ChatGPT could never get a PhD in geography

Re:ChatGPT could never get a PhD in geography by ben_w on Monday May 12, 2025 @05:29PM (#43967636)

I very much appreciate all the ways we're improving our ideas of what "smart" means.

I wouldn't call LLMs "smart" either, but with a different definition than the one you use here: to me, at the moment, "smart" means being able to learn efficiently, with few examples needed to master a new challenge.

This may not be sufficient, but it does avoid any circular arguments about if any given model would have any "understanding" at all.

[ Parent | on HN ]

Re:ChatGPT could never get a PhD in geography by yoko888 on Monday May 12, 2025 @09:42PM (#43969012)

I appreciate your point — I also don’t think LLMs are “smart” in the usual way.

But I’ve been wondering: maybe we don’t need to ask whether they “understand” like we do. Maybe it’s more useful to ask: can they build useful patterns from limited experience, and improve over time?

It’s a different kind of intelligence — but maybe not a lesser one. Just one that doesn’t look like ours.

[ Parent | on HN ]