Hi, Gabriel Wicki <[email protected]> skribis:
> On Mon, Jun 29, 2026 at 08:30:49PM +0200, Ludovic Courtès wrote: >> The policy is enforced by banning contributions, not contributors: >> [...] >> That way, a contributor who violates the policy but genuinely didn’t >> know about it gets a chance to learn about it and remains welcome to >> contribute, this time following the project’s terms. >> > >> I hope this makes sense! > Not really, no. I think you are missing my point, here. In your > proposal, this: >>> genAI MUST NOT be used for direct interaction with other participants > is a rule on how people interact with each other. The way I understand > it, it should prevent contributors from letting genAI interact with us, > project members. > > I am not in favor of punishing unknowing people—on the contrary! But I > don't see why we should not *ban* genAI bot accounts that feed us slop. > Or wasn't that one of your main concerns? It wasn’t one my concerns, but it’s definitely a valid one! I was under the assumption that people would have the bot use their own account to occasionally interact with others (I’ve seen that before), but it’s true that it could be a separate account. I think we could treat it as spam (be it on mailing lists or Codeberg). WDYT? > >> Well, yes and no. The rule that, when someone submits code, either it’s >> non-copyrightable or they have a valid copyright claim on it, has always >> existed, but it’s an unwritten rule (like in most free software >> projects). >> >> The GCD adds this new rule about LLM-generated code, which has to be >> treated specially. > IMO we need to write down the rules for human contributors, before we > ratify special cases for machine generated ones. IIUC when specifying > these rules adequately the whole GCD008 debate evaporates, since it > seems to not add much, that could not have been specified otherwise. This GCD specifies commitments and a policy towards genAI contributions. Many other projects have done the same because of a common desire to adapt to this new technology, to set boundaries, and to mitigate risks. I think it’s important to do so; I spelled out some of the risks associated with the status quo—i.e., non-regulation of genAI use: https://codeberg.org/guix/guix-consensus-documents/pulls/13#issuecomment-16470989 I hope this clarifies why, to me, this GCD does have value. Ludo’.
