I can't quite call this a good book. Michael Lewis isn't the person to write the story of Sam Bankman-Fried. He had the luck—you decide if it's good or bad—to be in the room when FTX collapsed. Although Lewis has financial chops, he's primarily a teller of stories about people, and someday a better book, written by someone with investigative chops, will be written.
Instead, what we get is a portrait of Sam Bankman-Fried (hereafter known as SBF). Lewis plays it straight, never fully giving his opinion of SBF—you have to guess. I would not call the portrait flattering, though some have called this a hagiography. Whether that was Lewis' intent, or because his portrait unintentionally played into all of my preexisting prejudices, is an open question. What I will say is that if the last chapter or two of the book had played out differently—if the FTX house of cards hadn't fallen apart—people would be hailing SBF as an eccentric genius. (I wouldn't, but more on that later.) I will say that the book reads a bit like Lewis started writing it liking SBF, and then the "oh shit" came in during the editing process. Based on interviews I've read, he was entirely too credulous about both SBF and crypto in general.
The short version of the book is that SBF is a major fucking asshole who has no regard for anyone but himself. His childhood vignettes could be viewed as cute and quirky. I didn't see them that way. Lewis almost sets him up as a "look maybe he's autistic!" in the way that's always applied to white men who have bad social skills. (I would question if he has some other pathology going on, but I don't diagnose.) SBF is so much more than that, though. He's the coddled son of Stanford professors whose self-regard is barely punctured till he's not the smartest kid at math camp. He decides literary analysis is bullshit because he doesn't get it.
SBF is the perfect synopsis of the paradigm that is the Rational Man: He claims to be deciding everything via data and logic, while ignoring that he's actually manipulating logic to produce the result he already wants, or that ethical claims cannot be entirely decided via statistics. Everything, to SBF, is just a puzzle that needs to be solved. Nothing is real, and there are no human stakes to consider.
The apparent exception to this rule is his belief in Effective Altruism, which holds that people should maximize the utility of their giving. That sounds great, but it breaks down quickly. For example, using EA logic, one should try to earn the most money possible, to give away: the perfect conscience-salve for working in finance. In reality, people don't necessarily have such a choice, and it only works as long as a minority of people do it. But the real hole at the heart of EA comes later, when it's all about all the money they will give, but never actually do... and when they use their logic to persuade themselves that the best use of their money is long-termism, or investing in highly improbable scenarios like preventing an asteroid from hitting the earth. It's a very convoluted way of avoiding having to help people today: having to take moral responsibility for choices.
Incidentally, the finance industry and academia don't come off very well: the best goal for MIT physics majors is apparently to go work for firms like Jane Street so they can play games with other people's money.
In the end, as much as I hated SBF at the end of the book, I also felt Lewis hadn't actually given me anything solid. Why did SBF do anything he did? Was it all just a game? An employee accurately sums him up as entirely lacking in empathy, and that's probably correct. Is there more there? I don't know.