If he was actually committed to effective altruism, he should be trying to send himself to jail as an example to others to not do fraud.

A utilitarian effective altruist absolutely should defraud anyone not already on their team.

They precommit to others who are in on the take that 90% of the profit will go to the Cause, and then all that excess that was gonna be wasted on luxury or low-QALY retirement/dying years will instead go to places that generate more utils.

I wouldn't be shocked if a document (or, more likely, testimony in a trial) eventually turns up explicitly saying that.

(To be clear, I think the plan I have described is reprehensible, but that's because I'm not a utilitarian.)

And what might we expect the second order effects of “EAs are philosophically committed to fraud” be? Like this only looks rational if you kind of take a first approximation of a dynamical system and call it good.

“Worse than a crime, it’s a mistake”

It only looks rational if you think you can get away with it. People considering it at all are at least somewhat likely to think they can.

Similarly, inside the utilitarian mode of reasoning, it's only a mistake if you get caught.

SBF consistently demonstrates some serious next level thinking:

- Next level risk management: don't worry about it

- Next level defense strategy: talk as much as possible

- Next level EA: save the planet by killing the crypto space which is wasting energy