Tyler Cowen: evaluation

Transcript context

…I don’t think there’s a negative correlation. I think some of the big names have been really good conversations as well, but we’re not in the Dwarkesh Patel territory of he will say that his most popular episodes tend to be the more obscure. For whatever reason, at least this year, that’s not our experience. Okay. Let’s move on to listener questions. First one from Patrick McKenzie, past guest on the show, “What are the biggest surprises in AI this year and the biggest changes to the Tyler Cowen et al production function as a result of AI this year?” Maybe the biggest surprise in AI was how good the O3 model was, which I called AGI at the time. Maybe tongue-in-cheek, but I also meant it, that it can’t do every job. The old definition of AGI, that it’s just as good as human experts, is largely true. I don’t know if it was a shock. By that point, you were more or less expecting it. Had it come one year later, that would not have been a surprise either. Just how much better reasoning models got from, say, O1, to me, was somewhat of a surprise. Otherwise, mostly, the lack of surprises was the surprise. Things just kept on going. There’s been great progress. They used to think there were two R’s in strawberry, and now they know there’s three. In their spare time, they win gold medals at math Olympiads. To broaden the question to the production of the show itself, I don’t think AI has been transformational this year. I assume that’s what Patrick was referring to when he asked about “et al.”…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.