this. The whole licensing of software stands on copyright. If the content is already in the public domain because it was generated by an LLM (copyright only protects human works) then it can't be licensed.

However, there's no compunction to publish the generated code, even if it's public domain.

We end up in a strange nega-OSS world where all code can be used by anyone for any reason, if you can get your hands on it.

If you believe this, then this was always the case. LLM's only made it cheaper and more accesible.

Sorry I don't understand, what am I believing that has always been the case?

That you could reimplement a software and then be license free. Or did I read you wrong?

That's a separate thing. I mean, valid, and a good point to raise, but I wasn't raising it.

I was more talking about the weird space we're getting into where code is completely open, because it was generated and cannot have any copyright protection, but also unpublished, and so effectively proprietary. The opposite of OSS, where code is available but protected by copyright-based licenses.

That's true. You can write software that duplicates the functionality of existing software and by doing that you can evade the copyright of that software.