Login
You're viewing the techdon.dev public feed.
  • Sep 24, 2026, 2:30 PM

    RE: toot.cat/@zeenix/1173263986563

    I'm sure this will be dismissed by some folks as "tone policing" but I'll once again try to make this case: the overheated, straight-to-rage reactions are so detrimental to the cause of keeping LLMs/AIs out of FOSS projects.

    I am not saying that people aren't justified in their feelings. I am saying that every person is responsible for what they do with their feelings, and I have less than zero sympathy for the argument that when a person is angry they are not responsible for managing their response.

    One of the reasons I'm passionately in favor of projects adopting anti-LLM policies is that I still believe in the Apache Software Foundation's motto of "community over code". My love of FOSS has always been mixed with the appreciation of the people and communities behind the code. LLMs are, at best, rejection of community ethos. They are the "go alone" in the saying "If you want to go far, go together. If you want to go fast, go alone".

    But hate and contempt are always, always, always corrosive. You cannot build on that foundation, and you cannot persuade people when you're demonstrating naked contempt for them. You cannot. The attempt to create a hostile environment for FOSS maintainers in order to coerce them into anti-LLM policies is not going to preserve a FOSS culture we want to be part of.

    It is counter-productive to swarm project issue trackers and discussion fora with angry messages calling people names. Believe it or not, if you've never spoken up in a project before, now is probably *not* the time to create an account and start angrily posting that the project is "mask off" because it's considering an LLM policy shy of absolute LLM banishment.

    I think the "KDE for people" thing is fine. I think KDE Eco's statement was good. Talking on social media about this is fine - without the contempt.

    I get that there's so much to be angry about right now that it's really hard to keep it focused where it needs to be. But if we turn the anger on each other, we're hosed. Just absolutely hosed. I'm sure that the Sam Altmans, Elon Musks, and DHHs of the world are absolutely thrilled to see the anti-LLM folks going full PETA in these discussions, because they fracture communities that could stand up against them.

    💬 1🔄 6⭐ 18

Replies

  • Sep 24, 2026, 2:49 PM

    @jzb

    > LLMs are, at best, rejection of community ethos

    Firstly, you're more than entitled to that opinion. As long as you don't expect this to be shared by everyone, all is good.

    The problem is that the reality doesn't care about our ethics. The undeniable fact is that LLMs give you an advantage over your opponents and this advantage is growing at a spectacular rate (so far at least). FOSS saying "no" to LLM is going to be a death sentence to it, in the (not so) long run, given that fascists are taking full advantage of LLMs.

    💬 0🔄 0⭐ 2
  • 💬 0🔄 0⭐ 15
  • Sep 29, 2026, 2:09 PM

    @hbons @zeenix @jzb i tried to write about this last weekend. I think there is a ln approach to this that is more positive than utterly denying the use of LLMs. I don't think that line of thinking is tenable for software development in the very near future, even if you freeze the technology where it is today.

    If we can accept that, then we can engage with the problems more realistically -- how do we collaborate and keep the community part of the process, can we work with and improve open models, can we train/refine models or figure other patterns for license and copyright awareness.

    We also adopted a similar humans-in-the-loop-onlu policy in PipeWire, but the maintainers are very aware this is too restrictive and we likely will need to adapt as we figure out good ways to work with the tools without adding maintainer burden (maybe even reduce it!)

    This was what I wrote: arunraghavan.net/2026/09/on-ll

    💬 0🔄 0⭐ 2
  • 💬 0🔄 0⭐ 2
  • Sep 29, 2026, 3:55 PM

    @hbons @zeenix @jzb it was more a call for positive engagement, i.e. "how can we make this work" rather than "shun because of all the negative impacts".

    Because it's not just sometimes useful, it is changing how software development happens overall.

    Perhaps this looks like a fad, a couple of years ago I might have agreed, but now we are well past that.

    💬 0🔄 0⭐ 2
  • 💬 0🔄 0⭐ 0
  • 💬 0🔄 2⭐ 3