On Tue, Jul 7, 2026 at 5:02 PM Ludovic Courtès <[email protected]> wrote: > > Hi Greg, Jason, > > Greg Hogan <[email protected]> skribis: > > >> Unless I'm misunderstanding, it sounds like you're proposing that > >> that project members *should* be allowed to use LLMs for > >> contributions. I thought that the doc's intent was rather > >> different, but perhaps Ludo' can clarify. > > > > Ludo' previously stated 'In my mind “contributors” in the policy is > > any contributor, including “project members”' [0] and my request is > > simply to clarify this in the consensus document so that there is no > > confusion when harmonizing the new GCD into the existing project > > documentation. > > WDYT of this: > > We propose that contributions from any participant to Guix be subject > to the following policy: > > ? > > Now, a “project member” is also bound by the commitments above, so in > practice project members are not expected to use genAI “to author code > or packages, to interact with other participants (e.g., to explain code > changes or to review code) […]” (commitment #1). > > That’s the intent, and I thought it was clear, but let me know if it > wasn’t.
This is under the "Policy for Contributions" section? I am not seeing any change from adding "from any participant". Do you see how the first commitment still conflicts with the policy for contributions, which permits contributions from and exploratory analysis by project members? I do think that we should judge contributions on quality and merit rather than discriminate based on the contributor. If we say that a contribution is legally, ethically, socially, technically permitted and acceptable then it should not matter who made the contribution. But if this is not to your satisfaction then let us return to [0] and simply remove the reference to "less than 15-line-long": --8<---------------cut here---------------start------------->8--- > Given the confusion over commitments #1 and #2, could we simplify > these to something like: > > "The project will not use nor encourage use of genAI to author code or > packages, to review code, to produce artwork, translations, or any > other artifacts. These prohibitions do not constrain exploratory > analysis, running guix commands, or authoring non-legally significant > contributions. This is to ensure the contributor has a valid copyright > claim on the code. For example, this includes code less than > 15-line-long, or package definitions that are evidently not creative, > similar to those that guix import and similar tools might produce." OK, sounds good. I find the second sentence ambiguous and would be tempted to summarize it as: Thus rule does not prevent use of genAI for exploratory analysis or to produce non-legally-significant contributions. (“Running guix commands” is fine, whether or not it’s happening via an agent, so I’m not sure what you had in mind.) Ludo’. --8<---------------cut here---------------end--------------->8--- [0] https://lists.gnu.org/archive/html/guix-devel/2026-06/msg00152.html
