Dropsafe

by Alec Muffett

Clade of “Human Emacs” developers want Free/OpenSource software to be small-batch, handmade, organic, untouched by LLMs

To me it’s rather ironic that the people having this argument are largely working in the archetypal AI research language of the 1960s, 70s, and 80s:

indeed, i have decided that llms are, at every level, an affront to justice, from who makes and profits and pushes them, to how they are trained, to the resources used, to the way they are used, to what they do to people who use them, to what they do to the communities where their use is tolerated. everywhere, and in profound ways, i have looked and seen llms as a destructive force that only exist to empower the few at the incredible, immoral expense of the many.

Quoted in:

https://yhetil.org/emacs-devel/86mrrmknn7.fsf@gnu.org

Via Royce:

Royce Williams

Since the GNU listserv is getting Slashdotted (probably in part due to people using LLMs to search / summarize it, ironically)

lists.gnu.org/archive/html/ema

(Please don’t use this thread to discuss the thing, I’m just trying to inform the thing. And to be clear on my own stance, I’m not just presenting this neutrally: I think there’s a forest for the trees problem here, to put it mildly — freedoms can’t be isolated the way he wants them to be)

Re: Human Emacs From: Eli Zaretskii Subject: Re: Human Emacs Date: Fri, 06 Nov 2026 17:03:40 +0200 > From: brian cully <bjc@spork.org> > Cc: jporterbugs@gmail.com, bugs@gnu.support, phil@hagelb.org, > emacs-devel@gnu.org > Date: Fri, 06 Nov 2026 08:55:29 -0500 > > Eli Zaretskii <eliz@gnu.org> writes: > > > No one explained what kind of hostility will happen, and people here, > > myself included, don't really understand what is alluded to. It > > almost seems like you've decided that LLMs have adverse effects in > > some of their popular uses, and therefore using LLM-generated code > > will somehow "infect" Emacs even though those problematic aspects of > > using LLMs are completely irrelevant for our case. I personally hope > > that this is not the reason, but in the absence of any other > > argumentation, what else is left? > > indeed, i have decided that llms are, at every level, an affront to > justice, from who makes and profits and pushes them, to how they are > trained, to the resources used, to the way they are used, to what they > do to people who use them, to what they do to the communities where > their use is tolerated. everywhere, and in profound ways, i have looked > and seen llms as a destructive force that only exist to empower the few > at the incredible, immoral expense of the many. So basically, you are saying that just because LLMs can and perhaps do harm elsewhere, we should not let them anywhere near Emacs, regardless of what they can or cannot do to Emacs and our community. > in my view, while there may be some benefits to llms in some > circumstances, their harms vastly outweigh those, and, if i want to live > in accord with my principles of justice, i am left with no choice but to > reject them at every opportunity and in every way. i cannot come to > another conclusion based on my sense of ethics. There's a lot of evil in this world, and we meet it all the time. In the Free Software movement, we only deal with evil that directly affects or threatens software freedoms. Fighting the rest should be left for other arenas. > > That's very different. Here, we are asked to reject a completely > > valid, free and correct code just because of its progeny. It's like > > saying that code from left-handed or red-haired people should not be > > accepted. > > my point is not the progeny, but that the gnu project often rejects > technically and legally valid contributions because of gnu > ideology. progeny can be factored into that. No, progeny _cannot_ be factored into that. It shouldn't be. It must not be. > as a hypothetical example: if i live in a territory that permits > slavery, and by law i am granted sole ownership of the products of that > forced labor, and then i use that to create contributions to gnu emacs, > complete with copyright assignment, and it is technically correct, and, > above all, you, the emacs maintainers are aware of how this was > produced, would it be ethical to accept those contributions? Our answer must be YES. Any other answer is a first step towards splitting our community along the many lines of political and ideological divides that have nothing to do with software freedoms. Right now we have contributors from all over the world, despite the differences in their regimes and political systems, differences in ideologies, and even in the very interpretation of the notion of "freedom". You are suggesting to subject the contributors to some examination of the principles they uphold, the ethics of their day-to-day activities, their views on matters of policies and important international events, etc. This will immediately fracture our community into many different enclaves. And that's exactly my point, and the main reason why I must reject the "Human Emacs" initiative, as it is presented to us. > it is my worry that the fsf and gnu's apparent narrow focus on legality > and technical correctness in the framing of llm contributions ignores > the much larger issues with supporting human freedom, and, as such, > becomes hypocritcal: free software is free in name only when it is used > for subjugation or created from subjugation. There's no hypocrisy here, only the realization and understanding that we cannot fix everything in this world, even if we tried, not via developing and distributing software. So we limit ourselves to solving a few of the problems, and leave the rest to other communities. (Of course, people who contribute to Emacs can well be members of those other communities as well.) > this example is hypothetical and not intended to display any real or > imagined link to how llms are created and operated. it is only intended > as a philosophical point on how a narrow focus on license legality and > technical correctness are, at times, inadequate in the pursuit of the > larger goal of liberty for all humanity. The Free Software movement doesn't pursue the goal of "liberty for all humanity". If you think it does, you are gravely mistaken. Our goals are different, and are well described elsewhere.
2026/11/08, 18:09 0 boosts 3 favorites
⊞

Comments

4 responses to “Clade of “Human Emacs” developers want Free/OpenSource software to be small-batch, handmade, organic, untouched by LLMs”

  1. Daniel Kochma?ski

    @alecm

    That tension is not anything ironic nor new in the AI discourse. You would do good reading Weizenbaum's "Computer Power and Human Reason" to get some context. As a bonus point, he wrote the very first chatbot.

    If you are interested in reaching even further into the past, the founding idea of AI was symbolic computation, and that was its differentiating feature from cybernetics that were concerned mostly with feedback loops (neural networks came from this field), and it only later absorbed the latter.

    There is also a strong argument on strictly technical merits, that until the copyright status of LLM training and output is properly codified within a law, it is in practice proprietary technology with a sprinkle of wishful thinking that "we are too many to render this not a fair use".

    1. alecm

      Dude. I once met John McCarthy, albeit briefly, and I know a few MIT & CSAIL alumni.

      Reply
      1. Daniel Kochma?ski

        @alecm don't dude me dude 😉 I'm judging your knowledge, not your contact list.

        1. alecm

          You’re judging something, but it’s not my knowledge.

          Reply

Leave a Reply

Your email address will not be published. Required fields are marked *

More posts