"Thompson, David" <[email protected]> writes: > On Wed, Jul 29, 2026 at 8:53 PM Maxim Cournoyer <[email protected]> wrote: > > Hi David, > > I understand you are unhappy about GCD 008 having been withdrawn, but > reacting to it with a satire (GCD 009) is not a productive use of our > time. Please note that trolling is not an acceptable behavior according > to our CoC, which I believe this satire qualifies as.
I don't believe David's actions qualify as a code of conduct violation. Nobody has been targeted negatively in any of David's writings, and an important point was being made. And, in a sense, I believe David (or perhaps Grok's) submission was fairly close to what I think a pro-LLM GCD would look like. If anyone can be accused of a Code of Conduct violation, it's closer to accuse me as collaborator on this; my satirical inline review was intentionally pointed. However, I do not think I committed a CoC violation either; I did not target anyone directly, and was fairly up front within a short period that this was satire... though I think that honestly should have been obvious to everyone by the primary author being Grok. (I *do* think there has been intentional sealioning on the list lately from a certain member, and that does border on if not cross over to a CoC violation for me, personally. But let's put that aside for now.) > This is far from trolling. I am drawing an important contrast between > policy expressed in the negative vs. the positive. If LLM > contributions are neither prohibited (because GCD 008 has been > withdrawn) nor permitted (if GCD 009 fails to make it, which has yet > to be determined) then what does that mean? Are we okay with > Schrödinger's LLM policy? Some are under the impression that the GNU > moratorium applies to Guix, but I was under the impression that Guix > makes governance decisions of its own and thus the policy does not > apply to us. Some are even under the impression that GNU Kind > Communication Guidelines apply here and not the aforementioned > CoC. Some clarity is needed! > > If this GCD goes too far in the pro-LLM direction then is there a GCD > 010 that, say, implements a moratorium that could achieve consensus? > I think we need to come to a consensus on *something*, however > temporary, so that further discussion can happen without so much > uncertainty about the present. > > - Dave It's no surprise that I agree with everything Dave said above. I also want to highlight that there is a history of satirical specification proposals that can be useful and illustrative. Certainly, you don't want them to be common form, but when few other options are available (and I will make it clear why I think this was so), I believe they can be important. The most well known example of this is RFC 3514, "The Security Flag in the IPv4 Header", also known as the Evil Bit RFC. This RFC was never meant to be intentionally implemented, but it has an enormous and even positive effect on technical standards by being something that could be pointed to. The spec *looks* like a serious document, but upon reflecting on reading it, the absurdity of having an IP packet flag that is instructed to be turned on when something malicious is happening becomes clearer and clearer. However, over and over again in standards spaces, people begin to think of things that *resemble* the Evil Bit RFC, and so in standards and technical protocol circles, "the Evil Bit RFC" serves as an important and useful reference point to this day. Without the Evil Bit RFC, many worse standards decisions would be made (and I can say this with confidence as a woman who has had to sit in the room where such standards are being decided) because it is easy to come to think of *non-satirical* solutions which, upon reflection, just turn out to be the Evil Bit RFC. The intent of submitting GCD 009 was not to waste time or antagonize. The intent was something much more serious. By having Grok draft the GCD and be primary author, and by submitting something that appears in the shape of what *would* be a more likely pro-LLM document, we have a reference point to begin a conversation which is incredibly critical at this moment, which is the following: "If instead of GCD 008 which asks to forbid LLM use, we had a GCD which asks to permit it, what would the result be?" I think we all know what the result would be. The point of GCD 009 is to make that abundantly clear. Furthermore, it is meant to demonstrate the question: *where is the onus of consensus on LLM usage?* For this reason, I say with utmost seriousness: I believe that until someone can develop a GCD that gains consensus in the inverse, that LLM usage *should* be permitted, the de-facto LLM moratorium stands. And that is the statement GCD 009 is making, and despite the shape of it being satirical, it is a serious one indeed. - Christine
