He also said he was even more bullish on FTX and SBF after the podcast that quote came from

There's that famous quote about the tide going out and seeing who isn't actually wearing swim suits. I think a similar thing applies to all the financial chicanery that has been in play in this 0% interest rate environment. In reality, while the financial system is irreducibly complicated in some important ways, there's also an important sense in which the entire thing is really built on just some simple primitives. Basically, the financial system can move risk spatially and temporally, and charge a fee for the activity. That's really all it can do. I think from the proper point of view it can't really even do that primitive without at least some net increase in risk, though I admit defending that mathematically would be a challenge. (For now we could just start with the proposition that if there was an easy way to simply reduce overall risk, it would have already been taken, so all in all there can't be very many such things just lying around. And they've been building a lot of things on this primitive, so it's not a hard guess that there's a lot more use of the primitive than there could possibly be opportunities for such uses to result in true net reduction in risk, whatever exactly that may mean... I know I'm handwaving here.)

If you think about it mathematically, it's important to understand that no amount of concatenation of the primitive of "moving around risk" can do anything more than that. And, in particular, no amount of concatenation of things that leave net risk either the same or somewhat greater can ever end up reducing the amount of risk in the system.

However, each such composed operation introduces a place for someone to misprice the risk, especially as the risk is communicated across lossy channels, let alone channels with a certain amount of incentive to misrepresent the risk to the buyers on the other end. So while no amount of combination of those primitives can ever reduce risk overall, it sure is full of opportunities to convince people the risk has been reduced, and for them to take various actions based on that. Combine that with one of the most popular operations being to take a nice risk gradient that gradually ramps up from "high probability of small bad event" to "low probability of big bad event" and shoving all the risk into "super low probability of total utter unrecoverable catastrophe", and the whole thing is just destined to explode hopelessly.

And yet... perhaps some readers are saying well, duh, jerf, how else could it be, I would say to you I'm firmly middle aged now, but those castles based on financial engineering have been floating in the sky my entire adult life now, even counting the so-called "crises" I've seen. And to the naked eye, those castles have gotten bigger, higher, nicer, and more numerous the entire time. It takes... something... some pretty big cajones to stare up at those things that have been floating in the sky for so long and saying "they can't possibly do that forever". Especially when you may well have bankrupted yourself seven times over trying to trade on that presumption, even if it is in fact true in some abstract sense.

The financial system is rapidly simplifying. It's going to be a painful process for quite a lot of people. A lot of people who think they are on solid ground are going to discover they've been herded onto the flying castles without realizing it. It will, if nothing else, be very educational I suppose.

Yes, I read the article where Matt Levine explicitly mentions that he missed this.

There's nothing wrong with being an entertainer, it's just something to keep in mind.

In other words, don't make major financial decisions based on Levine not saying it is a good or bad idea.

he's a former M&A advisor at Goldman Sachs, so he has a lot of background and knowledge in everything he writes about.