On 30/07/26 at 12:57 +0200, Johannes Schauer Marin Rodrigues wrote: > Hi Lucas and everybody else, > > Quoting Lucas Nussbaum (2026-07-30 09:16:14) > > On 30/07/26 at 07:48 +0200, Johannes Schauer Marin Rodrigues wrote: > > > But I > > > think it became clear from the other mega-thread that there are a lot of > > > people > > > with very strong feelings about this topic. In fact, their feelings are > > > apparently so strong that they open the big can of worms that is a GR, we > > > spend > > > a *lot* of our time on very difficult discussions and we risk a hard > > > division > > > of our community or even some of our members leaving. That's tough stuff. > > > Apparently, some in our community feel strong enough about this topic > > > that they > > > are willing to risk all of that and put so much trouble on their fellow > > > DD's > > > limited Debian time. Even if this whole LLM topic is fine for you, you > > > probably > > > see that this would not've been made into a GR if some of our friends > > > would not > > > feel *very* strongly (as in: badly) about that topic. > > > > This paragraph works both ways: > > Some of us are also convinced that AIs is fundamentally changing > > software engineering, and want to use that technology in the context of > > Debian, to the point that they would lack motivation to continue working > > on Debian if it was not possible. > > my paragraph was meant in the context of the Debian related work we do right > now. I am not using LLMs to do anything Debian related and I fail to come up > with anything that my refusal to use LLMs could result in you feeling bad over > my work. Me not using LLMs likely does not make you feel bad.
I'm not trying to convince you (or others in the anti-LLM camp) to use LLMs. I agree that there are valid concerns about LLMs, that need to be weighted and compared to their usefulness. I understand that this process is person-specific; I respect the willingness not to use LLMs, and I also respect the willingness not to read LLMs output, up to the point where it clearly harms my own ability to use LLMs in the context of Debian. What I meant with "This paragraph works both ways" is that "a lot of people" in your paragraph could apply both to pro-AI and anti-AI contributors. I'm not trying to diminish the feelings of those in the anti-AI camp, but there are also people in the pro-AI camp that would reduce or stop their Debian activities if it was not possible for them to use AI in that context. > Sorry, I explained myself poorly there. Your message was fine (for me). I like > that you disclose your use of genAI and your disclosure of your prompt. I > have > read your SPEC.md and you have clearly put a lot of work into this. I don't > want to belittle that work. Your SPEC.md has been useful to me for example > because of the 12 bullet points which I can use to rate my own understanding > of > each ballot entry. So what I would've preferred is, if you had not spent time > on writing this prompt and then send it to some LLM to generate HTML and > markdown pages. Your message itself was fine. Full disclosure: SPEC.md is also AI-assisted. Basically, in an interactive session using a local agent (OpenCode), I started by throwing an LLM at the list of proposals, iterated a bit (such as manually adding some comparison criterias), also fed it the whole discussing to let it identify additional comparison criterias, and then asked it to generate SPEC.md to keep a trace (~ "source code") I could reuse later. That process of saving the user's intent is something I've also been doing for code, by asking the LLM to include a "Prompt-Summary" trailer in commit messages (summarizing the various exchanges that happened interactively so that the commit message captures the original motivation for the change, not just a description of the change itself). > I see you saw my "Kitten" example in the other mail. I used this example > because I think (hope!) that we can agree that killing baby kittens is really, > really bad and we should not be doing that. Going back to being more > realistic: > if I would now try to weigh costs versus benefits of use of genAI then on the > one hand we have for the costs a long list containing items like: all the big > and powerful models being owned by billion dollar corporations, big enough to > sway politics, big enough to sway public opinion, that make the rich richer > (with a surprising number of rich people with clearly anti-democratic fascist > view points), the poor poorer (no everybody can be an artist!), support the > climate crisis with data centers requiring more electricity than nation > states, > use and pollute water, requiring supercomputers to train (not something we do > as part of a source package build), scraping our web because suddenly > everything is free game (server admins hate this trick), making me want to buy > faster hardware because my browser has to solve anubis riddles on every > website > i use, oh no the hardware is now super expensive so you can only afford it if > you have enough cash around, the whole copyright dilemma, the whole > hallucination issue, the issue with inherent sexism and racism (why are these > points so low on my list?), the problem that if the LLM is wrong I cannot "fix > it", pushing back on the idea of "free software needs free tools", the deep > fakes, the problems for my students with learning because the LLM just hands > them the answer (yes the smart ones will still be smart but many will just not > put the effort anymore) and so on... Oh and there is also the small thing of > my hobby being replaced by a machine but that issue really pales in contrast > to > the long list of *real* problems. > > And on the list of benefits we have, very cynically speaking: it now was much > easier to create a 17 kB HTML table. > > Now you can go and you can challenge any of my points above. And yes, your > HTML > table will not make the oceans boil or democracy fail. But the things we do do > not happen in a vacuum. The things we do influence others and they normalize > how all of us (even outside Debian) do things going forward. The free software > movement also started small and see where we are now. I would not like this to > become another thread where we discuss how local models are more ethical than > claude/chatgpt (they are) or how things will be so much more efficient in the > future (they probably will be). The point I'm trying to make is this: > > Most of us are in a very, very privileged position. Many of us can work for > Debian because we either have secure jobs where we are paid to do this or > because we live a sheltered and financially secure enough life to do this in > our free time. Compared to most of us on this planet, we can count us as > *extremely* well off. And I understand how for probably all of us, creating > software is something very important to us because otherwise we wouldn't put > so > much of our time into it. But at the end of the day: it's just software. It's > just an interesting sequence of zeroes and ones. Being able to produce this > sequence faster is a first-world-problem of such a high degree of privilege > that it pales against most of the issues I have with LLMs even if all of the > issues in my list above would only be half as bad as I think they are. > > For myself this means that I rather "limit myself" and write software "slowly" > or that I don't do some things at all. That way, I'm not part of the problems > I > see and instead am part of the world that I want to live in. > > Sorry Lucas that this long (and emotional) mail was now a direct reply to > yours. I *highly* respect your work and your technical expertise. Just to > affirm, I do not think that you acted in bad faith and I don't think that your > messages were written badly. It is useful for me to read your messages. Thanks a lot for this write-up. I really don't want to sound as if I'm trying to convince you that the benefits of AI outweight its costs. But there are a couple of things I'd like to write to explain my point of view. I still need to organize my thoughts so I'll do that later. Lucas
signature.asc
Description: PGP signature
