RT if you want a CLEAR statement about AI from all GNU/Linux distributions and are ready to quit any distribution that is ok with integrating AI slopware.

[shaft@piaille.fr]

@dallo @breizh @Khrys I want tools usable by regular human beings with little or no technical knowledge. Not niche stuff for people with 30+ years of sysadmin experience

[paradegrotesque@mastodon.sdf.org]

@breizh @Khrys @shaft

#NetBSD has explicitly banned AI code from being incorporated into the system.

I suspect both #FreeBSD and #OpenBSD will take similar positions, probably for the same reasons (dubious provenance and quality issues, problems with code licensing, etc.).

[dallo@pouet.chapril.org]

@shaft @breizh @Khrys I met people running Hurd/GUIX. They were not people with 30+ years of sysadmin experience. I never met anyone running Plan9 though.

[shaft@piaille.fr]

@ParadeGrotesque @breizh @Khrys The door is not 100% closed in NetBSD

“Code generated by a large language model or similar technology [...] is presumed to be tainted code, and must not be committed without prior written approval by core.”

An approval can be given

https://www.netbsd.org/developers/commit-guidelines.html

[paradegrotesque@mastodon.sdf.org]

@shaft

What you say is true, but I don't see NetBSD core team give an approval of this kind any time soon.

As far as I know, nothing has been approved so far, but I'll welcome information and corrections from the #netbsd developers on the Fediverse (and there are quite a few here).

@breizh @Khrys

[crell@phpc.social]

@Khrys With low level system libraries getting on the AI bandwagon, it may very soon be impossible to have an AI free distro.

We are well and truly screwed.

[shaft@piaille.fr]

@ParadeGrotesque @breizh @Khrys Anyway: you have your AI-Free kernel. Nice. Then all the softwares that will run on this kernel, ranging from the shell to the web browser have to be AI-free

[mathew@universeodon.com]

@dallo @Khrys We can probably assume everything maintained by RedHat is going to include LLM-extruded code, since IBM calls the shots.

[milagemayvary@mstdn.social]

@fmasy @Khrys

AFAIK GNU doesn't, got a source?

[paradegrotesque@mastodon.sdf.org]

@shaft

BSD user land (shell, utilities, etc.) is part of the distribution, and is often created and maintained by the same developers as the kernel.

So no AI there either.

Which means BSDs, in general, are more consistent and have slightly lighter utilities than Linux.

For example: ksh is usually a version of pdksh, and is maintained by several developers from all 3 major BSDs (if I remember well).

@breizh @Khrys

[faithfulljohn@mastodon.scot]

@dallo @Khrys Go Inkscape! @doctormo

[grahamperrin@mastodon.bsd.cafe]

RE: https://mastodon.bsd.cafe/@grahamperrin/116220642823558416

@breizh re: Linux, thank you for providing the missing link to docs.kernel.org.

There are multiple BSD distros.

For FreeBSD, the quoted post might help.

cc @ParadeGrotesque

[csgraves@turtleisland.social]

@dallo @Khrys I just saw elementary OS on the list, but they do use systemD so we will see what happens. If the Wayland session can be refined a bit more I could run it, but I do miss having an actual desktop with icons.

[i@toot.pouyan.net]

@Khrys why? For security concerns? For ethical reasons?

[only_ohm@mas.to]

@milagemayvary @fmasy @Khrys

GNU Guix has its own init system, so wouldn't be immediately scuppered by the contrary position of systemd.

[joacim@mastodon.melin.org]

@Khrys It's Linux. And yes.

[troed@masto.sangberg.se]

@i

'Now we come to the third and final novel AI psychosis, the critics' psychosis, that AI is an abnormally terrible technology. This is a species of "criti-hype," which is when critics repeat the hyped-up claims of the companies they're targeting, but as criticism'

https://pluralistic.net/2026/03/12/normal-technology/#bubble-exceptionalism

@Khrys

[crazyeddie@mastodon.social]

@dallo @Khrys That ship sailed a long time ago. Systemd has always been slop. To the point of actually proposing changes to the kernel command line to hack around bugs in systemd rather than fix them, which earned quite the colorful response from Linus.

But that's what is widely used so ... well, whatever. I figure if that's how you be you may as well use AI. Do the JIRA shuffle on fast forward.

vOv

[fmasy@piaille.fr]

@milagemayvary @Khrys

Source that Linux kernel project may receive AI-assisted contributions ?

I cannot point to an example of such a contribution but at least the documentation indicates that it is now explicitely allowed:

https://www.kernel.org/doc/html/v7.0-rc1/process/coding-assistants.html

(github source:
https://github.com/torvalds/linux/blob/master/Documentation/process/coding-assistants.rst)

[milagemayvary@mstdn.social]

@fmasy

I'm familiar with the contributions to the kernel Linux, but the GNU utilities I don't believe is in the same situation.