Accelerating research on increasing reasoning / intelligence capabilities of AI? That is opposite of what Emmett seemed to advocate recently, so it would be strange to pick him as the CEO if they really wanted to accelerate that.

Those are not the only things to research, but I wouldn't assume that capability-base research would stop. I mean, they want to create AI to perform AI alignment, so their road map includes capability research.

The for-profit bender that OpenAI was on appears to have been more of the issue. It's one thing to create greater capabilities for enhancing the rate of research, and another rushing those capabilities to market.