On 2026-05-26 at 11:01+02:00, Hugo Buddelmeijer wrote:
> On 5/26/26 03:19, pelzflorian (Florian Pelz) wrote:
> > Don't talk about use of LLMs.  Writings by LLM are a problem,
> > not use.  Hints given by consulting LLM on packages code cannot be,
> > since according to c't magazine, all the Web search engines
> > internally used language models for query interpretation
> > since a long time ago.
>
> Half of the argumentation is that using genAI is bad
> for the environment and is enabling fascism.  It does not matter
> that the genAI is hidden behind a search engine
> for those arguments to matter.

I wholely agree with the sentiment here.  It should not be
within the scope of Guix contribution guidelines to dictate
how a contribution is done, but only what contributions it accepts.

It is indeed important to have a policy in place though,
because it might not be obvious to all contributors that piping LLM
into one's editor may result in low-quality contributions
or copyright infringements, and that Guix is not open to such risks.
The same for reviews.

It might be a cutural clash but I also find the GCD's title and framing
distasteful and the opposite of its disclaimer:

> This pledge is *not* about claiming a "moral superiority"
> of the project and its members for not resorting to genAI,
> nor is it about "virtue signaling".

Attachment: signature.asc
Description: PGP signature

Reply via email to