Being Against LLMs is Against the Spirit of FLOSS
Published:<br>15 August 2026 / 28 Thermidor 234<br>Last Edited:<br>15 August 2026<br>/ 28 Thermidor 234
Being Against LLMs is Against the Spirit of FLOSS
Earlier this year I wrote Vibe-coding Brings the Power of Emacs to<br>Everything in a more positive manner surrounding this topic. This is<br>the negative version. I am not a radical in either direction<br>surrounding this debate. But the internet does not reward subtlety or<br>nuance, as I also pointed out in Three Thoughts on Crypto and nfts.
The debate surrounding llm-generated code is more intense than<br>ever. There are some people who have recently tried to turn this<br>debate into action as an attempt to stem the tide of llm contributions<br>that are flooding maintainers. Most notably Vim “Classic” and my own<br>adopted public git forge of Codeberg. This is a shame, since it flies<br>against the spirit of Free (Libre) and Open Source Software (floss).
Linus Torvalds has effectively been vibe-coding the Linux kernel for<br>decades (his models being the world’s best kernel programmers and his<br>agent harness email) and that has clearly turned out fine, but I still<br>write most my code the old-fashioned way (at least in a raw loc-way)<br>because I think that the hardest parts about programming are not<br>solved by llms. I also do not like to refer to llms as “ai”.1
You are of course free to not accept contributions made by llms (or<br>any contribution for any reason) to your own project, and you are also<br>not forced to host someone else’s vibe-coded projects on your<br>infrastructure; the beauty of floss is that they are then also free to<br>fork your project.2 But rejecting code not for technical reasons (of<br>which there are often many for vibecoded code) but for ideological<br>reasons is unfortunate.
At the same time there is of course a genuine problem for maintainers<br>happening that I do not know how to solve. One possible solution may<br>be to only accept contributions that you yourself believe to be<br>written by a human, while nominally accepting llms to be used in any<br>capacity. This means that llms can be used when they are useful, but<br>that code must remain readable and succinct enough for humans to<br>process and develop.
llms are fundamentally a tool. Trying to ban llms is giving them too<br>much credit; it is contributing to the framing made by companies like<br>Anthropic and Openai that llms are a new class of software that is so<br>powerful it must be banned, and entirely unlike all the other tools<br>that make software development easier. Imagine banning the use of<br>compilers and only accepting contributions written in Assembly as a<br>way to reduce the number of contributions. It works, but is it in any<br>way reasonable?
Such a policy — that contributing to a project because of the<br>circumstances of the contribution’s creation, and not due to its<br>quality — is directly contradictory to the spirit of free<br>software. Being able to read, spread, and easily customize a program<br>is just as useful as being able to merely run it. And the long-term<br>strengths of free software is that it can adapt to everyone’s<br>(collective) needs instead of each person building their own custom<br>software.3
Arguments that llms fundamentally a part of a system of political<br>oppression fall flat against the case of open-weight models that can<br>be run at home on (high-end) consumer hardware. Are you going to be<br>against online search in general because most people use Google as<br>their provider?
The Russian-born French philosopher Alexandre Kojève pointed out that<br>— when studying the history of Hegel’s geist — it is uninteresting<br>to study the existence of the nihilist, because the path of nihilism<br>is one of (intellectual) suicide. This is what is happening to these<br>projects ideologically opposed to llms; projects like Vim<br>Classic. They are intellectually oriented not towards improving the<br>program either through bug-fixes or new features, but merely towards<br>making sure that every line of code was written by hand (wilfully<br>ignoring other automated tools such as lsp, the aforementioned<br>compilers, macros, automated config tools, even editors<br>themselves!). There are features that could be in the next release of<br>my beloved Emacs right now because they have been blocked because of<br>the initial use of an llm.4
If you actually care about the direct impacts and externalities of<br>llms, this is not the path forward. It shutting your eyes and plugging<br>your ears to the reality that llms are widely used to develop<br>software, akin to what the Qing empire did during the heyday of<br>European colonialism. Instead the solution might be to solve the<br>underlying causes of overconsumption, fossil fuel use, and increasing<br>inequality. ❦
Footnotes:
“ai” is a very broad term, that today mostly refers to the<br>specific technology of Large language models. Other forms of<br>algorithms (both deterministic and not) perform acts of “intelligence”<br>— even very simple...