General Discussion

in the OP. I've also seen articles and studies on people using AI to code being proven NOT to be saving as much time as they think they are, and in some cases spending more time on it than they would doing the coding without AI.

And I wince every time I read about someone using AI to code in a language they know little about, because I'm not sure they can catch any problems with the code.

Which is why vibe coding is particularly a mess.

I've read that people using AI to help them with coding in a language they should be more familiar with don't learn/remember that language in the way they would if they didn't use AI for it.

And I've seen lots of posts from security firms on AI code creating security risks.

I don't doubt that some people might find it useful for some smaller coding projects. But I don't think that begins to outweigh all the harm genAI does.

And AI coding tools, like other genAI tools, are built on a lot of stolen intellectual property. Even when code was posted online with a Creative Commons license, those licenses have certain requirements the genAI industry ignores. They're fundamentally unethical tools. Those AI coding tools' training data is almost certain to contain a lot of stolen books, since those are the basis of the chatbot's ability to communicate.