Forgot your password?
Close
typodupeerror
Unix AI BSD

AlpineBSD Announced - an AI-free FOSS OS (theregister.com) 29

Posted by EditorDavid from the AI-vs-Unix dept.
Lproven (Slashdot reader #6,030) shares his article about a new OS in the BSD family "combining NetBSD's kernel and libc with a port of the Alpine Linux userland" — and promising it won't use AI-generated code: AlpineBSD is so new that it barely exists yet. There are no installation media available, though it's possible to manually bootstrap an installation. What AlpineBSD does already have is an AI policy, and the plan is to avoid "the output of generative AI models (e.g. LLMs)" as far as possible:

"Any software whose code, documentation, binaries, or other assets include, in any part, the output of a generative AI tool such as an LLM, is considered 'tainted' for our purposes...."

Back in 2024, the NetBSD project made its position on AI plain. Its commit guidelines say that "code generated by a Large Language Model [...] is presumed to be tainted code...."

AlpineBSD's ports repository and source code are hosted on Sourcehut, under the account of sircmpwn, better known as Drew DeVault... We will follow AlpineBSD with interest. There's strong appeal in the plan of combining perhaps the best-known, longest-established AI-free Unix-like kernel with the user-facing parts of pretty much the smallest and cleanest Linux distro there is. The AI advocates often warn LLM skeptics that they risk being left behind. Well, here's an OS for those who don't want to surf the AI wave.
The article does acknowledge difficulty installing software using the pkgin package manager (an optional extra), explaining that since 1993 the NetBSD community has had "its own ideas about what is normal... Some folks on the mailing list told us that they never use binary packages anyway — they compile what they need from source code."

AlpineBSD Announced - an AI-free FOSS OS More | Reply

AlpineBSD Announced - an AI-free FOSS OS

Comments Filter:

  • The AI advocates often warn LLM skeptics that they risk being left behind. Well, here's an OS for those who don't want to surf the AI wave.

    How articles like this leave off the big AI threats of ending the world of the effects on the environment is beyond me.

    I read these things and I feel like the people against AI are given a false choice. Like having someone asking you when you stopped beating your wife.

    • Well, pull up a chair, because you're going to hate this conversation!

      There are two parts to the issues around AI, at least when we're only talking about the ones you brought up.

      The first part, which most people quietly shove under the floorboards and try to ignore, is that the current AI ecosystem is heavily characterized by the hypercapitalist mess. Much of what LLMs actually are does not necessarily pose a threat to the world order, nor does it necessarily involve environmental degradation. Saying "AI ba

  • Reminds me of SQLite's stance (Score:5, Interesting)

    by Sebby ( 238625 ) on Saturday November 07, 2026 @08:14PM (#66414810) Journal

    This stance reminds me a bit of SQLite's stance on contributions: [sqlite.org]

    SQLite is open-source but it is not open-contribution. All the code in SQLite is written by a small team of experts. The project does not accept "pull requests" or patches from anonymous passers-by on the internet

    The main reason in their case is that they want to maintain the code as pure public domain, so don't want it "tainted" by contributions of unknown origin that might later cause issues if used.

    • Re: (Score:2)

      by gweihir ( 88907 )

      The main reason in their case is that they want to maintain the code as pure public domain, so don't want it "tainted" by contributions of unknown origin that might later cause issues if used.

      And given the litigation we have seen against Linux and others, that is an excellent reason. The GCC folks do the same thing for the same reason. Linus, unfortunately, has missed the problem or mistakenly thinks the kernel is not affected.

    • Re:Reminds me of SQLite's stance (Score:5, Interesting)

      by arglebargle_xiv ( 2212710 ) on Saturday November 07, 2026 @10:43PM (#66414890)
      I contribute to a project with a similar policy, and it's not because we're worried about PRs sapping and impurifying our precious bodily fluids but because the code base does something that's quite difficult to understand for most people and so we analyse the issue in a code change request and then rewrite it in the appropriate way. I was at a talk by the SQLite devs years ago and the said that was why they were a small team, you need a huge amount of knowledge of low-level database internals just to begin to work with SQLite, so they don't actively exclude people but there's just naturally a very small number of people able to work on the code.

    • Re: (Score:3, Interesting)

      by karmawarrior ( 311177 )

      And this is a massive concern when it comes to LLMs. To date, there's no evidence that the output of LLMs really is "public domain", the courts are divided on it, and the sloptivists aren't going to be able to assure us LLM code is free of copyright risk until literally the UN intervenes and convenes an international AI copyrights treaty that declares it public domain. Until then different countries may legislate, but as long as any significant jurisdictions tag LLM output as belonging to the people whose w

      • The neat thing about copyright is that it's such an inconsistent issue. People are pro-copyright when they're worried about their work or jobs being at stake, and anti-copyright when they think copyright holders are making too much money. Nobody seems to have a balanced, much less nuanced, concept of what copyright law should actually be. It's just a cudgel that keeps getting passed around, picked up and refashioned to meet the moment into whatever someone needs to wage the war they want to wage.

        Crucially,

        • The balance was found sometime prior to this century. Possibly prior to the 20th century as well. All the recent modifications have just made copyright horrible. I blame the senators bought by the mouse in the 90s that really did us in, but I'm sure there were other "gotcha" moments prior.

  • Footgun... (Score:5, Insightful)

    by x0ra ( 1249540 ) on Saturday November 07, 2026 @09:54PM (#66414862)

    Back in 2024, the NetBSD project made its position on AI plain. Its commit guidelines say that "code generated by a Large Language Model [...] is presumed to be tainted code....

    That's a hell of a way to dig your own grave... not that NetBSD was relevant in any way...

    • Re: (Score:1)

      by caseih ( 160668 )

      I agree. On the other hand they are not wrong about the code being tainted by borrowing from copyrighted code. While you can argue the human brain also taints code in the same way, human wetware injects enough originality and creativity into it that it's not generally considered a problem, except in a few rare cases (for example the contributing to the wine project). AI Llms systematize and automate the tainting on a very large scale.

    • macOS is BSD...

      • macOS's Unix-like userland (ie the command line tools that weren't written for macOS, tools like "ls" and "rm") are mostly from FreeBSD (at one point I think there were some GNU tools mixed in, but I have a vague memory they started to remove those in earnest). None of the rest - the kernel, the GUI, the file systems, device drivers, the equivalent of init/systemd, et all are from BSD derived code, they're mostly proprietary Apple/NEXT.

  • And hopefully they get enough people backing up all the work that it takes to quickly keep the artisanal code fresh. Because they are working against bad people using tools to make it easy to find bugs and instantly turn them into vulnerabilities.

    • This. If I was writing a piece of software and I wanted it to be secure, I would definitely use a local AI to look over my code. It's just another set of eyes and an advisor, and like any advice that comes your way, you decide what to do with it.

  • Unlikely to happen (Score:3)

    by bill_mcgonigle ( 4333 ) * on Saturday November 07, 2026 @11:01PM (#66414908) Homepage Journal

    The repo will just shrink as more packages take on some LLM patches. This doesn't make practical sense.

    It's worth reviewing the ongoing battles between DeVault and Lunduke and you decide who is telling the truth. e.g. Stallman-report.org .

    Just do a web search - DDG gets the headlines at least.

    • That it would shrink seems obvious, but I'm not sure "you decide" is a viable approach to dealing with pedophiles in the OSS community, or something that normal people are going to want to both-sides or "I dunno."

      Great initial point, but highly dubious comparison that makes me question your character.

  • AI works, really well (Score:3, Informative)

    by NynexNinja ( 379583 ) on Sunday November 08, 2026 @02:40AM (#66414996)

    Given the right prompts and direction, AI engineered software is faster and more bug free than what a human can provide at this point. All the grey hairs wearing the tinfoil hats, sorry to bust your bubble, but I would trust a libc written in Rust by AI a hundred times more than one written in C at this point. We have moved beyond hallucinating LLM's generating slop code. Linux kernel just fixed over 2000 verified bugs just in the past few months directly related to AI. OpenAI on november 4th wrote a story about how they have a pipeline that finds, validates and generates working exploit PoC's and also a patch fix along with its submission and has resolved over 1700 serious security vulnerabilities in Linux kernel just inthe month of October. When used and validated correctly, pipelines like this can be used to add features, fix vulnerabilities, find new optimizations, etc. Anyone denying this is not keeping up with current events.

    • Re:AI works, really well (Score:4, Interesting)

      by serviscope_minor ( 664417 ) on Sunday November 08, 2026 @03:49AM (#66415040) Journal

      Given the right prompts and direction, AI engineered software is faster and more bug free than what a human can provide at this point.

      That's a marvellous little concoction: AI is better by construction. Any time it's not, it's the human's fault for not providing the right prompts and direction.

      We have moved beyond hallucinating LLM's generating slop code. Linux kernel just fixed over 2000 verified bugs just in the past few months directly related to AI.

      One of those doesn't follow from the other.

      I've been setting up agents for scanning for security hole at work. It's pretty good, and as someone pointed out recently AI is making us pay back 40 years of technical debt. It's also quite good at finding security holes in its own code. But also it still has a tendency to produce slop and vastly overengineer solutions to simple problems.

      OpenAI on november 4th wrote a story

      OpenAI say lots of things. Jut because they say it doesn't mean it's definitely not true, but also doesn't mean it's true either. They will definitely round up "done by a team of security experts with million of token credits and AI tooling with unlimited compute" to "AI did it".

      When used and validated correctly, pipelines like this can be used to add features, fix vulnerabilities, find new optimizations,

      I didn't find it very good at optimizations, so far. It's... OK. Well, probably better than most programmers because most programmers are allergic to profilers and will hyperfocus on std::unordered_map because they know it's slow, rewrite it and assume their work is done. AI agents will use the tools and find the hotspots and will make something a bit faster. But they don't do a great job on the whole I've found. At some point it's better to stop the agent churning for hours and repeated re-prompts and simply write the faster version by hand.

      • Re:AI works, really well (Score:4, Interesting)

        by ArmoredDragon ( 3450605 ) on Sunday November 08, 2026 @05:18AM (#66415088)

        I've been setting up agents for scanning for security hole at work. It's pretty good, and as someone pointed out recently AI is making us pay back 40 years of technical debt. It's also quite good at finding security holes in its own code. But also it still has a tendency to produce slop and vastly overengineer solutions to simple problems.

        I've found that if you don't use weakly typed, anti-pattern riddled languages like C++ or python, it's a lot less sloppy. C++ is like trying to build a 6502 out of high-voltage uninsulated wires anyway, and no longer has any compelling use case other than you've been told to maintain code somebody originally started writing 20 years ago, so it's not as if anybody's going to use that crap for anything new unless they just plain suck as a developer. Python still has some compelling use cases, such as quick logic glue, especially for people just learning to code and haven't yet discovered much of the patterns they thought were useful were actually footguns. Either way, these languages are prone to slop with or without AI.

    • Re: (Score:3, Insightful)

      by kertaamo ( 16100 )

      Grey hair here. Stop being agist. It's silly and stupid and the assumption you have made there is not true anyway. Plenty of us are generating code with AI for a long time already. In fact a few of us even created the possibility of doing so.

      Now, about that AI generated libc you would trust. I have to ask why you would trust it?

      Presumably you are not going to review all the C or whatever language code it is compiled from. Even if you tried it may well be so contorted and unfathomable you would not be any wi

  • It's easy to make bold claims for a project that is still on the drawing board. It's much harder to make bold claims for a project that is widely used and has to support the real craziness that exists in the hardware world today.

  • I don't think Alpine had AI before. It's a minimalist system, it often doesn't even have bloat like systemd. AI free is here a label like "0% fat" for winegums. It's certainly true, but misleading about the competition.

  • The same Alpine Linux with a recent *sshole issue where any very long password gives root?

    Well, that issue was human made, no AI required.

Slashdot Top Deals

VMS must die!

Close