[[[ To any NSA and FBI agents reading my email: please consider ]]] [[[ whether defending the US Constitution against all enemies, ]]] [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
If specific code generated by an LLM can be understood by humans, there is no inherent reason to why we could not use it. However, this is a separate question from the effect of the LLM use on the copyright on the code. There are two possible kinds of issues. 1. That code might infringe one or more of the works that were used to train the LLM. 2. That code might be unable to carry a copyright of its own. I think it is possible for both of these to occur for a single piece of LLM output. At present we do not know how we should address these two potential problems. Therefore, for now, we do not want any LLM-generated code in Emacs. -- Dr Richard Stallman (https://stallman.org) Chief GNUisance of the GNU Project (https://gnu.org) Founder, Free Software Foundation (https://fsf.org) Internet Hall-of-Famer (https://internethalloffame.org)
