What is the policy/attitude towards LLM contributions in CCL?

[zyd-n]

@xrme @sellout @billstclair @gmpalter @mikelevins @gzacharias

Hi, I'm looking over all the active Lisp, Scheme and Lisp-adjacent languages and implementations to see how projects contributors and maintainers are responding to the proliferation of LLMs. I don't see anything that resembles an official 'policy' or guideline to LLM-generated/assisted/enhanced/based/etc contributions.

Are LLM contributions welcomed in CCL by the maintainers and contributors?

[gmpalter]

There's an individual using Claude to port CCL to Arm64. See #11.

[mdbergmann]

Yeah, this question comes up now everywhere. Would be good to know.

[xrme]

Ultimately, contributions need to come from humans who know what they are doing.

It's fine if the human uses tools such as LLMs, but I wouldn't like to see PRs like "here's a patch my LLM generated; it seems to work for my issue, but I don't understand why."

If I had to adopt a formal policy today, I'd probably copy Ghostty's. https://github.com/ghostty-org/ghostty/blob/main/AI_POLICY.md