zlacker

Based on the report, it seems like he was kicked for focusing too much on profits instead of developing and improving AI. This is purely speculation, but I've always suspected that the guardrails they put on ChatGPT to make it "safe" (i.e. corporate-friendly) essentially acts as a lobotomy for the AI. Hopefully we can start seeing a less censored ChatGPT and see if it really does perform better.

>>matteo+(OP)
It does, but Ilya had to have been one of the votes against Sam and he's spoken about AI safety quite recently too.

If this was about differing visions on the core product, it may have instead related to the open/closed aspect of progressing.

Sam may have been the driving force behind keeping things closed in the name of safety, and others at OpenAI might have been ready to rebel if it continued that way in the future and prevented general advancement in the field.

Scientists don't like closed research.

The safety alignment part is unlikely to be the core issue even if there are underlying issues with it.