zlacker

No he didn't fire Sam over AI safety concerns. That's completely made up by people in the twittersphere. The only thing we know is that the board said the reason was that he lied to the board. The guardian[1] reported that he was working on a new startup[1] and that staff had been told it was due to a breakdown in communication and not to do with anything regarding safety, security, malfeasance or a bunch of other things.

[1] https://www.theguardian.com/technology/2023/nov/18/earthquak...

>>seanhu+(OP)
The Atlantic Article makes it pretty clear that the fast growth of the commercial business was giving Ilya too few resources and too little time to do the safety work he wanted to do: https://archive.ph/UjqmQ

>>seanhu+(OP)
Spoken to a bunch of folk at OpenAI, it really does seem to be regarding safety. Ilya was extremely worried, did not like the idea of GPT’s as users can train AI’s to do arbitrarily harmful stuff.

>>frabcu+p2
I can buy that they fired him over a disagreement about strategy (ie are we going too fast/we are concentrating on the wrong things etc), because in general of course board members get fired if they can't work together on a common strategy. But the narrative that has taken over lots of places at the weekend is more along the lines of he got fired because they had created a sentient AI and Ilya was worried about it. That just makes no sense to me.

Additionally, no-one (not insiders at OpenAI and certainly not a journalist) other than people in those conversations actually knows what happened, and noone other than Ilya actually knows why he did what he did. Everyone else is relying on rumor and heresay. For sure the closer people are to the matter the more insight they are likely to have, but noone who wasn't in the room actually knows.