Terence Eden on Nostr: One of the things I've noticed with #ChatGPT, is that it is reluctant∗ to say ...

One of the things I've noticed with #ChatGPT, is that it is reluctant∗ to say "that's not possible".

If you give it various computing / linguistics problems, it would rather say the wrong thing than admit defeat.

I wonder how much hallucinating could be removed from #AI if it were allowed to show ignorance?

∗ I know, don't anthropomorphise computers; they don't like it.