Sam Bankman-Fried's infamous analogy for yield farming helps explain the tech boom

A lot of people were yield farming without knowing it during the 2010s.

On November 11, the cryptocurrency exchange FTX declared bankruptcy, having lost (or perhaps stolen) around $8 billion in customer funds. After FTX’s collapse, a lot of people have been talking about an interview that CEO Sam Bankman-Fried (SBF) did on Bloomberg’s Odd Lots podcast back in May.

At one point in the interview, Bloomberg’s Matt Levine asked SBF to explain yield farming, an investment strategy that enables cryptocurrency owners to generate earnings using their digital tokens. When SBF finished his explanation—which I’ll describe in more detail shortly—Levine and the podcast’s co-hosts seemed stunned.

“I think of myself as like a fairly cynical person, and that was so much more cynical than how I would have described farming,” Levine said. “You're just like well I'm in the Ponzi business, and it's pretty good.”

SBF responded that this was a “pretty reasonable response” with “a depressing amount of validity.”

SBF now faces accusations that he ran a literal Ponzi scheme, so many people have understandably interpreted this as effectively a confession of his own guilt.

But SBF was actually making a broader point about the speculative dynamics that drove last year’s cryptocurrency boom. And the dynamic he identified didn’t just happen in crypto—something similar was happening in the broader world of venture-backed technology startups during the late 2010s.

The Box and the X Token

Sam Bankman-Fried. (Photo by Alex Wong/Getty Images)

Yield farmers deposit cryptocurrencies with decentralized finance platforms and earn returns in much the same way that you can deposit dollars in a conventional savings account and earn interest. In 2021, yield farmers were earning a lot more than you could get from conventional financial products, which made some people wonder if something shady was going on. In that May interview, Levine asked SBF to explain how yield farming worked.

In response, SBF sketched out a “toy model” of yield farming involving a box that could hold digital assets. “For now ignore what it does,” SBF told his Bloomberg hosts, “or pretend it does literally nothing” aside from storing people’s tokens.

This hypothetical box was associated with an equally hypothetical X Token, SBF said. X Token holders get to vote on changes to the box, and if the box ever turns a profit, X Token holders would get a share. Every day, the box creates a bunch of new X Tokens and gives them to anyone who has stored money inside the box.

In SBF’s telling, this leads to a speculative flywheel effect: people put more money in the box to get more X tokens. The flood of cash into the box makes people more optimistic about the box and its associated token, so the value of the X Token rises. As X Tokens get more valuable, people put still more money in the box and the cycle continues.

While SBF’s hypothetical box does literally nothing, that’s not quite true of the real boxes SBF was trying to explain. Certainly the creator of a box doesn’t portray it that way. According to SBF, a box creator will “probably dress it up to look like a life-changing, world-altering protocol that's going to replace all the big banks in 38 days or whatever.”

SBF said he was specifically thinking about crypto platforms like Uniswap, Aave, and Compound that allow people to trade or borrow cryptocurrencies. Let’s take Uniswap as an example. Uniswap is a decentralized exchange that lets people trade one cryptocurrency for another. Rather than using a conventional company as an intermediary, Uniswap runs its code automatically on the Ethereum blockchain.

Anyone can lend tokens to a Uniswap pool to provide greater liquidity and more attractive prices for traders. People who do this are rewarded with UNI, tokens that work a bit like shares of stock in Uniswap. UNI holders have the right to vote on proposals to change how Uniswap works. In SBF’s analogy, Uniswap is the box and UNI tokens are the X tokens. 

Uniswap doesn’t literally do nothing—lots of people use it to trade one token for another. But SBF’s point was that there are lots of platforms out there that employ this kind of token reward scheme. And in many cases the crypto industry wound up with a tail-wagging-the-dog situation where people were participating in new platforms mainly to earn tokens, without caring very much about the underlying functionality or value. The result was a speculative bubble that has been deflating over the last few months.

Yield farming in the real world

As I was pondering this analogy, I realized that I experienced something like it way back in 2014. That year I spent a week driving for Lyft and wrote about the experience for Vox.

Over 50 hours of work I completed 54 rides and earned $594 in fares. At the time, Lyft had a promotion guaranteeing new drivers at least $1,500 for their first week of driving. So Lyft kicked in another $906.

This was obviously not a sustainable business model. But at the time Lyft’s priority was growth, not profits. And Lyft had recently raised $250 million, so it could afford to spend lavishly to attract new drivers.

SBF’s box analogy fits this situation pretty well: Lyft accelerated its growth by offering subsidies to drivers in much the same way that Uniswap accelerated growth by offering subsidies to liquidity providers. In both cases, the subsidies were financed by people who were making speculative bets on the future of the platform. For Lyft that was venture capitalists who held Lyft shares. For Uniswap it was people who held UNI tokens.

An investment like this isn’t necessarily irrational. Lyft investors were betting that there were powerful network effects in the taxi business—that the more drivers on the road, the more valuable the network will be for passengers, and vice versa. Today Lyft is still struggling to turn a profit, but I think it will ultimately prove to be a viable business—albeit not as profitable as early investors hoped.

Other tech platforms pursued similar business models during the late 2010s. One of the most hilarious examples was Doordash, which sometimes offered meals to people below cost. In 2019, one pizza shop owner discovered that Doordash had started offering a delivery service for his restaurant without his permission and was offering a $24 pizza for $16. His friend realized that this was an arbitrage opportunity.

If someone could pay Doordash $16 a pizza, and Doordash would pay his restaurant $24 a pizza, then he should clearly just order pizzas himself via Doordash, all day long. You'd net a clean $8 profit per pizza.

Of course, the pizzas themselves weren’t free—the ingredients cost around $7 per pizza—so the profit per pizza was initially small. But the pizza shop owner eventually figured out that he could just throw uncooked pizza dough into the box, allowing him to earn something like $7 per order.

This is the brick-and-mortar equivalent of yield farming. Lyft and DoorDash were paying people above-market rates (or offering them below-market pizzas) to incentivize more people to join their platforms. And as more people joined their platforms, the value of their stock went up, which allowed them to raise more money and pay more subsidies.

As in a Ponzi scheme, this kind of growth seems unsustainable. But there’s also an important difference: Ponzi schemes involve stealing the money of unsuspecting customers. In contrast, DoorDash and Lyft were perfectly transparent about what they were doing. The companies’ investors willingly dumped billions of dollars into these companies in hopes of earning longer-term profits.

By the same token, the phrase “Ponzi scheme” doesn’t fit many of the speculative cryptocurrency investments SBF was describing with his box analogy. People who chose to buy and hold UNI tokens weren’t being defrauded. They were making a speculative bet on the future of the Uniswap platform in much the same way Lyft investors were betting on Lyft.

With that said, I do think there’s a potentially significant difference: taxi rides and restaurant meals have clear utility to ordinary people. In contrast, platforms like Uniswap seem to mostly facilitate cryptocurrency speculation, a zero-sum activity that produces little to no value for the broader society. It’s hard to see how it can support a thriving cryptocurrency industry over the long run.

To be fair, there is lots of speculation in conventional financial markets too. But conventional financial markets also fund the creation of new businesses, new homes, new factories, and so forth. As far as I can tell, crypto’s “decentralized finance” industry doesn’t do anything like that.

I’ll be thinking about this question more in the next couple of days as I write a longer piece about the state of cryptocurrency regulation. I’d love to know what you think. What do you see as the most significant applications of cryptocurrency? What would we lose if cryptocurrency networks shut down?