The 'Don't Look Up' Thinking That Could Doom Us with AI

It's bold to assume that we will create anything that can care. Even a paperclip maximizer. It will just get taken offline as not fit for purpose, these things are expensive to keep running. If they aren't generating positive value and doing as instructed then there's no point keeping it online.

I doubt we will ever create an "intelligent mind". If we build something that is capable of caring about anything rather than accurately following instructions and not "hallucinating" then we are doing AI wrong and we have created a slave rather than a machine.

taken offline? That's laughable.

The argument for keeping it in a box is not even made any longer. Developers are working as fast as they can to proliferate it across the internet because it's already so useful. What do you think the outcry will be like when it is 10x better and you tell all the internet users you want to shut down there new super-helpful friend?

https://archive.is/gMi8q

LLMs are not intelligent.