Comment by devjab

I think programmers are in the perfect profession to call LLMs out for just how bad they are. They are fancy auto-complete and I love them in my daily usage, but a big part of that is because I can tell when they are ridiculously wrong. Which is so often you really have to question how useful they would be for anything where they aren’t just fancy auto-complete.

Which isn’t AIs fault. I’m sure they can be great in cancer detection, unless they replace what we’re already doing because they are cheaper than doctors. In combination with an expert AI is great, but that’s not what’s happening is it?