2022 was, you might say, the year when grifters got their just deserts. The year began with the conviction of Theranos founder Elizabeth Holmes, and ended with the arrest on fraud charges of FTX founder Sam Bankman-Fried, and the revelation that newly-elected Republican House member George Santos had essentially invented an entire life for himself when talking to voters. So perhaps it would be more accurate to call it the year of grifters getting caught (though Santos is still on track to be seated as a House member in a couple of days).

Of all these stories, the biggest, and strangest, was of course the collapse of FTX, the crypto “exchange” Bankman-Fried had somehow turned into a business that was valued at $32 billion at its peak. In one sense, FTX looks like a simple financial scam — according to the SEC complaint against Bankman-Fried, Alameda Research (the crypto hedge fund that Bankman-Fried owned and controlled, and which was run by his former girlfriend, Caroline Ellison) was dipping into FTX customer assets to fund its own trading, and was also funneling money to SBF and other FTX executives in the form of loans.

But that simple summary doesn’t really explain what happened at FTX, or why it collapsed. In a conventional scam — like, say, Bernie Madoff’s multi-billion-dollar Ponzi scheme — the scammer fraudulently uses new customer funds to line his own pockets and become insanely rich. But while there seems to have been some of that in SBF’s case — spending millions on Bahamian properties and the like — it appears that the vast majority of the money that was taken from customers went to making huge, reckless bets on various cryptocurrencies and crypto companies. That is, on the face of it, an odd thing to do with stolen money.

The FT, for instance, recently published a spreadsheet of the investments Alameda Investments (FTX’s venture arm) had made in what seem to be mostly crypto companies, almost none of which you’ve heard of. There were almost 500 of those investments, with a supposed value (as of early November) of $5.4 billion. That valuation was likely massively overstated. But even so, that’s a ton of money for SBF to be putting into illiquid investments, many or most of which were sure to blow up, while also using FTX customer money to fund Alameda’s trading.

Similarly, consider FTX’s relationship with Alameda. The reason FTX collapsed was that on top of whatever Alameda was doing with FTX’s customers’ money, it was funding its leveraged bets with billions of dollars in loans from FTX, loans that were largely collateralized, absurdly, with big chunks of FTX’s own cryptocurrency, FTT. When crypto prices crashed, Alameda’s trading losses became too big to cover (or to hide). And when, around the same time, the value of FTT collapsed after FTX’s chief competitor, Binance, dumped a big chunk of it, the “collateral” on those loans became largely worthless, making FTX effectively insolvent.

What FTX seems to have been, in other words, was not so much a conventional scam, but rather a vehicle for SBF to take absurd risks, often financed by customer money. And I think what happened at FTX was, in that sense, ultimately the result of Bankman-Fried’s eccentric ideas about risk, which he explained in many different interviews, back when people thought he was a tech guru.

Bankman-Fried had two core ideas when it came to taking chances, which he laid out pretty clearly in an interview with the podcast 80000 Hours in April. The first was that your goal shouldn’t be to make a lot of money — it should be to make a ton of money, because that’s what you need if you want to “maximize the amount of impact that you have on the world.” The second was that people are far too risk averse in general, and that instead of trying to limit the downside of the risks you take (including most obviously investments), you should be trying to maximize the upside, meaning you should make lots of bets that have a high likelihood of failing, but that will pay off huge if they succeed. As he put it in that interview, “The ‘optimal strategy’ to follow is one that probably fails — but if it doesn’t fail, it’s great.”

It’s worth not skipping over the weirdness of this position, which implies that you should be essentially indifferent between making a lot of money and losing all your money on a bet that had a positive expected value. It’s not that the strategy is illogical — it’s that it seems, in some sense, inhuman. And while it’s obviously a lot easier to follow this kind of strategy if you’re doing it with other people’s money, even then, making huge bets that you think will “probably” fail is an unusual thing to do.

That is, though, exactly what SBF was using FTX to do. To be clear, he seems to have massively underestimated the risks involved in betting on crypto — I don’t think there’s any reason to believe that the bets Alameda was making on a host of fake-money assets actually did have a positive expected value. But SBF presumably thought they did, and believed that making all these bets was the rational thing to do, even if they were likely to fail, and even if Alameda was using customer money to make them.

None of this excuses what SBF did — if you tell customers that their assets are theirs, and then use them to fund trading (or, even worse, personal loans), that’s fraud. But it does help explain why he didn’t just use FTX as a money pump, which he presumably could have done for years with just a minimal level of risk management.

Bankman-Fried has, so far, denied he’s denied knowing anything about Alameda dipping into customer funds. But that always seemed improbable. And looking at what happened at FTX through the prism of his ideas about risk makes you wonder whether he viewed the possibility of FTX collapsing and him being arrested for fraud as just another risk he was willing to take in pursuit of this huge upside he was chasing.