To your edit: A compiler will make the exact same output from the same input. An LLM will make different outputs from the same prompt and may give you someone else’s code verbatim
A compiler doesn’t have to make the exact same output, nor does an LLM have to make differing outputs, and a compiler uses someone else’s code to transform your code into something else.
I believe it was a product of the earlier conflict between copyright owners and AIs on the training side. The compromise was that they could train on copyright data but lose any copyright protections on the output of the AI.
https://reuse.software/faq/#uncopyrightable
The REUSE specification recommends claiming copyright even if it’s machine generated. Is this incorrect information?
EDIT: Also, how is copyrighting code from an AI different than copyrighting an output from a compiler?
To your edit: A compiler will make the exact same output from the same input. An LLM will make different outputs from the same prompt and may give you someone else’s code verbatim
A compiler doesn’t have to make the exact same output, nor does an LLM have to make differing outputs, and a compiler uses someone else’s code to transform your code into something else.
I believe it was a product of the earlier conflict between copyright owners and AIs on the training side. The compromise was that they could train on copyright data but lose any copyright protections on the output of the AI.