Hi Iván, Iván Rodriguez <[email protected]> writes:
> I was working with Grok 4.6 (I don't remember the exact model and I'm not > sure if I need to include the assistant tool in the email, etc. I would > appreciate it if you could inform me about this) mostly for > stress/pathological tests and as a secondary reviewer of the patch in the > source. I don't think we have an official policy on anything. I suspect we are, for the most part, insulated from the LLM barrage since we still primarily use mailing lists instead of GitHub. Of course, there are exceptions; for example, the Linux Kernel is definitely not insulated from it, despite using mailing lists. With that said, I'd be for following Gnulib's policy since I am still concerned about the copyrightability of LLM-generated outputs [1]. The policy doesn't try to prevent you from using LLMs, but it prevents you from using them to write anything more than a few lines of code for you. E.g., for Anthropic's OSS scanner, I have been pretty much ignoring the patches and skimming the start of the summary and looking at the command-line invocations used to trigger the bugs. That helps avoid the posibility of my patch being uncopyrightable, or copyrighable by someone else. Also, I find LLM-generated tests to be a bit wonky; e.g., they will assume Linux behavior is portable. Perhaps they are fine for simple stuff, but I'd be a bit skeptical of them. Collin [1] https://github.com/coreutils/gnulib/blob/8f5547126aedc70a9c2cdb7c63829c87c3d71df1/HACKING#L240-L276
