> You’d think this MIT graduate would wise up, but he never did.
I think there were potentially two problems. First, for the previous 10 years, lying and taking massive risks actually did work for him: Even if some people became disillusioned and stopped buying what he was selling, he could just walk away from them and start working with people who would.
Secondly -- well, what else could he have done? It's a bit like back when we saw those AIs playing Go or Starcraft or whatever at the point where it was clear they were going to lose: They'd just start making random moves, because terrible moves had as much chance of winning as "minimize loss" moves.
There's an interesting article I read which is half book review, half critique of SBF and "Effective Altruism" (EA), that has some interesting connections[1]. Basically, EA encourages "cost / benefits utilitarian analysis" thinking; every time SBF lied (or committed fraud), he could tell himself that it was actually moral because it maximized some other good towards thinking people somehow. It also encourages poker-type gambling in some circumstances, where it makes sense to take a bet where you lose 75% of the time, as long as the potential payout is larger than 4x.
And if that article's reading is right, FTX actually had a profitable business. All they had to do was sit back and take x% of trades and they would have been fine. But that wouldn't have been a huge massive multiple that SBF was looking for, so he went in for a bunch of poker-style bets with his customer's money -- ultimately in the name of making more money to give that money away. (And he did actually give a load of money away, even when it didn't make financial sense to do so.)
SBF's choices were:
1. Accept that you're going to jail for a long time no matter what, and try to minimize the damage. Maybe you can cut your jail time in half.
2. See if you can weasel your way out of it. Maybe your jail time is a bit worse, but maybe you don't get any at all.
First of all, he'd always been able to do #2 so far, so why would this time be any different?
Secondly, #1 means a huge emotional downside for current SBF; #2 only means a potential downside for future SBF. The attitude he got into, where he screwed everyone else in favor of "current SBF", carries over into screwing "future SBF". So there's a sort of poetic justice to it.
[1] https://www.lesswrong.com/posts/AocXh6gJ9tJC2WyCL/book-revie...