Been a long time Debian user but with the recent AI decision I am planning on moving on so anyone have an anti-AI recommendations? I understand its being used in the kernel as well but are there any distros taking a full anti-AI approach to the parts of the code they control?
Since folks seem a little confused on what I am referring to here is the mailing list page regarding the use of gen ai without disclosure vote: https://lists.debian.org/debian-vote/2026/08/msg00360.html
Phoronix article on it: https://www.phoronix.com/news/Debian-Votes-Responsible-AI-Use
I’m not sure why you think Debian is embracing AI
They took a nuteral stance on it
I wouldn’t call that neutral. Allowing it with no disclosure sounds like embracing but trying to hide it.
Given Debian’s slow development and focus on stability above all else I doubt this will make any practical difference to contributions that actually end up in releases. And honestly at this point every OS or piece of software needs to have AI reviewing its codebase for vulnerabilities, even if it’s not making any code contributions.
Gentoo?
For the people who really don’t want AI in their OS, check this.
Ooh, that’s one I hadn’t seen. Thanks for sharing!
Arch, Void, Gentoo, Chimera and Aeryn all have anti-AI policies.
Do you know where I can find the Arch policy?
I must have misremembered it because I also can’t find an official AI policy.
But the devs have banned LLMs in their issue tracker, at the very least:- https://gitlab.archlinux.org/archlinux/packaging/packages/onnxruntime/-/work_items/18#note_460864
- https://gitlab.archlinux.org/archlinux/packaging/packages/mkinitcpio-systemd-tool/-/work_items/3#note_503832
- https://gitlab.archlinux.org/archlinux/packaging/packages/python-huggingface-hub/-/work_items/43#note_504464
I am probably gonna ditch linux all together because the creator has been infected by genAI ( used LLM to heko debug, which is a big sin in my codebook ), so I have been trying NetBSD. I got it installed last night, but now I need to figure out how to get xfce4 to be default desktop environment I boot into.
Someone else listed a no-ai OS website, so I am gonna look into some of the other BSDs there as well.
You may as well abandon life altogether, because life has been “infected” with reality.
And reality is shapes by our actions. We work hard enough and we may and genAI forever best case scenario. Worst I’ll take is societal values equating genAI with CP levels of degeneracy and depravity.
You’re deluding yourself, friend, and will only cause yourself more pain in the long run than if you just accepted reality now. Eradicating AI is as likely as eradicating computers themselves. I mean, I’m not going to delete the dozens of models off my devices, and I’m just one person.
Accept that innovation is here, and find a winnable battle to focus on.
I believe Fedora requires disclosure on the use of LLMs in contributions, and that fully generated submissions are not allowed, so slightly better than Debian in this respect (https://docs.fedoraproject.org/council/policy/ai-contribution-policy/)
Linux Mint is always a nice option, but I can’t find an LLM policy for them, but given that they are removing snaps and other Ubuntu bits that nobody likes by default, I would imagine they are more resistant on LLM contributions. They don’t have a policy set in place for this though.
Arch Linux (and derivatives, I currently use EndeavourOS but many like CachyOS too!) are awesome too if you are alright with rolling release. Again, no LLM policy I can find (I didn’t look very hard though, I have better things to do), but I imagine they would be more resistant to it.
Mint is still downstream of Debian though, which will make them taking a stance against it pretty tough.
i wouldn’t say they’re “embracing AI” or ‘sloppified’… but anyway.
i think slackware isn’t using llm stuff for development. it’s my daily/fav distro. but it’s quite different from debian.
What about commits by Russians? Or Israelis? Or MAGAts? Or Libertarians? Or Climate change deniers?
What you are doing is not okay. You’re deliberately engaging in divisive action for no logical or beneficial reason.
They don’t care. They’re hypocrites, filled with irrational hatred. Card carrying members of the “AI Hate Boner” club and that’s all they care about. Their hatred for this narrow facet of technology gives their lives meaning.
Oh my god, you cannot be serious.
Get off the cross; we need the wood.
You may not have noticed, but this is literally a thread about people who want to change operating systems because of the tools the creators use to develop it.
That’s like wanting to move out of your house because you found out that the builders used Ryobi or DeWalt power tools, instead of defunct manual tools by ‘Spear & Jackson’ or ‘Klein Tools,’ because in your mental illness, you object to the power tools that are taking jobs from human hands, completely ignoring the fact that there is still a human in the process, operating the tool. Or you argue that the power tools ‘stole’ all the knowledge of how to be used by the manual tools that came before them. Or whatever hilarious excuse you’ve made up to justify boycotting innovation.
It’s lunacy.
Nah, that’s a bad analogy.
It’s more like wanting to move out of your house because you found out that the builders were in fact subcontracting out their work to unaccredited randos off the street who made a bunch of building code violations and don’t truly understand how to build structurally sound buildings. It’s not about power tools. It’s about the competence of the builder themselves.
And you’re also handwaving away the very real ethics and licensing issues that stem from using models that were trained by scraping all of GitHub without any consideration for project licenses whatsoever, in addition to potential legal issues down the road, stemming from said usage.
It’s not lunacy. It’s a real thing.
Your analogy doesn’t track because there’s no one new being introduced into the process, just the same people using newer tools.
There’s zero ethics and licensing issues, in my humble opinion, any more than there are ethics/licensing issues when a builder watches documentaries on public broadcasting to learn how to build. The knowledge is ingested and incorporated, not copied and redistributed. However, I’ve argued with enough anti-AI zealots to know I’ll never convince you otherwise, so I won’t even try.
It is very real lunacy. Time will tell, I’m just trying to save people the heartache ahead of time. It’s likely a lost cause, but at least I tried.
@Epp @gravitas_deficiency Oh, you’re being an absolute prick to everyone in order to help them? How selfless of you, surely no one appreciates the sacrifices you make to help others.
Thanks, Scott, I’m glad you recognize my contribution. Your opinion, specifically, means a lot to me.
Edit: you’ve had your account for four years, you’ve made a grand total of three comments, and one of them was to chastise me? A third of your Lemmy existence has been devoted to me… I’m touched.
There’s this list of distros that actively reject AI/LLM contributions.
There’s also this list with a few more not listed in the first.
Also found that AntiX rejects AI usage based on a conversation about it in their forum, but they are based on Debian, so…
Appreciate the list! Definitely going to be grabbing as many programs who actively reject AI as I can. Gotta support those who are keeping up the good fight! Thanks again ProdigalFrog!
Wait until they find out AI is used in the Linux kernel.
Asahi Linux
Oh, that’s nice
I’d say wait and see what they actually do. People are super trigger happy with shitstorms these days. But maybe wait for them to actually do something bad?!
LLMs were created with the mass theft of others work by corporations that are now reselling that same information while making billions off it. Thats enough for me to want to step back from any project utilizing these tools.
making billions off it
Lol, no its all circular investing and will soon fail to prop itself up. Then all that stolen work gets sent to open weights for everyone to use. Still not the best, but you can’t close Pandora’s box.
The 2008 financial crisis was the largest upwards transfer of wealth in US history. There are people who still haven’t financially recovered what they lost to this day, but the US economy had regained more than it lost within 2 years. It’s just that 90% of it went to those who lost the least.
With AI being forced into general stocks and federal bailouts practically guaranteed, it’s going to be the exact same scenario. The AI companies will be “too big to fail” and we’ll all be footing the bill.
Theres a ceiling to this shit though, and we are reaching it. Your population won’t carry on when they can’t eat, and it doesnt matter how militarized you are the US is too big to protect everything by force. You have to keep your population chill.
Still not the best, but you can’t close Pandora’s box.
Then mandate local ai only. Fuck big ai company with an anchor.
No one is gonna do that when they are on the cusp of eternal surveillance
We talk about opensource coding project what has to do with surveillance?
I mean if want to have “ethic” ai use ban any non local ai from use.
mandate local AI only
You said this. No one with the kingdom keys is going to do this because the actual purpose is surveillance.
Yeah but has nothing to do with context.
Just like with every financial crisis, when the bubble bursts they will get absurd amounts of money to save their unethical shite companies with money from all of us (through taxes). The rich only ever get richer.
When the bubble bursts, do you think Altman and Amodei won’t be richer than when they started?
I think there will be two new piñatas in the water wars of 2035
I know we’re talking about corps, but damn, humans as water piñatas is so mad max
Its moreso their bunkers will be filled to the brim with supplies. I dont think we have that human-water converter we saw in Dune Part 2 just yet.
It’s wild that Amazon alone has already pissed away more on their own than even Nvidia has earned back. I’m also surprised that Anthropic still isn’t remotely profitable given the sheer volume of money the rest of the corporate software world dumped into Claude Code and how much pressure Anthropic applied to monetize it. That alone should be enough to show that it is impossible for AI to be profitable in any traditional sense.
Their actual plan is to monopolize information and basically use (preferably their own) AI as the middleman and broker for all data and information and knowledge and advertising that any human being ever sees. In that, they might eventually be successful. At least if they don’t run out of money trying (almost certain), and we don’t kill them off or blow them up first.
The meme about Nvidia selling pickaxes to gold diggers really is accurate
That does seem to be their strategy, at least on its face. Although I’m not even sure I even really believe that is profitable, or if they’re cooking the books with such incestuous partnership agreements that it’s to the point where Nvidia might be effectively “buying” chips from themselves with their own money on their partner’s behalf. I guess time will tell if they’re actually the next Enron or not, but I don’t trust any business that’s AI related at this point.
God damn love to see it
Yeah. But my point is: They’re not utilizing those tools. At least not yet.
Looking back at my Linux experience, I have a lot of trust in the Debian community. Including to do the right thing. Even if it’s difficult. They don’t always get everything right. But they’re pretty alright. I wouldn’t just throw it all in the bin because of some hypotheticals. And democracy isn’t easy. We judge people by what they do. (At least I do.)
I understand they may not be utilizing them yet, but the vote was approval of the tools without disclosure. So this isn’t hypothetical, they are going to begin using them and therefore it makes sense for those with concerns regarding the technology to start looking for alternatives. It doesn’t require them to start adding AI chat bots to Debian or anything like that for me to make this move. The tool itself is the problem.
That remains to be seen… As I said, they earned my trust over many years. Maybe they’re able to tackle it, as they faced other challenges. I mean you’re free to throw away democracy on a whim. I think it needs to be addressed and tackled. Votings have turned out with bad results before. It’s ugly. But I’m still subscribed to the ideology. My red line is something bad actually happening not just on paper and in some hypothetical future. Until then, there’s plenty time to repent. I still have my hopes up. They’re good people and have been doing a phenomenal job until now. I don’t think I want to ignore that. But I’m not stopping you from having some alternative plan ready. 😅 That’s not a bad idea either. All I’m saying is: Don’t borrow trouble. (And “Don’t throw the baby out with the bathwater.” Democracy might indeed be a good thing, despite it’s downsides.)
The tools used dont change the core principles which in Debians case is the social contract.
Slop is never going to get through any serious quality control.
If it comes, it comes. Will be a shame. But until then I will happily keep using it.
There are plenty (and I mean plenty) of free open-source AI, including some with open datasets that are not stolen.
I really wish people wouldn’t use “AI is theft” as a blank term because the technology itself is not inherently bad. We can hate OpenAI and still run local models.
By the way, there’s like two open-source LLMs, which are large enough to be usable in some form. And one of them claims not to contain pirated training data. It’s pretty much the opposite of “plenty”.
Which ones are those two?
Yeah, good question! I bet nobody even knows their names. I had hoped to get a reply by @[email protected] or @[email protected] who seem to know plenty of them?!
I was talking about the RedPajama project who recreated the dataset of Meta’s LLaMA and proceedingly trained a model on it. And Apertus by the ETH Zürich. They also factored in ethics when compiling the dataset.
I made a post asking on !localllama: https://programming.dev/post/55905594
In case anyone knows about the existence of such models, please educate me over there!One more than i knew thank.
If you believe all that then you need to stop using linux all together. Linus sees AI tools as just tools so they can be used in the development if the Linux kernel. So Linux is a project using them. You must now stop using Linux.
Mentioned this in the description of the post already.
All you said is that you know they use it. You didn’t answer why it’s ok there? Also, Linus didn’t put it up to a vote.
Never said it was okay there. Purpose of this post is to locate a linux distro that isn’t using AI with what is under its own control. Can’t control the kernel but I can look for distros that share similar ethics.
Correct.
They just recently did something bad.
What did they do? All I saw was them talk and do democracy. All of that is allowed from my side. I don’t think they incorporated any AI stuff (yet). At least I’m not aware of anything?!
They decided to take a vote on whether or not to “Ban LLM contributions” where the proposed policy wasn’t all that well thought out, and ended up settling on a subsequent proposal which was even worse, enshrining in official policy the standard that LLM contributions are just fine actually and don’t even need to be disclosed. The governance system, in this case, failed spectacularly. The contributors they will lose as a result of this are in my view likely to be more valuable ones than those they’d have lost had the vote gone the opposite way. It’s substantially worse than having done nothing at all. It does not inspire confidence in the future of Debian, and is enough to have many of us looking around to see what alternatives might exist.
@kbal @WorkingPie @hendrik Banning LLM contributions would be, at best, a symbolic gesture about a heavily politicized topic. For a team like debian, the concern should be about the quality of the contribution not what tooling was used.
Yes. And it goes further than that: Maintainability should also be a concern to maintainers. Doing good in the world is what pretty much all Free Software community projects are trying to achieve. They have to deal with licensing issues… It’s complicated.
I think democracy is a good thing. Full stop.
It ain’t easy though. Not even for the Debian community. But what’s the alternative? Some corpo distro? They’re going to adopt AI as well. And benevolent dictator might be alright… Yet, I personally still prefer a democracy. Even if it’s difficult. Sometimes voting results are sub par. But we need to deal with that, not drop it once the first difficulty arises?! What kind of strategy is that? I don’t think anyone is getting anywhere if they just keep dropping everything once they’re faced with problems.
Distro hopping was a hobby before people had reasons to distro hop. If þey have time and some personal motivation to look around, why discourage þem? It’s not as if having a few fewer users is going to hurt Debian; on þe contrary, þe justification for embracing AI had been “we’re overworked.” Risking AI troubles is fixing þe wrong problem; instead, encourage users to be a drain on some oþer distro’s maintainers!
I recommend:
No AI there.
This is the only answer.
What about
?
No, that harms the environment somehow, I’m sure of it.
Oh! I know. It’s “fascist technology” because Texas Instruments is owned by rich people.
Texas Instruments make weapon system.
There you go. Stick with the abacus. It’s the only way to be moral in his world. Of course, that’s assuming you don’t have a problem with the trees that died in its creation. To be WHOLLY moral, you should use rocks on the ground, in the outline of an abacus you made in the sand with your finger. And HOPEFULLY you didn’t disturb the micro ecosystem of the sand when you drew it.
It does use plastic. 🤔
By jove, you’re right! 😱 Clearly it’s fascist technology.
Better stick with the abacus.
Maybe this is the push you need to . . . Yes, maybe it’s time for . . . FreeBSD.
Search your feelings. You know it to be true.
I’ve been tempted to give it a go a few times. Mainly for fun tbh. Appreciate the recommendation!
I believe that there is now LLM generated code in the Linux kernel, so BSD is probably going to be where this path leads you. Solaris or AIX maybe ;p
FreeBSD has been slop for a while, the only slopfree BSD remaining is NetBSD, but that’s probably not practical to actually use.
What about openBSD? I doubt the most secure OS would want llm code.
FreeBSD does not include AI features in its core operating system, but the FreeBSD community and Foundation have explored AI-assisted tools for specific tasks like vulnerability discovery and driver development.
https://freebsdfoundation.org/blog/freebsd-ai-assisted-vulnerability-discovery-project-launch/
I’ve been meaning to revisit FreeBSD for a while, I just haven’t gotten around to it.
I used FreeBSD extensively from versi9n 3.3 to 5something and barely touched linux back then. Only reason why I moved to linux was the larger community.
Then suddenly Linux became a good desktop OS without much hazzle at all, so it replaced both desktop and server-side use for me.
How is FreeBSD as a desktop OS these days? I vaguely remember compiling xfree86 from /usr/ports some 25 years ago, and dicking around with monitor sync rates to get Gnome to display properly.
To be honest that was sort of my story too - played around 15-20 years ago with it, then Debian (and the rest) Got Good and I just haven’t gone back.
Maybe I was just talking to myself.
Hmm.
In purely utilitarian terms, are there any interesting differences in using freebsd or other bsds instead of linux?
NetBSD is more portable than Linux and supports a lot of CPU architectures. OpenBSD is secure by default. I think the BSD’S have jails, which are better containers, but I do not know much about it.
For desktop daily use, it only matters if you really want that Unix taste.
Thank you!
Well, back in the day the guy who was really into it would bug out his eyes at you and go “Oh YES! There’s . . . “ and he’d immediately list two or three huge differences and then downshift to a bunch of reasons why that’s better™ and then you’d go home and not remember a single one very well.
These days I don’t really know. Is there a little red devil in the house?
I had another look and I think it’s basically down to Gentoo. If you’re in for the BSDs, you got NetBSD as well. And OpenBSD seems to be anti AI-contributions as well. I didn’t really find anything else with a strict No-AI policy.
I’ll take a look at them! AerynOS seems to be the primarily distro I have found with a no AI policy but sadly its atomic. Thanks for the recommendations!
I missed AerynOS. Thanks! Let’s hope you’re going to find another one. Given there’s lots and lots of people who don’t appreciate AI written code… I feel there has to be a market for a mainstream Linux distro with a no AI policy.
Agreed! With how many linux distros we have out there some choices on AI usage would be nice!
Fantastic it is atomic. Very promising future with this independent distro in these times.
Gentoo gang rise up!
There are dozens of us! Dozens!
