Regardless of how you personally feel about it, it seems the AI coding assistants are here to stay.Although there are plenty of problems when it comes to AI slop code, vibe coding by people who don't understand the code itself, or the illusions of productivity destroyed by excessive quality control, coders are not letting this one go.I've argued before that AI slop coding could destroy open-source software, and my colleague Jorge Aguilar has shown how Linux has an AI problem with no clear solution.
From the outside, looking in, this seems like a very precarious situation, but what are major Linux groups actually doing? Like most things in the world of Linux, it seems everyone has decided to just go their own way.Linux's biggest projects are suddenly taking very different positions on AI Do not go Gentoo into the night Some of the biggest names in Linux have started to plant their flags on different sides of the AI debate.Gentoo Linux has a clear anti-AI policy in its wiki, last updated in 2024 as of this writing.
It is expressly forbidden to contribute to Gentoo any content that has been created with the assistance of Natural Language Processing artificial intelligence tools.This motion can be revisited, should a case be made for such a tool that does not pose copyright, ethical and quality concerns.The council voted this way based on concerns around copyright, quality, and ethics.
Arguing that LLMs have the potential to breach standards on all three fronts.Representing serious liability for the project.NetBSD (which is not Linux, I know!) has an interesting stance from its side of the fence.
The NetBSD commit guidelines effectively consider AI-generated or assisted code as "tainted" by default, and you need written permission to contribute it.Code generated by a large language model or similar technology, such as GitHub/Microsoft's Copilot, OpenAI's ChatGPT, or Facebook/Meta's Code Llama, is presumed to be tainted code, and must not be committed without prior written approval by core.A study by Hora and Robbes found that across 1,000 general open-source projects, "." This indicates that in the FOSS community as a whole, things seem to be leaning toward "yes" for AI, albeit with explicit rules and conditions.
Fedora is trying to split the difference, and that's where things get interesting Trying to tip your hat to both sides is tough Close Linux distro titan Fedora is taking a more measured stance than Gentoo or NetBSD.Fedora's stance is "AI as a Tool to Advance Free Software" and this key line from the council policy proposal stands out to me: We encourage the use of AI assistants as an evolution of the contributor toolkit.However, human oversight remains critical.
The contributor is always the author and is fully accountable for their contributions.This puts the onus for quality control on the contributor, and if you contribute sloppy code, you will be treated the same as someone who wrote that slop manually.Linus Torvalds has basically told the anti-AI camp where to go “Fork it.
Or just walk away.” Linus Torvalds, the father of Linux, is on-record as being somewhat annoyed at the trivial and sloppy pull requests people have been sending using AI.But don't think that Mr.Torvalds is anti-AI, far from it.
As reported in Ars Technica, Linus thinks the blanket banning of LLMs is wrong-headed, saying, "" It seems that after experimenting with vibe coding himself, the penguin godfather is sold on the technology.Though, as critics point out, the problems with LLMs in FOSS might not be technical, but rather ethical and philosophical.The real battle isn't about AI.
It's about who pays for AI's mistakes In 2024, I wrote that AI search engines only work if you can trust them, and without a human in the loop, that requirement can never be met.The AI policies we've seen for Linux so far seem to emphasize personal responsibility above all, or simply bans the use of LLMs entirely.In the end, the biggest problem here isn't that the code is generated by an LLM, but that some people are using LLMs to review that code as well.
Once code starts entering Linux that has never been eyeballed by a human, and is treated like a black box, the entire OS becomes corrupted.In 2025, CNBC reported that Microsoft CEO Satya Nadella says up to 30% of Microsoft code is written by AI.I can't help but feel this fact is at least part of why I left Windows behind years ago, and it would be a tragedy if the Linux community mismanages this issue.
Read More
