ellison or Zhang, both the initial huge transfers to Modulo and the later discussion of shutting down Alameda have the outline of a man using his wealth to replace one CEO/lover with another.

The amount invested in Modulo was initially reported as $400 million, but Peter Easton found transfers totalling $450 million. Easton concluded that all of these funds came directly from customer funds - and that, as with other series of transfers we’ll see below, they accelerated dramatically as the inevitability of FTX’s collapse would have become more apparent to Bankman-Fried.

The first payment, of a mere $50 million, was initiated by a Slack message send by Bankman-Fried on June 27th of 2022 to an employee named Jen. “We should send it via the Signet account,” Bankman-Fried instructs, indicating that customer funds should be tranferred to a specific Alameda Research-linked bank account. “Paid,” Jen replied. That was followed by another $50 million transferred to Alameda on July 29. Two months later came the true deluge, though: a series of 5 payments totaling $100 million on September 19, and seven transfers totaling $250 million, on September 25th and 26t of 2022.

Bankman-Fried’s directions of these final transfers hint at some urgency. “We’re putting an additional $250 million into Modulo,” he instructs a subordinate on Slack. “Would you be able to send over the funds to the same [Alameda] Signet address. It would be great if this done [sic] tonight, and ideally, $250 million tonight.”

Ultimately $292 million was transferred from customer depository accounts to Alameda in that final sprint, with $250 million then forwarded on to Modulo - Alameda seemingly pocketing the change. “We've got a total of $292 million,” Easton summarized, referring to the initial larger transfer to Alameda. “We don’t know exactly what customer this $292 million comes from, but we know it is only customer funds. It’s not any other source of funds. And so $292 million of customer funds was used to [fund] Modulo Capital.”

Philanthropy:

The Center for Applied Rationality: $5 million

Though far smaller in dollar terms than the half-billion funneled to Bankman-Fried’s allies at Anthropic, a total of $5 million given to a Bay Area nonprofit called the Center for Applied Rationality reveals even more of the ideological commitments that drove him. In this case, tracing the financial connections also pulls back the veil on some of the uglier ideas underlying such seemingly benign concepts as “rationality” - at least as practiced by the Bay Area movements Sam Bankman-Fried funded. They ultimately implicate Bankman-Fried - a non-religious man of Jewish heritage - in funding white supremacism and eugenics.

The Center for Applied Rationality is a California non-profit organization, founded in 2012 and meant to provide workshops on combating cognitive bias, and which overlapped substantially with promoting Effective Altruism.[ See for example: https://www.effectivealtruism.org/articles/ea-global-2018-cfar-workshop] CFAR is deeply tied to the broader Rationalist movement, and was promoted through Eleizer Yudkowsky’s blog Less Wrong.[ https://www.nytimes.com/2016/01/17/magazine/the-happiness-code.html] Like Yudkowsky, CFAR also has ties to Peter Thiel through its contributions to the Thiel Fellowship, a program that provides smart young people with incentives to not pursue a traditional college education.

Also like Effective Altruism and the Rationalist movement more broadly, CFAR has borderline cult-like features: its rationalism workshops were four-day, on-site intensive workshops in reportedly spartan conditions, despite a hefty $3,900 price tag. This evokes “encounter sessions” once run by groups like EST and the Landmark Workshop, who similarly claimed to offer secrets to better living. At its most extreme, this format of isolated, rigorous training is frequently used by cults to indoctrinate followers.[ https://aeon.co/essays/how-cult-leaders-brainwash-followers-for-total-control]

According to a civil suit filed by the bankruptcy administrators of the FTX estate, FTX’s payments to CFAR began in March of 2022, with an initial tranche of $2 million send directly from FTX proper. Soon after that, payments began flowing from the FTX Foundation. The flow of these payments were not detailed by Peter Easton at Bankman-Fried’s criminal trial, but the civil suit argues that the “FTX Foundation’s primary source of funds was Alameda monies that had been commingled with FTX customer deposits.” The suit further suggests a broader agenda: “In reality, very few of FTX Foundation’s donations directly benefitted the needy. Its largest donations went to associates of FTX Insiders in the ‘effective altruism’ movement.”

In addition to the specifically fraudulent source of funds, the estate sought clawbacks, in this and other cases, on the basis of “undue enrichment.” Undue enrichment occurs when payments are not made in exchange for services of value, or when a payer is insolvent - that is, when a payment is actually a disguised theft of corporate funds.

The pattern of payments to CFAR are themselves quite suspicious. The initial March payment of $2 million was followed by a fairly tidy series of payments in July, August, and September. Then on October 3 of 2022, everything changes - instead of one lump sum or a series of smaller gifts, that single day saw ten different transactions, each sending either $150,000 or $160,000, plus one payment of $100,000, for a total of $1.5 million in a single day. Given insiders’ dawning awareness of FTX’s fragility in the fall of 2022, it is exceedingly difficult not to interpret this flood of payments as a rushed attempt to exfiltrate as much money as possible before the jig was up.

Two of the FTX Foundation’s payments, the $500,000 tranches sent on July 13 and August 18, were of particular interest. According to the estate’s suit, these payments were sent from North Dimension bank accounts directly to a title company as a deposit for the purchase of a building called the Rose Garden Inn by a subsidiary organization of CFAR called Lightcone RG. The balance of the purchase price of the Rose Garden by Lightcone, totalling $20m, came via Slimrock Investments. According to Lightcone founder Oliver Habryka, this is an entity controlled by Jaan Tallin, one of the principals of another nonprofit research group called the Future of Life Institute.

These connections are relevant because of what Lightcone RG did with the Rose Garden Inn after the purchase was complete. The hotel, renamed Lighthaven, became the site for workshops and events - including events featuring advocates of scientific racism.

The Manifest conference was hosted by a firm tied to so-called “prediction markets,” which allow gambling on real-world events, and which are largely illegal in the United States. Manifest 2023 was held at Lighthaven, and as discovered by The Guardian,[ https://www.theguardian.com/technology/article/2024/jun/16/sam-bankman-fried-ftx-eugenics-scientific-racism] featured speakers including Richard Hanania, who had written for avowed white supremacist Richard Spencer’s Alternativeright.com; and Malcolm and Simone Collins, a “pro-natalist” couple with professional ties to Peter Thiel,[ https://www.entrepreneur.com/leadership/the-bizarrely-authoritarian-us-education-system/425668] and who openly referred to themselves as “hipster eugenicists.”

The next year, Manifest 2024 hosted a wide range of speakers, including Eleizer Yudkowsky and an array of tech-world gurus. But further down the agenda were less savory figures including Jonathan Anomaly, author of a 2018 paper called “Defending Eugenics”; Razib Khan, contributor to virulent extreme-right outlet VDare; and Brian Chau, an affiliate of the “effective accelerationist” offshoot of Effective Altruism, whose history of racist comments[ https://www.transformernews.ai/p/alliance-for-the-future-director] included disparaging police murder victim George Floyd.

This star-studded crossover event between rationalism, effective altruism, and eugenics is not as odd as it might seem. Daniel HoSang, a professor of American studies at Yale University, told the Guardian that “the ties between a sector of Silicon Valley investors, effective altruism and a kind of neo-eugenics are subtle but unmistakable. They converge around a belief that nearly everything in society can be reduced to markets and all people can be regarded as bundles of human capital.”

More specifically, as we’ll see, eugenic thinking has deep and powerful roots in the Bay Area, and at Stanford. That legacy has been carried forward by generous funding from the likes of Peter Thiel - a legacy Sam Bankman-Fried willingly joined.

That the event was hosted by a prediction markets firm further points to the Rationalist movement and Effective Altruism’s shared belief in markets as a source of truth. This premise hinges on classical economic principles of supply and demand, according to which Adam Smith’s “Invisible Hand” creates a “correct” price for goods and services. This same logic, as we’ve seen going back to Peter Singer’s original choice between a child and an expensive coat, is constantly at play in Effective Altruism’s calculations of “expected value.” Dollars are a simple, rational way to follow inputs through to a future impact.

In a July of 2024, CFAR filed a response to the FTX estate’s attempted clawback of the $5 million which, in part, supported Lightcone’s eugenics-inflected programming. CFAR’s argument had two interesting points. First, CFAR posits that because most of the funds flowed through the FTX Foundation rather than directly from FTX, the funds were remote from any claims by the FTX debtors. Second, against the idea that the contributions constituted unjust enrichment, CFAR argues that “debtors received value from their philanthropic efforts.”

These are refreshingly honest articulations of the implicit logic of Sam Bankman-Fried’s understanding of philanthropy: That it was a convenient means of conveying money to his allies, and that it was a long-term strategy ultimately intended, less to help people, than to create more “value” for FTX.

Marketing and Relationships

K5 Global Holdings &c. - $500 Million. Customer Funds: At least $173 million

If the $500 million tranche of customer funds sent to the AI startup Anthropic represents Sam Bankman-Fried’s ideological commitments and his willingness to criminally enrich allies within the Effective Altruist orbit, the same amount distributed to firms linked to former Hollywood agent Michael Kives speaks to different elements of his character: his staggering naivete and sheer incompetence.

Bankman-Fried was effectively seduced by Michael Kives’ connections to celebrity and politics, a seduction marked by an invitation to the Super Bowl match between the Los Angeles Rams and Cincinatti Bengals on February 13, 2022. An uncanny photograph taken at the event[ https://nypost.com/wp-content/uploads/sites/2/2023/01/celebrities-sbf-09.jpg] speaks volumes to what happened next.

With Inglewood’s SoFi stadium in the background, Bankman-Fried stands in the center of a scrum of seven other people. Michael Kives, tall and broad, has one arm around Bankman-Fried. An unidentified, smiling woman, Hollywood beautiful, has her head and hand on Bankman-Fried’s shoulder affectionately. At the bottom right of the photo is actress Kate Hudson. To the left stand the actor Orlando Bloom and the musician Katy Perry. At the center of it all, Bankman-Fried looks simultaneously awed and joyful. He smiles broadly as he leans into the scrum of celebrities and influencers around him.

Both Perry and Hudson had been clients of Kives at Creative Artists Agency (CAA) before he left the agency. It was Perry who posted the Super Bowl photo to her social media feed. In a separate post the day before, on February 12, she had posted that “im quitting music and becoming an intern for @ftx_official ok”[ https://www.instagram.com/katyperry/p/CZ5Qw2-vVoW/?img_index=1]

Whether spontaneous or directed by Kives, these gestures of friendship from Perry generated immediate headlines and, it seems unavoidable, contributed to Bankman-Fried’s sudden and overwhelming enthusiasm to work with Kives. According to a lawsuit later filed by FTX bankruptcy administrators, Bankman-Fried on February 15 sent an internal message describing Kives as “probably, the most connected person I’ve ever met,” and “something of a one-stop shop for relationships that we should utilize,” whose “infinite connections” could lead to “partnerships with celebrities” and “[working] with them on Democratic politics.” Kives had been a onetime aide for Bill Clinton, and a major bundler for Hillary Clinton’s disastrous 2016 campaign for the U.S. Presidency.

Bankman-Fried concluded, almost incidentally, that Kives and K5 “Maybe [want] us to invest in them or some stuff, idk.” A relationship with a connector like Kives made undeniable sense to a growing company like FTX, and indeed, Kives reportedly made at least one meaningful contribution very quickly - connecting Bankman-Fried to Bill Clinton, who would appear on stage with the FTX founder at the Crypto Bahamas conference just weeks later, in April.

According to the New York Post, Clinton was paid more than $250,000 for his appearance.[ https://nypost.com/2023/01/19/sam-bankman-frieds-ties-with-the-clintons-helped-dupe-investors/]

But Clinton’s payoff was barely a speck next to the truckload of mostly-stolen cash Bankman-Fried dumped onto Kives. Within less than three weeks of the Super Bowl, $300 million would be wired to K5. In May, another $200 million was funneled through Alameda to K5, and a final tranche of $200 million was transferred in September, just one month before FTX’s collapse. Though some of this money was characterized as an investment, at least $250 million was specifically earmarked as personal payments to Kives and his partner, Bryan Baum.

In his testimony at Bankman-Fried’s criminal trial, Peter Easton only spoke to the path taken by the initial $300 million transfer made in May drawn directly from FTX customer funds. Of this transfer, Easton found that at least $173 million was drawn directly from FTX customer funds.

The subsequent $400 million in payments detailed in the civil suit by the FTX estate were framed as investments in an entity referred to as Mount Olympus, separate from K5 but also controlled by Kives. In addition to the transfers that occurred, Bankman-Fried committed to transferring $1 billion per year to Olympus over the following three years - a commitment that seems disconnected from any budgeting or due dilligence.

This was one of a numerous investments Bankman-Fried directed to outside venture capital or hedge funds, also including former Trump staffer Anthony Scarramucci’s Skybridge Capital. These were all strange, in that even Bankman-Fried and FTX weren’t realistically in a position to responsibly make venture investments, much less farm them out to third parties.

But the Kives-linked transfers are perhaps strangest of all, because there is no evidence whatsoever that Kives was an expert or even skilled investor. Emblematic of the total lack of serious analysis backing Bankman-Fried’s decision, part of the deal outlined in a term sheet allocated $214.5 million for a 38% stake in a Kives-owned entity whose only asset was a minority stake in 818 Tequila, a brand promoted by Kives associate Kendall Jenner. According to an SEC filing made by Kives’ organization in March of 2022, the total asset value was only $2.94 million. The FTX bankruptcy estate concluded that Bankman-Fried effectively paid Kives and his allies more than 200 times what this tequila-slinging entity was worth.

In sum, Bankman-Fried’s dealings with Michael Kives and K5 were a farcical and fraudulent theft in the superficial guise of investment. But where other fraudulent uses of FTX customer funds generally benefitted Bankman-Fried’s longtime associates in the extended Effective Altruism universe, the enrichment of Kives showed more complex, and frankly tragic, motivations. While connections to figures like Bill Clinton had immediate benefits in legitimating FTX in the public eye, Sam’s apparent enthrallment by power, influence, and celebrity accords with John Ray III’s later assessment that, with many of his outlays, Bankman-Fried was “buying friends.”

But even this doesn’t quite get to the sadness of it. Bankman-Fried, after all, had friends, and lovers, in surplus - but he seems to have often found these real, immediate confidantes burdensome and annoying, as with Caroline Ellison’s constant demands to formalize their relationship. Other devoted close associates, like Nishad Singh and Gary Wang, he was happy to manipulate into committing crimes for him. What made Sam happier than real friends, it seems, was celebrities - people whose power and value was affirmed in the public realm of media and symbol. Much as dollars reduced the complexity of existence to manageable and fungible units, Sam found it easier to buy connections to important strangers than to navigate the complexities of people actually close to him.