Considering Ilya regrets his decision, implying this wasn't an e/acc vs safety debate, the two most likely scenarios to me appear to be:

A) DeAngelo getting revenge for custom GPTs killing his company Poe's core product.

B) The board discovering Sam working to use OpenAI resources/IP to further his own personal business/political goals.

> Considering Ilya regrets his decision, implying this wasn't an e/acc vs safety debate

Does it?

If it was exactly that but he misjudged the external response and resulting consequences, that would also be a reason for regret.

That's very well possible but there is as of now no evidence to substantiate that. But that is what it ultimately could have boiled down to and if there is even a remote chance that it can be spun that way you can be sure that chance will be exploited - even if it wasn't originally the case!

> These are not idiots.

They shouldn't be. But short of temporary insanity I'm running out of ideas on what it could be.

Yeah there has been plenty of explanation. Sam Altman has competing interests and was looking to leverage OpenAI’s tech in those interests.

Source??

My guess, and it is just a guess, is that there's some deep and fundamental position that the board members hold to, which Sam cannot describe well enough to pass the Ideological Turing Test[0], and this failure was perceived as being deliberately dishonest.

Happens a lot. I know some on the very extreme outskirts of the left wing who look at the "splitters!" scene in Life of Brian and see themselves; I've seen it in arguments over programming language features; there's plenty of religious disputes that look like this from the outside; most of the Brexit arguments painted everyone who didn't agree with the speaker as being part of Team Other and merely an attempt to undermine Team Us; and so on.

People can get very seriously angry about things that seem totally pointless from the outside.

[0] https://www.lesswrong.com/tag/ideological-turing-tests