Magic beans, Bahamian penthouses, old-fashioned fraud and other important SBF-inspired insights. A postscript to Bloomberg Businessweek’s

The Crypto Story.

By Matt Levine

On Oct. 25, Bloomberg Businessweek published “The Crypto Story,” a cover-to-cover issue of the magazine that I wrote about what crypto is and what it all might mean. Over the summer prices had crashed and several prominent crypto companies had failed, and it looked like the popular enthusiasm for crypto was finally fading. People called it a “crypto winter.” Still, I wrote, “it’s a good time to be talking about crypto. There’s a pause; there’s some repose.”

That was true for, like, an afternoon? Just two weeks later, in early November, FTX—one of the biggest and most prominent crypto exchanges—imploded. By Nov. 11 it was bankrupt. Its founder Sam Bankman-Fried and other executives were soon charged with fraud.

The crypto winter got colder and darker. FTX and Bankman-Fried—“SBF,” everyone in crypto called him—had been important to the crypto industry. FTX had positioned itself as a well-run exchange that wanted to work with regulators; SBF often spoke to regulators and to Congress about how crypto should be regulated, and they tended to listen to him. When crypto companies failed over the summer, SBF often ended up bailing them out, shoring up confidence in crypto. That confidence is now doubly betrayed.

Possibly another good time to reflect? So, here, I will. Consider this a postscript to “The Crypto Story.”

A Game

One imperfect but useful way to think about crypto is that it allowed for the creation of a toy financial system. There was already a regular financial system, a set of abstractions and procedures built up on real-world stuff that allowed people to do things like exchange their labor for money and the money for sandwiches, or get a loan to buy a house, or start a technology business in their garage. That system grew up over time, in path-dependent ways; it was fragmented and complicated and embedded in society and history. Different bits of it had different cultures and practices and were regulated differently; the regulation had also accreted haphazardly over time, and it could feel arbitrary and constraining.

And then crypto came along with a new set of stuff to do finance to. This stuff is so clean and new and shiny. It lives entirely on computers; you never have to worry about how to foreclose on a house or take delivery of 5,000 bushels of soybeans. And it lives on really user-friendly computers: The assets are created and sent between users on permissionless blockchains, and anyone who has a clever idea for how to trade them can implement it. The culture of crypto skews young and tech-savvy and optimistic; people want to try stuff, and they want everyone’s stuff to work. Also, for most of its history, the overall value of crypto has kept going up, which meant it was pretty easy to make your stuff work. If your strategy was “Buy a lot of crypto and then do some mumbo-jumbo,” the mumbo-jumbo might have been good or bad, but you probably made money just from buying a lot of crypto.

Also, because it has so little history, crypto came with very little regulation. If you wanted to build a new system for trading US stocks, there were a lot of detailed technical rules that you’d have to work through, rules that might get in the way of your ideas, rules that you might think were arbitrary and outdated and bad. If you wanted to build a new system for trading crypto, you could kind of just code it up and see what happened. Then you could go to the regulators and say, “Here’s how the rules for crypto should work,” and they might listen to you. (Or they might not. They might argue, as many regulators did, that crypto is largely covered by existing rules, and that you were breaking them. But you might go ahead anyway, or move to a different country with friendlier regulators.) There was a sense of freedom—freedom from regulation, but also freedom to invent new regulation—that was very exciting.[1]

Also, a pretty prominent cryptocurrency is Dogecoin, and nonfungible tokens—pixelated images of monkeys, in some cases—sold for millions of dollars. Crypto is perhaps a bit more accepting of absurdity than the traditional financial system is.

A result of all this is that if you were a young trader or developer working in traditional finance, it would not seem insane to quit your high-paying job and move to a friendly island with a few of your friends to build a crypto exchange. There you are, you and your friends, hanging out, playing board games together, coding up your exchange. And then there it is, out in the world, with people using it. You don’t have to run it by some boss with years of old-school financial experience who doesn’t get your vision. You don’t have to run it by regulators or auditors or lawyers. You just do it, on your computer, with your friends.

And it could all feel like a game; it could all feel unreal. It is unreal. You are trading tokens, they live on computers, many of them didn’t exist a year ago, none of them existed 15 years ago, some of them are Dogecoin, and what makes them valuable is just people’s shared agreement to ascribe value to them. You do not have to figure out how to interface with the real world, how to perfect security interests in oil cargos or evaluate the earnings quality of a ball-bearings manufacturer. The tokens beep and boop, and the balance in your account goes up.

This game was played by young people who came from the world of traditional finance, from banks and hedge funds and quantitative proprietary trading firms, people who already liked finance and wanted to play with a toy version of it they could shape however they wanted. And—because it’s the game they knew—they ended up replicating much of the world that they came from, only with crypto as the subject matter. Margin lending and futures exchanges, hedge funds and proprietary trading firms, shadow banks and over-the-counter derivatives, just like the stuff they were used to, but for crypto. It was like a fantastical version of their old jobs, a new financial system with all the bits of the old system that they liked, none of the bits they disliked, and some new bits that they dreamed up because they thought they might be cool.

Or, as I wrote in “The Crypto Story” (about DeFi, but it’s largely true of centralized crypto exchanges like FTX, too):

In some crude sense, what decentralized finance is is a big community of people who get together to pretend to trade financial assets—or, rather, who trade financial assets in a sort of virtual world. They’ve built derivatives exchanges and secured lending protocols and new ways to do market-making, but instead of trading stocks or bonds they trade tokens that they made up. And those tokens are valuable … in part because DeFi is itself an online community, or cluster of communities, and the tokens it trades are points in that community. If you build a cool trading platform or execute a cool trade, you’ll earn tokens, which you can spend on other cool trading platforms or trades. Talented financial traders are willing to work on projects to get those tokens. If you had some of those tokens, you could hire those traders.

A Dangerous Game

There are two reasons this might be bad.

One is that, if you are playing a game, you might not take it very seriously. You might call up your trader buddies at your old firm and say, “Hey, come over here, it’s so fun, we can just make stuff up, and the money is good,” but you will not call up the compliance people at your old firm and say that. For one thing, that is not an appealing pitch to compliance people. For another thing, the compliance people are what made your old firm less fun, always nagging you about compliance. Now you don’t need to comply! Now you get to make stuff up.

Same with accountants, it turns out. It’s very hard for crypto firms to produce audited financial statements.

You might find yourself building out a snazzy user interface and a fast, clever trading algorithm, because those are fun and profitable things to do, but you might find yourself neglecting the accounting department, because that’s boring. You might get really good at attracting customer money, with your snazzy interface and your sense of fun, but also really bad at keeping track of the customer money, with your lack of accountants and your sense of fun.

Also, if the game stops going your way, you might be tempted to reprogram it, to cheat. In traditional finance, there are exchanges and clearinghouses and prime brokers and market makers, and they tend to be separate companies serving different purposes. This is part of what makes the traditional finance system feel slow-moving and annoying. To trade, you need relationships with all these different entities; there is so much bureaucracy, so many contracts, so many people who can object to what you are doing.

In crypto it is common for one exchange to do all of these things, to run the exchange that matches trades and also the website that takes customer orders and also the bank that lends customers money and also the market maker that buys what customers are selling and sells what they’re buying. This, in many ways, will feel like a better user experience; the customer can go to one website that does everything. It is also a better experience for the finance people building the game: You can just think of the best possible way to trade and offer it to customers, without dealing with any middlemen.

But then, if your market-making firm stops making money on the exchange that you run, you might tweak how the exchange is run so that you can make more money from your customers. Or, if your market-making firm loses a lot of money, you might tweak how the margin-lending function works so that, uh, you can take a lot of money from your customers and “lend” it to your market-making firm. This will not be a very good experience for your users in the end. But it’s all a game, anyway, to you.

Play Money

Here’s another reason this is bad. The regular financial system is built up from things in the real world, things that have some practical value and produce some reliable cash flows. A synthetic collateralized debt obligation of mortgage-backed securities is a very abstract bit of financial engineering, but it is the output of a complicated machine, and the inputs to the machine are people who live in houses sending monthly checks to pay for those houses. There’s a lot of math and judgment involved in structuring and pricing the thing, but there is also a house. The value of a synthetic CDO tranche can go to zero, but the value of all the stuff that goes into the machine can’t go to zero as long as people need houses.

The crypto financial system—this game—is built on crypto tokens. Some of those tokens are time-tested fixtures of the economic system that have performed robustly through many cycles, by which I mean that Bitcoin has been around since 2009? Others have less of a track record.

Some are exchange tokens: If you run a crypto exchange, trading crypto tokens that people just made up, it might occur to you to make up some tokens of your own. If you run the Matt Exchange, you might make up a Mattcoin, and then let people trade it on your exchange. Mattcoins would have some economic link to the functioning of the exchange: People who own Mattcoins might be able to use them to pay trading fees, or they might get a discount on trading fees for owning a lot of Mattcoins, or you might promise to use a portion of the trading fees to buy Mattcoins to prop up the price. The cash flows are not very important, though, at least not in the good times; the basic point is that Mattcoin is an indicator of confidence in the Matt Exchange, and when confidence is high so is the price.

The good news, for you, is that if you invent Mattcoins you can give yourself as many as you want. Issue a million Mattcoins to customers, keep a billion for yourself, see what happens. If a few customers trade the Mattcoins for a few dollars, well, technically you’re a billionaire. (The simple math is: The latest token price times supply of tokens equals market value.) If a few customers want to sell a few Mattcoins, and you buy them for a few dollars, well, you’ve spent a few dollars to become a billionaire.

These are obvious points—crypto tokens are worth what people will pay for them, the market capitalization of a lightly traded crypto token is not necessarily proof of its real value—but they make it uncomfortable to build a crypto financial system modeled after the real one. The real financial system is built on debt. The basic thing that banks and broker-dealers and hedge funds and proprietary trading firms do is borrow money to buy and sell more of the things they are buying and selling. Specifically, they borrow against those things: They put up stocks or bonds or commodities or mortgages or whatever as collateral to get money to buy more of them.[2]

The sophisticated young people who came to crypto from traditional finance also want to borrow, except they want to borrow against crypto. They created ways to lend against crypto. Some were lending platforms—BlockFi, Celsius, Voyager—that attracted customers with the promise that they could lend out their crypto and get high returns; those platforms loaned their customers’ crypto or dollars to crypto trading firms that wanted leverage.

But there are also crypto exchanges—prominently, Binance and FTX, before its implosion—that let customers buy crypto with leverage, often using futures contracts. Intuitively, the exchanges borrow money from some customers to lend to other customers, or rather they lend the customers’ crypto to each other. One customer will deposit $1,000 worth of Bitcoin to borrow $500, another customer will deposit $1,000 worth of dollars to borrow $500 worth of Bitcoin, and the exchange will take the first customer’s Bitcoin and lend it to the second and vice versa.

This creates a lot of risk for the exchange. As I wrote in “The Crypto Story”:

A crypto exchange may have customers with big leveraged bets on Bitcoin rising (they’re “long,” in the language of finance) and customers with big leveraged bets against Bitcoin (they’re “short”). If Bitcoin moves too far in one direction too quickly, then the long (or short) customers will be out of money, which means there won’t be money to pay back the short (or long) customers on the other side. The exchange has to think about how volatile its assets are, set leverage limits so blowups are unlikely, and monitor leverage levels to ensure no one is in imminent danger of blowing up. If someone is likely to blow up, the exchange has to seize their collateral and sell it, ideally in an intelligent way that doesn’t destabilize the market too much. And in periods of high volatility the exchange might shut down trading rather than deal with all this.

That’s true of traditional exchanges, too; this year the London Metals Exchange had a very similar set of problems with nickel. It had to shut down trading in nickel futures for more than a week, because big customers were in danger of blowing up, and because it concluded that the price had gotten too far away from economic fundamentals.

What makes this problem so hard in a crypto financial system is that there are no economic fundamentals. There are cases of popular well-known crypto tokens that are worth billions of dollars one day and nothing a week later. If someone comes to you and says, “I have $3 billion worth of Mattcoins. Will you lend me $1 billion against them?” you might say yes, because that’s the sort of thing you do in finance; you lend money against collateral at some discount to its market price so that even if the market goes down a bit you’ll still get your money back. But Mattcoin might go to zero tomorrow! And then where will you be?

Here’s a worst case:

1. You run a crypto exchange, the Matt Exchange.

2. You issue your own token, Mattcoin.

3. You give some Mattcoins to customers, they trade a little bit, they have a market price, whatever.

4. But you give like 95% of the Mattcoins to your own affiliated hedge fund, the Matt Fund, which does the market-making on your exchange.

5. The Matt Fund says, “Hey, we have billions of dollars of Mattcoins. Can we borrow billions of dollars of real money secured by our Mattcoins?”

6. They say this to you, since you run a crypto exchange and have a lot of money to lend out.

7. You are like, “Sure, these Mattcoins are good collateral! Invented them myself!”

8. You lend the Matt Fund billions of dollars, dollars that effectively belong to your other customers.

9. The Matt Fund loses the money on bad trades, or spends it on political donations and philanthropy and snacks.

10. People find out.

11. Now the Mattcoins are worthless.

12. The Matt Fund owes you the money, but doesn’t have it. And you have the collateral, but it’s worthless.

You have invented some play money, and then you have used it to lend yourself real money. And then the game ends and you don’t have the real money.

Well, FTX

Sam Bankman-Fried founded Alameda Research, a crypto trading firm, in 2017. [3] Alameda started out doing crypto arbitrage trading, buying Bitcoin on exchanges where it was cheap and selling it on exchanges where it was expensive. This is not too dissimilar to what SBF was doing at his old job at Jane Street Capital, a prominent quantitative firm that trades things like stocks and exchange-traded funds. Later, Alameda expanded into riskier and less clever trades. Famously, by 2021, Alameda was buying Dogecoin when Elon Musk tweeted about it. Any idiot could do that, but you needed certain rare skills to borrow millions of dollars to do it.

Bankman-Fried founded FTX Trading, a cryptocurrency exchange, in 2019, because he knew the kind of exchange that he wanted to trade on, and because it was crypto so he could just make it himself. FTX apparently stands for “Futures Exchange,” and it focused on offering futures and other leveraged trades: You went to FTX because you had a million dollars and wanted to bet $20 million on Bitcoin, and FTX would very much let you. FTX quickly developed a reputation as a good cryptocurrency exchange. It had good technology, a good website; trading was fast and efficient. It offered desirable products and lots of leverage. It seemed to have a good risk-management system. It gave crypto people what they wanted.

But it also gave the public what it wanted. Bankman-Fried was, I’m sorry, the kind of colorful character I tried to avoid writing about in the “The Crypto Story”; he wore shorts everywhere, had messy hair, and projected a nerdy enthusiasm. Everyone called him “SBF.” FTX advertised extensively, with big celebrity endorsers. And SBF was a compelling advocate for better crypto regulation, meeting with regulators and Congress (in a suit) to push his vision. The regulators and politicians liked his ideas. To be fair, the politicians also liked his money: As FTX grew, SBF became a billionaire while still in his 20s, and became a prominent political donor.

So FTX attracted a lot of customer money, and seemed like it might be a good and upstanding crypto exchange. Oh sure, it was located in the Bahamas because its offerings were not legal in the US (though it had a subsidiary, FTX US, that was regulated in the US), and sure, it didn’t have public audited financial statements. And sure, there were questions about the relationship between FTX and Alameda, which was still around, still doing a lot of trading on FTX, and still mostly owned by SBF, though he stepped down as Alameda’s chief executive officer. But SBF was rich and famous and on magazine covers, and he gave a lot of interviews where he sounded like a good guy.

Then crypto prices crashed this summer, and a bunch of crypto lenders—Celsius, BlockFi, Voyager—blew up, freezing customer withdrawals. In several cases, SBF got involved, lending them money or offering to bail them out so that customers could get their money back. This helped FTX’s reputation: It was a stabilizing force in crypto, an exchange that stayed strong when weaker firms broke, and that used its strength to help the weaker firms’ customers and to maintain confidence in the crypto financial system generally. People started calling SBF “JPEG Morgan.”

FTX also had an exchange token, called FTT. Actually it had another one, called SRM; FTX and Alameda developed a decentralized finance exchange protocol called Serum, and issued SRM tokens for that protocol, and kept most of the tokens themselves. Alameda ended up with piles of FTT and SRM tokens, which it got for free, and which technically had a market value of billions of dollars, based on recent trading prices. The value of Alameda’s FTT and SRM tokens was much greater than the total market value of all FTT and SRM tokens held by anyone other than Alameda; the market value was based on just a tiny stub of tokens that traded freely. Also, Alameda did a lot of that trading: If you were selling FTT or SRM, there’s a good chance that Alameda was buying. That helped to keep the prices up.

Then FTX collapsed. On Nov. 2 crypto publication CoinDesk published a story reporting that Alameda’s biggest asset—the thing propping up its ability to borrow money to do its trading—was FTT. On Nov. 6, Changpeng “CZ” Zhao, CEO of Binance, the biggest crypto exchange, tweeted that Binance was going to dump its holdings of FTT. The price of FTT dropped, and FTX customers got nervous. They started withdrawing their money. “FTX is fine,” Bankman-Fried tweeted.

And then it wasn’t. On Nov. 11, FTX filed for bankruptcy. On Nov. 12 the Financial Times published FTX’s balance sheet, and it was a thing of nightmares. By Bankman-Fried’s account, FTX owed customers about $8.9 billion, and had about $9.6 billion in assets. But billions of dollars of those assets consisted of FTT, SRM and other tokens that FTX had invented itself. Those tokens were valuable when people believed in FTX; now they were worthless. FTX had no way to pay its customers back.

On Dec. 12, Bankman-Fried was arrested in the Bahamas on US fraud charges; he was soon extradited to the US. Two of his associates, Gary Wang, who ran technology at FTX, and Caroline Ellison, who succeeded him as CEO of Alameda, have pleaded guilty to fraud and are cooperating with prosecutors.

As US prosecutors, the Securities and Exchange Commission and the Commodity Futures Trading Commission tell it, the core of the fraud was that FTX gave Alameda free use of FTX customer money. It was allowed to “go negative in coins,” in Ellison’s phrase; if the money in its account on FTX was less than zero—even billions of dollars less than zero—then FTX would not complain. As far as FTX was concerned, if its regular customers had $10 billion deposited on the exchange, and Alameda had negative $8 billion, and there was $2 billion of actual cash, then the books balanced and everything was fine: $10 billion minus $8 billion equals $2 billion of net liabilities to customers, and FTX had $2 billion to cover it. But the books were not balanced. If all the regular customers wanted their money back, there was only $2 billion of actual money to give them. Unless Alameda paid off the negative $8 billion. It did not.

Earlier this year, Alameda was borrowing some money from FTX (by “going negative”), and it was also borrowing a lot of money from other lenders, using its big stash of FTT tokens as collateral. As crypto prices fell, the other lenders wanted their money back, and Alameda didn’t have it, because it had used it to make long-term venture capital bets, or because it had lost it. Faced with the risk of Alameda going bust, SBF and his lieutenants allegedly decided instead to bail it out with FTX money—with FTX’s customers’ money. Alameda’s debt to FTX ballooned, and when FTX’s normal customers started asking for their money back, the money wasn’t there. Just a pile of FTT and SRM tokens, magic beans that had lost their magic.

It’s still not entirely clear where the money went. The new CEO of FTX, who came in after the implosion and is in charge of sorting out the mess, told the court that “Never in my career have I seen such a complete failure of corporate controls and such a complete absence of trustworthy financial information as occurred here.” Sounds bad! His career includes overseeing the bankruptcy of Enron.

But the money didn’t go anywhere good. Alameda lost some money—possibly quite a lot—on bad crypto trades. It invested some money in bailing out other crypto firms. It made a lot of venture capital investments. It spent money freely on perks and real estate for employees. It made personal loans to SBF and other executives, some of which seem to have funded their philanthropic and political donations.

What happened at FTX? “They stole the money” seems to be a true but insufficient answer. I think that part of the answer is that they found, and helped to build, a toy financial system, and they played with it. They didn’t take the game too seriously; they didn’t spend a lot of energy hiring accountants and compliance people, because that is not the fun part of finance. They built clever systems for margin lending and risk management, because it is fun to build an idealized trading system from scratch. But they also exempted themselves—Alameda—from that system, because it was just a game. In the real world, if you run a hedge fund and your balance becomes negative, the game is over. At FTX, when Alameda’s balance became negative, it got to keep playing.

FTX’s and Alameda’s senior employees lived together in a big penthouse in the Bahamas, far from outside influences, spending company money freely. SBF became rich and prominent virtually overnight, on stages with Bill Clinton and in ads with Tom Brady. How could it not have felt unreal?

Also, they made tokens—FTT, SRM—that they could trade for real money. Did they believe that those tokens were real? I mean, they ran a crypto exchange. What does “real” mean, really?

You start with a company that builds a box and in practice this box, they probably dress it up to look like a life-changing, you know, world-altering protocol that’s gonna replace all the big banks in 38 days or whatever. Maybe for now actually ignore what it does or pretend it does literally nothing. It’s just a box. So what this protocol is, it’s called “Protocol X,” it’s a box, and you take a token. …

So you’ve got this box and it’s kind of dumb, but like what’s the endgame, right? This box is worth zero obviously. … But on the other hand, if everyone kind of now thinks that this box token is worth about a billion-dollar market cap, that’s what people are pricing it at and sort of has that market cap. Everyone’s gonna mark to market. In fact, you can even finance this, right? You put X token in a borrow/lending protocol and borrow dollars with it. If you think it’s worth like less than two-thirds of that, you could even just like put some in there, take the dollars out. Never, you know, give the dollars back.

“You’re just like, ‘Well, I’m in the Ponzi business and it’s pretty good,’” I said. That’s what FTX ended up doing! It printed its own tokens, pretended they were valuable, and ended up trading its customers’ real money for those tokens: “Never, you know, give the dollars back.” They were playing a game, a game where everything felt fake and arbitrary and driven by sentiment and confidence. The sentiment about them was good, their confidence was high, they had a world-altering crypto exchange that was gonna replace all the big banks in 38 days.

Sure, they were taking customer money and trading it for magic beans. Sure, when you put it like that, it sounds like fraud. But when you put it like that, it also just sounds like running a crypto exchange?


A central fact about crypto is that it is just a thing someone made up. There is a white paper. Someone—“Satoshi Nakamoto”—wrote it. A lot of people took up the cause, decided Bitcoin was valuable, traded cryptocurrencies for real money, built new financial systems and technologies on the basis of crypto. It is a collective social project.

This is not a bad thing. Everything, really, is a social project. A company is, or a country, or the stock market, or the whole global financial system: all people, doing stuff, linked by some forms of coordination and incentives, and by a shared belief in what they are doing.

But a company probably makes stuff; a country has an army and a police force; the global financial system provides your mortgage and your checking account. You do not participate in the global financial system because you believe in Jamie Dimon’s vision of the future; you just need money to buy a house.

Crypto is, now, less necessary. Nobody has to have anything to do with crypto: You don’t need it to pay your mortgage or buy groceries, and if you want to ignore it, you just can. [4] Crypto is like a—large, distributed, decentralized—startup; as a social project it has big plans to change the world, and some cool prototypes, and a lot of hype, but it has not yet made itself essential to most people’s lives. If you go to work at the startup, it’s because you think the stock options will pay out.

Buying crypto is a choice. The main reason to do it is that, at some level, you believe in the social project, in some aspect of crypto’s vision for the future, a distributed web3 or a censorship-resistant financial system or whatever. Or you think cryptocurrency prices will go up—because other people will buy into the vision—so you want to make money.[5] You are betting on the social project.

This is not, in the abstract, a crazy thing to bet on. Social projects can create lots of value, and there are things about the crypto project—its technical ideas, its widespread and rapid adoption by technologists and financial people, the enthusiasm of its supporters, the price of Bitcoin—that suggest it might be promising.

But, paradoxically, crypto is much more reliant on trust than the rest of finance and business. It only works if people believe in it. There is no external source of value.[6] Crypto prices go up when more people become more interested in crypto; they go down when people turn away from crypto.

One thing that this means is that, if you are running a scam, you will be drawn to crypto. You are running a confidence game, and crypto offers the most efficient market for turning confidence into money. “These people just made these tokens up and sold them for money,” you will say to yourself. “How do I get in on that?” There are more sophisticated versions. “You make a box and issue a token and get some trading action; everyone marks to market, and then you can borrow against it and never give the dollars back” would be one.

But another thing this means is that, if you are in crypto, and you are not running a scam, you rely on the trustworthiness of everyone else in crypto. If the highly trusted operator of a big crypto exchange—the public face of trustworthiness in crypto!—turns out to be running a big fraud, you can say, “Well, I wasn’t running a big fraud” or “That guy’s big fraud says nothing about the underlying blockchain technology,” and, fine. But, like it or not, you are in a collective social project, and that project is crypto as a whole, and you will be judged by the other people in that project with you. If everyone thinks “Ah, yes, crypto, that’s for scams and crime,” that is bad for the project.

It is not, by any means, the end of the project. One possible future for crypto is to return to Satoshi Nakamoto’s vision of trustlessness and decentralization. Decentralized finance, DeFi, has come out of crypto winter looking relatively good; in fact, open-source smart contracts on the blockchain do seem less likely to steal customer money than centralized exchanges. The centralized exchanges do keep attracting money, though; people want somebody to trust.

Another possible future is for crypto to do better at earning users’ trust, or more realistically for regulators to force it to. You get some audited financial statements, some leverage limits, some regulators checking up on the risk management, and maybe crypto exchanges stop blowing up so often. One problem with this is that FTX was a leading advocate for it, and look what happened. It’s not a great look for regulators and politicians and auditors to work with crypto firms right now.

If the crypto project is going to work, that’s probably its best chance: to be more regulated, more grown-up, more like the regular financial system. All that stuff in the regular financial system, all the accretions of rules and customs and requirements, all the intermediaries and checking—it turns out all that was there for a reason. It can be fun to get away from it for a bit, to build a fantasy financial system without all those boring rules, but that’s just a game. In the long run, you want your system to work in real life.

But that is not an inevitable outcome. Crypto might want its system to work in real life, but why should anyone else? After how the crypto financial system performed this year? Crypto’s toy financial system managed to have itself a toy financial crisis: The collapse of crypto lending firms and exchanges in 2022 was in many ways worse, faster and dumber and more complete, than the global financial crisis of 2008. But it did much less harm, because the damage was confined mostly to crypto. Crypto speculators, people playing in the toy financial system, lost a lot of crypto. But banks and savers mostly did not lose money, because banks and savers mostly steered clear of crypto, because it was so obviously unregulated and full of scams. More and better regulation would be good for crypto, in that it might give more regular people the confidence to invest in crypto. But that might be bad for the regular people!

Over the last few years crypto built a toy financial system. That was an accomplishment, both in a technical sense (crypto found smart ways to do financial trading) and a social one (crypto attracted a lot of smart finance people). It is an accomplishment that I personally appreciate, since I love a clever financial system. But it is in important ways a bad place to start. A cleverly designed exchange for trading magic beans will never get around the basic problem that the magic beans don’t work, and people might stop believing in them. If crypto is going to work in the long run, it will need to prove its real usefulness outside of finance. Finding new ways to trade the tokens is fun, but it is not enough; the tokens have to mean something, too.

More On Bloomberg