What this covers
Sam Bankman-Fried discusses what constrains high-impact work across philanthropy, finance, and technology with Tyler Cowen. The core argument is that [coordinator scarcity](concept:Coordinator Scarcity), not money, is the binding constraint on ambitious projects—the person who plants the flag and forces resources and people to converge. Bankman-Fried weaves this claim across three domains: effective altruism (where billions remain unspent because innovation in deployment is scarce), trading (where the fundamental skill is consequentialist quantification of messy systems despite known blindspots), and cryptocurrency (where meme-driven pricing has always been present, even in ostensibly rational assets).
The conversation ranges widely in specificity and terrain. On asset valuation, Bankman-Fried argues that discounted cash flow explains perhaps only 30 percent of a stock like Tesla's price, with social-demand accounting for the rest—a principle made explicit in Dogecoin but present everywhere. On institutional risk, he locates hidden implicit leverage as the true source of fragility, from unreported OTC derivatives before 2008 to nested rehypothecation in DeFi protocols. On utilitarian philosophy, he surface several tensions: the infinite utility problem that breaks down when comparing outcomes with unbounded magnitudes, the replacement cost of life argument that life extension may rank below new life creation, and risk neutrality over social welfare paradoxes. On pandemic preparedness, direct cash transfers, stablecoin regulation, and organizational culture, his positions are concrete and often heterodox but grounded in quantitative reasoning applied to institutions as they actually work.
Bankman-Fried argues that the binding constraint on high-impact altruism and ambitious projects is not money but 'coordinators'—people who will take responsibility and force a project into existence—and that a quantitative, consequentialist mindset applied to messy real-world systems (trading, crypto, philanthropy) is the core scarce skill.
- Effective altruism is less funding-constrained than people claim; the real scarcity is project drivers and coordinators who make things happen
- Good trading and good decision-making both involve quantifying messy situations where you know you are omitting factors
- Crypto asset pricing reveals that meme-driven supply-and-demand pricing was always present even in 'rational' assets
Utilitarian decision-making requires converting outcomes into numbers and estimating quantitative properties despite inherent uncertainty.
- Viewing the world from a utilitarian/consequentialist perspective is inherently quantitative: you ultimately turn outcomes into numbers and decide which number is biggest, doing the best you can to estimate a quantitative property even when you can never be sure you're calculating it correctly.
“there's something very quantitative about how you view the world if you're thinking about things from a utilitarian perspective, from a consequentialist perspective. There's this really big notion of, in the end, you turn things into numbers, and you decide which number is the biggest.”
- Utilitarian expected-value reasoning works cleanly only with finite numbers; once you admit the possibility of infinite utility, you must compare hierarchies of infinity, linearity breaks down, and addition stops behaving normally—a genuinely relevant unknown that no one has a good account of and that leaves serious thinkers more confused rather than satisfied.
“As soon as you say, "What are the odds that there's a way to be infinitely happy? What if infinite utility is a possibility?" ... Now, all of a sudden, we're comparing hierarchies of infinity. Linearity breaks down a little bit here.”
- A risk-neutral Benthamite would play a game offering a 51% chance of doubling the value of Earth and a 49% chance of destroying it, and would keep playing repeatedly, since expected value can continue to grow—though the conclusion depends on the universes being noninteracting and on whether the St. Petersburg-style iteration leads to nonexistence or to enormous value.
“Maybe you St. Petersburg paradox into an enormously valuable existence. That's the other option.”
- The fundamental skill behind successful trading is being able to get quantitative and put numbers on things in a world that is inherently messy, where you know you are forgetting an enormous number of factors but must still build statistical models and make the best probabilistic decisions you can.
“there's a huge amount of power in being quantitative. There's huge power in thinking about things, not just from a qualitative angle, but putting numbers on things, building statistical models.”
Stablecoin regulation should mandate full dollar backing in audited accounts, clarifying a framework problem rather than fixing product design.
- A reason to regulate stablecoins more than a similar undisclosed product is that they are stable in both directions: the holder bears downside risk if the issuer loses money but captures none of the upside if the issuer profits, so this one-sided exposure may justify consumer protection even if sufficient disclosure could otherwise let informed people accept some risk.
“if you're an investor, and you buy a stablecoin, you have downside risk but not upside. If somehow the stablecoin company makes money, you're probably not going to get any of that, but if it loses money, somehow, you're probably on the hook for that.”
- Requiring all stablecoins to be fully backed by dollars held in audited bank accounts would produce a safe, well-understood, well-regulated product just as good as current stablecoins, cleanly solving today's problems which stem from the absence of a clear regulatory framework rather than from the product design itself.
“If you just said, "Look, all stablecoins have to be fully backed by the dollar and have to have audits to confirm that they are in a bank account," that would get a pretty safe product that was well understood, well regulated”
- PayPal and a stablecoin like USDC are more similar than different—both essentially promise to give you back a dollar—which raises the question of why one is regulated as a stablecoin and the other is not, and whether banks' permission to rehypothecate deposits should extend to stablecoin issuers.
“what is the difference between PayPal and USDC? I guess there's some differences, but I think there are more similarities than there are differences, to be honest.”
- Because DeFi runs on-chain, most protocols only do what they literally say and rehypothecation is largely visible, so the leverage they introduce is roughly what they state; however, hidden leverage can be nested—for example DAI, an overcollateralized algorithmic stablecoin backed by volatile crypto assets, can itself be posted as collateral on borrow-lending protocols, a traceable but not easily traced form of rehypothecation.
“Most DeFi protocols are not doing things beyond what they literally say they're doing, and so the amount of leverage they introduce into the system mostly is what they say they are.”
All systems contain hidden implicit leverage beyond their explicit leverage that must be located to truly understand them.
- Any system or institution typically contains forms of hidden, implicit leverage beyond its explicit leverage, and you don't truly understand the institution until you locate where that implicit leverage lives—just as the 2008 collapse was driven by unreported, repeatedly rehypothecated OTC swaps whose total notional no one tracked.
“if you look at any system, any institution, typically there are forms of hidden implicit leverage in those institutions ... you don't understand the institution until you understand where's the implicit leverage in this game.”
- For a blockchain to replace substantial fractions of global infrastructure it must process on the order of millions of transactions per second, since a billion-user application generates roughly that volume; Ethereum today handles about ten TPS (six orders of magnitude short) and even sharded Eth2 yields only single-digit thousands per shard, which is insufficient because a single order book or company cannot be split across shards that don't communicate well.
“Ethereum cannot support millions of transactions per second. Today, it can support ten. Ten is the number of transactions per second it can support. Ten is off by six orders of magnitude from where it needs to get.”
- Crypto has shifted from being a near-total afterthought for financial regulators a year or two ago to occupying roughly half their attention today, making understanding the regulatory intentions and trajectories across some 200 countries a huge fraction of a crypto business's work.
“If you ask a random financial regulator somewhere in the world today, "What fraction of your time do you spend thinking about crypto?" "Ah, 50 percent," an enormous fraction.”
Effective institutions succeed by aligning culture around finding the truth first, then managing disputes through clear hierarchical decision-making.
- The biggest weakness of most managers is jumping straight into managing disputes without first determining the right object-level answer to the underlying problem; good management requires first doing the research to find the truth, then building a culture where everyone is aligned that getting the right ultimate decision is what matters above any other argument.
“That is legitimately, I think, the biggest weakness of a lot of managers. They jump straight into how to manage without first thinking, "Well, what's the right answer in terms of the object-level problem we're trying to solve?"”
- Disputes require a clear process with a known final decision-maker (an odd number of people, ideally one) so that if a civil war were to break out, everyone can foresee how it would end and therefore the conflict never actually occurs; the downsides of this concentration are outweighed by the upside of resolvable disputes.
“If a civil war were to break out, everyone could just look at how it's going to end, and now there's no civil war. There are downsides to that, but I think they're outweighed by the upsides from it.”
- Consistently overperforming organizations like the Patriots succeed because leaders like Belichick build a culture centered on outthinking opponents and choosing the play that optimizes winning rather than what is popular or conventional—evidenced by deploying obscure running backs who then lead the league in rushing.
“He cares about, "What's the play that will optimize our chances of winning, and how do we do that?"”
Pandemic preparedness is an urgent cause because COVID exposed and did not remedy the world's uncoordinated, disorganized response capacity.
- Pandemic preparedness is currently one of the most important causes because COVID revealed the world has no coordinated capacity to respond—our only mode is to freak out and shut things down in a disorganized way—and we have not improved since, so the next, potentially more lethal pandemic would find us just as unprepared with much higher death tolls.
“I think the single clearest takeaway from COVID is that we don't have our sh-t together as a world. That's what's become clearest, is that there is no magical wizard behind the curtain who makes everything go okay.”
- Beyond lobbying for pandemic spending, a high-leverage use of philanthropic dollars is funding the build-up of vaccine capacity, stockpiles, distribution networks, and early detection systems that could be deployed within a couple of days of a new pandemic appearing.
“it's funding efforts to build up vaccine capacity, vaccine stockpiles, distribution networks, early detection systems, and other things that could be done rapidly, that could be done within a couple days of a new pandemic appearing.”
Project success depends on a rare coordinator who forces convergence, funding and talent expansion beyond individual inputs.
- The truly scarce resource for new altruistic projects is not funding, founders, employees, or ideas individually, but the coordinator who plants the flag and forces them all to converge at the same place and time—someone who takes total responsibility, guarantees the pieces will be there, and thereby makes everyone take the leap together to start the project.
“What's it take for a new project to start up, a new altruistically minded project? It takes funding. It takes founders. It takes employees. It takes ideas, and they all have to come together at the same place and the same time and meet up.”
- A major component of successful crypto trading is operational tenacity—fighting through bizarre, Byzantine bureaucratic hoops, never letting them get the best of you, and continually generating and testing creative solutions until you reach the goal.
“You had to jump through some really bizarre, Byzantine hoops in order to get where you needed to go. That meant that you had to fight through bureaucracies. You could never let them get the best of you.”
- From a pure Benthamite standpoint, life extension research is not the most pressing problem because the replacement cost of human life is low—we have flourished without it and can grow utility more cheaply by creating new lives than by saving existing ones—though feedback mechanisms and international coordination effects complicate any simple birth-subsidy conclusion.
“I find that I'm not very compelled by life extension research for the exact reason you've said. I think that it is really cool, really f-cking cool, but I'm not sure it's the most pressing problem for the world.”
- Just as companies tend to become dysfunctional as they grow larger, the best-governed countries tend to have fewer than 20 million people, which is a consideration pointing against simply maximizing population.
“I will say that whenever companies start to get bigger, they become a sh-tshow.”
- Although a world where most meat comes from happy animals is conceivable, in today's environment almost all meat is inhumane meat—by count only a tiny fraction comes from happy animals—so the option to eat humane meat doesn't currently buy you much.
“today, almost all meat is inhumane meat. If you look by count, a very, very tiny fraction of it is actually happy.”
- Although the prevailing meme is that effective altruism is no longer funding-constrained thanks to large donors, this is overstated; there are billions a year that could be spent extremely effectively, and the apparent lack of funding constraint is partly a symptom of insufficient innovation in great ways to spend money.
“That's been a pretty big update. There certainly is a lot more funding than there was before, but saying it's not funding-constrained — I think people go too far in that, and to some extent that's a symptom of not thinking hard enough about great ways to spend money.”
- The underappreciated strength of the Bahamas is the combination of a nimble, small state that has a lot to offer with a robust democracy—evidenced by a recent extremely peaceful transfer of power after an election—which is not true of many places people put in a similar category, and which gives it real potential for sustained high economic growth.
“it's the combination of a nimble, small state that has a lot going for it, a lot to offer, and also is robustly democratic, which is fairly powerful.”
- Direct cash transfers are always within an order of magnitude of the best thing you can do with philanthropic money, making them a strong default for improving life in a place like the Bahamas alongside investment in healthcare infrastructure.
“direct cash transfers are always within an order of magnitude of the best thing you could do.”
- Over the last several years there has been a real step-change in fake meat from mediocre to decent, led by Beyond and Impossible burgers, which now account for a huge fraction of veggie burgers sold globally because targeted research made them taste much meatier than prior alternatives.
“In the last few years, there's been a real step-change in fake meat, from pretty mediocre to pretty decent, and Beyond and Impossible have been leading the way in that.”
Bitcoin and meme assets reveal that social demand, not discounted cash flow, drives much of asset pricing across all markets.
- Meme coins and meme stocks make visible for the first time what was always there in ostensibly rational assets: discounted cash flow explains perhaps only 30 percent of a stock like Tesla's market cap, with social-media-amplified, meme-like demand accounting for much of the rest, so crypto forces a reexamination of how everything is priced.
“It's when we look at the meme stocks and the meme coins that we feel like we can see the answer for ourselves for the first time, but it was always there in the other stocks as well, and social media has been amplifying this all over the place.”
- Pricing Bitcoin by asking what fraction of gold's value it could capture is genuinely useful as one sanity check: at 1% of gold it would be an obvious buy because the odds of Bitcoin becoming a new gold far exceed 1%, while at 10x gold it would be hard to justify—and a parallel check is ranking Bitcoin against world currencies, where it already plausibly outranks the Swiss franc.
“if you told me that Bitcoin was currently trading at 1 percent of gold, I would just be like, "F-ck it, I'm going to buy." ... What are the odds Bitcoin becomes a new gold? Is it 1 percent? No, it's more than 1 percent.”
- For most crypto assets there is not even a pretense of a rational pricing theory; Dogecoin is the purest meme coin precisely because it openly discards the pretense of discounted cash flow, leaving only supply and demand driven by memes—a feature that is bizarre and dangerous but also powerful.
“There is no sense in which any reasonable person could look at Dogecoin and be like, "Yes, discounted cash flow."”
- Elon Musk's most successful product is not the electric car or rocket but TSLA, the stock ticker itself—a product people want to own that is priced not primarily by discounted cash flow but by the same meme-driven demand seen in crypto.
“It's when we look at the meme stocks and the meme coins that we feel like we can see the answer for ourselves for the first time, but it was always there in the other stocks as well, and social media has been amplifying this all over the place.”
Categorical rules like complete vegetarianism reduce cognitive burden more effectively than gradual moderation attempts.
- Eliminating a behavior entirely (going cold turkey) can be far easier than gradually moderating it, because a categorical rule removes the recurring cognitive load of deciding case-by-case what is acceptable—illustrated by how going fully vegetarian succeeded after six months of failed gradual attempts.
“It was way easier for me once I went cold turkey than it was for me to try to wean myself off it because the cognitive load of every time that I went to eat, having to decide what I could and couldn't eat and should and shouldn't eat was just overwhelming me.”