Stefan Eissing1d ago

Lots of posts about AI. It‘s important that people exchange their views.

My own:
- the AI business plan is to dominate humanity, no exaggeration.
- LLMs can be a serious technology, code reviews have become good.
- AI is costly. The human brain is more efficient with superior capabilities.
- Idiots with LLMs suck and remain idiots. Probably forever.
- Bad code is bad code, who/whatever made it. Banning LLM code policies miss the point. Deny code you don‘t like, subjectively. Remain in charge.

Show threadAndreas Ahlenstorf
@icing How do you figure out whether a contribution is copyrightable or doesn't violate someone else's rights? Because that's what I ultimately have to do when reviewing LLM-generated code. I don't want to end up in an ReactOS-like situation.
Apr 3 at 9:05amWeb
Show threadStefan Eissing17h ago
@aha That applies to every code contribution. You don‘t know where humans get theirs from either.
Show threadAndreas Ahlenstorf16h ago
@icing But the rules are different. German copyright law is for humans, not machines. So you need to determine whether the human was involved enough in the creation of the work for it to be copyrightable. How do you do that? And how much is enough?