Post #1602374

2026-03-05 20:04 UTC

@foolishowl @zerodogg what is the objection to specifically, though? Large language models as a technology, or specific vendors? Debian is never going to hand out licenses for an Anthropic or OpenAI product simply on the basis that they're proprietary software. But it's not like the project has ever *banned* the use of a paid, proprietary IDE to support one's work—how would it even know? There exist LLMs trained on public, open data sets with public weights that can run on a personal machine, and would appear to be suitable for inclusion in Debian—are these tools also objectionable? If the goal is "completely halt use of LLMs as a technology used in any form", I don't think it's realistic to expect Debian Developers to be able to accomplish this.

Replies (2)

@foolishowl@social.coop 2026-03-05 20:06

@ehashman @zerodogg No, I agree that there's only so much Debian could do. I'd just prefer that the Debian Project puts up some resistance, even if all it can do is token resistance.

Open ##1602375
@zerodogg@hachyderm.io 2026-03-06 13:10

@ehashman @foolishowl To begin with your last point, the goal. It is, in my mind, for Debian to take a principled stance, and base that on its existing values and ethics. It's not up to Debian if LLM's continue to see use in society at large, just as it's not up to Debian if people keep using non-free alternatives. But Debian can make a difference, and I'd argue that Debian does so every day. Debian *is* different from many other distributions. One only needs to look at the social contract to see that, and Debian goes to great lengths to uphold it. Debian is important partly because it puts its values first. As for my objections, at this point it's hard to differentiate between the technology of LLMs and the harms that stem from the usage of them, mostly because those making said technology encourage harmful usage. Workers rights, a complete disrespect and active undermining of free software licenses, contributions to completely destroying the web by flooding it with nonsense and the environmental impact of training, are among my objections. I would say it's near impossible to use one in any kind of ethical way at this stage. If Debian accepts contributions made with LLMs, then I would argue Debian is contributing to the negative effects I list above, partly by legitimizing them. That said, If a model that respects licenses of the works it uses, a high bar it seems, has all of its training data, weights and other software licensed under a free software license, and is trained without significant environmental impact and without abuse of workers - then yes, that might be something that I could see be acceptable for contributions to Debian. I'll also accept that there might still be LLM contributions being made, but without disclosure. But that doesn't erase the impact of taking a principled stance. I don't need Debian to change the world on its own. But I don't prefer Debian on technical grounds alone, just like I don't write free software just because I need a particular piece of software myself.

Open ##1602376