OPINION Linux now being officially not anti-AI might be good news for projects that are explicitly anti-AI, but we foresee problems with conflict over where the money and development effort come from.A couple of weeks ago, reluctant geek superstar Linus Torvalds declared that Linux is not an anti-AI project. The pro-AI folks are jubilant, and they are especially loud on the internet. (They're probably using bots to write their posts.) However, Torvalds inadvertently issued a rallying cry for the people and projects that are anti-AI. The anti-bot resistance is growing.European volunteer-run code forge Codeberg has changed its Terms of Use to ban projects that "mostly consist of code written by 'generative AI' tools," as well as projects involving cryptocurrencies. Some have questioned the connection, but as Pivot to AI notes right in its name, many former blockchain maximalists switched to promoting LLMs instead after the 2022 "crypto winter."
This is just one sign that despite the legions of AI advocates, there is also a growing contingent strongly opposed to AI and its use. As Futurism reported back in March, Fewer Americans Like AI Than ICE.
Torvalds' position is pragmatic: his interest is in whether it works or not. That is not the only question. There are other important debates about the legitimacy of the vast data harvesting to build the LLMs; about the environmental costs, not merely of their use, but just as importantly, of their training; about how frequent LLM use impacts their human users; and about LLMs' effects on the job market, and the wider economy.Many of these debates come down to questions about the ethics of AI use. A lot of people in the software world are not accustomed to thinking about the ethical consequences of what they do and how they do it. One result is profound and passionate disagreements – and that could defeat the entire anti-AI cause.Anti-AI FOSS OSesWe've seen some superb responses to Torvalds' ex cathedra… well, we'd call it some kind of bull, but that seems unfair: even the Pope has serious misgivings about AI.Long-term observer and reporter on the world of system software OSnews summarized its take: "Asbestos is a tool, just like any other."On social media, one of the more on-target responses we saw was an enjoyably ranty thread by a Fediverse member with the wonderful username of [object Object] (see Bootnote). Among other comments, this one from R. L. Dane gave us a laugh: "Lemme see if I can wrap my brain around 2026 now… Linux is Windows, Windows is Facebook, Facebook is Palantir, and Palantir is basically the devil. Oh, and now BSD is Linux. 😁 🖖🏼 Carry on."The mention of BSD is because NetBSD has treated LLM-generated code as presumptively tainted since 2024. OpenBSD's project lead says that since LLM code can't be copyrighted, it can't be contributed – but as we noted when looking at OpenBSD 7.9, it does grandfather in other projects, such as tmux, which do accept LLM-assisted contributions. A year after NetBSD set its guidelines, FreeBSD was still considering the issue.There is already a list of Linux distros with no systemd, of course, but we feel that there are plenty more things that Linux distributions could do to move in an anti-corporate direction – indeed, we added our own speculations in the thread.
