On Thu, Jun 25, 2026 at 01:49:05PM +0200, λx.x wrote: > > Before this thread starts trading apocalyptic visions of future societal > > breakdown caused by LLMs and our eventual turn to cannibalism > > i do not see how this is relevant
OK, I was (trying) to make a joke, I'll be more explicit. The purpose was to encourage you to engage with the original posters argument rather than simply restating a position that is known. I laid out some reasons why a strawman of "but this (the environment) is something we should care about" is equally open to refutation. > > That directly negates the possibility of "taking stances" and "stating > > opinions" which are divisive, controversial and/or that don't achieve > > consensus. > > we can take stances, and state divisive opinions, and still reach consensus > about implementing those stances and opinions. the GCD process is not about > stating what everybody already agrees with, it is about finding consensus on > implementing everyone's opinions in a way that finds broad acceptance. if i > had to hazard a guess, what GCD 008 is today is much closer to finding broad > acceptance of contributors than either of completely allowing LLM > contributions and a blanket LLM ban, and this would not have been possible > if we only took stances and stated opinions that everybody else already > agreed with. failing that, we can still state opinions—as individuals, not > as a whole—that do not achieve consensus. No, you've missed my point entirely. If you're interested re-read what I said, if not lets move on. > > Second, if this is about 'political change' then I question what's > > special about LLMs? Guix doesn't have GCD's that cover use of public > > transport, cutting down forests or reducing domestic violence against > > women and girls. > > only one of those things is of immediate relevance to us; as far as i know, > there is no software that blows up busses, cuts down forests, or contributes > to domestic violence in a way that is usable as a tool for Guix and > contributors, or in any way relevant to the project otherwise—i would be > dismayed to be proven wrong. That's your perspective because the environment is immediately relevant to you. That justifies regulating it's usage. These are all in the same category of concern "societal". Other people, have other "societal concerns". There's no category difference. Justifying a rule due to something being a societal concern so we should regulate how it's used in reference to Guix is a big move. This is precisely what the original poster was pointing out. Age verification is exactly an example of this sort of "societal concern" being brought into technology. The belief that this societal concern is so bad is what justifies regulating how people use their own computer is exactly the same argument as stating that LLMs are so bad that they must not be used for contributions to Guix. The justification for regulating behaviour is the societal harm. > > We need to take "stances" we all agree on, and I (and the previous > > writer) do not agree that a Free Software project should be (a) be > > taking stances on societal topics, (b) constraining freedom of > > individuals. > > i do not see a way «LLM contributions are not allowed*» is constraining > freedom of individuals. if this is constraining the freedom of > contributors, then disallowing packaging and promotion of nonfree software > also does, as does the fact that we expect the contributors to have the > legal right to submit their contributions. i would think the last reason > alone is enough to reconsider whether allowing LLM contributions—at least at > this point—is a good idea, as we still have no idea about what the legal > status of LLM output is. This is a move from regulating outcome to adding process. We are a Free Software project which has a specific meaning, we follow a policy that regulates us (FSDG) within those freedoms. We don't allow proprietary software for those reasons. We have rules about our contributions which are technical and they're related to how something is delivered to us (an outcome such as a PR, a particular format for the log) so that we "as engineers" can do a better job. We do not regulate how someone uses our software as long as they follow the rules of Free Software. We do not regulate HOW they achieve a contribution, we don't tell them they can only use Free Software, we don't tell them what Web sites they're allowed to go to or search engines to use. This is a change in category, we'll be explicitly telling contributors what they may or may not do on their own computers. Steve / Futurile
