Live data from Hacker News

Codeberg Divides

lucumr.pocoo.org

241–242 of 242 posts

Re: Codeberg Divides

#241

Earlier quoted context omitted.

You're ignoring that adding non-copyrightable code to a collection of copyrightable code makes the collection copyrightable. The folks that wrote the policy are also ignoring that. It's simply a legal misunderstanding, because (I suspect) they didn't actually consult a lawyer, and their reading of precedent is overly broad.

> You're ignoring that adding non-copyrightable code to a collection of copyrightable code makes the collection copyrightable. That is true for public domain code. It is not true (at this time) for machine generated code.

Any work that is not copyrighted is public domain.

Anyway, that case did not establish that code written with an agent cannot be copyrighted by the human operating the agent. The whole problem is that the application was submitted listing the AI as sole author, so the court did not consider whether the operator could have listed himself as the author. And if he had no one would have had standing to dispute it.

Re: Codeberg Divides

#242

Earlier quoted context omitted.

So they have modified their policy because they do not understand either copyright law in general or Thaler vs. Perlmutter in particular. I think we will be seeing a lot more of this. There were plenty of examples of developers failing to understand copyright law and the implications of licenses before LLMs and that was with well established law. With some genuine uncertainty around LLMs, more variation between juris…

What aspect of copyright law do we not understand?

Public domain works incorporated into a copyrighted work result in a copyrighted work. And that case didn't determine Thaler could not have claimed authorship, only that because he didn't, his application was correctly denied.
Post reply on HN