Relevant excerpt:
[Ben Buchanan]: A canonical definition of A.G.I. is a system capable of doing almost any cognitive task a human can do. I don’t know that we’ll quite see that in the next four years or so, but I do think we’ll see something like that, where the breadth of the system is remarkable but also its depth, its capacity to, in some cases, exceed human capabilities, regardless of the cognitive discipline —
[Ezra Klein]: Systems that can replace human beings in cognitively demanding jobs.
[Steve]: Yes, or key parts of cognitive jobs. Yes.
The interview was really interesting, and it's very revealing to see what's in the head of somebody advising on the future policy. That said, most policy-making actors have some kind of stake in the results. That should probably be kept in mind when looking at the narrative that's coming out.
