Leaving aside arguments about whether or not the LLMs are actually AI and how (in)accurate they are, he clearly fears what bad actors could do with AI. Which begs the question of why he was involved in freaking developing one in the first place!
I am left feeling more annoyed at his preaching, both about technophiles and technophobes, than feeling like I learned from the conversation. Also, it probably could have been half the length.