If you had a language model whose responses were indistinguishable from a human, would you call that understanding?

I agree GPT-3 is * just * a language model, but then again, so are you.

Not correct. I can write a computer program (logic) and gpt-3 can't.

GPT3 can write correct computer programs. There are many demos online.

Stepping back it is just a translation problem e.g. English idea -> Python code, which an ideal language model would be able to do.

I think what I’m picking a gripe over is that “language model” is a very general term. Calling GPT3 just a language model is like calling a human just a bunch of atoms. It’s not actually limiting.

What I’d love to talk about is how useful of a language model is GPT3? And how useful do you think GPT4, 5, etc. are going to be?

No, GPT-3 is a lang model. I can call it like than since there is a logical definition of a lang model. We do not yet have model of human or human brain definition