@KimSJ @voltagex Non-Linux I guess as the Linux kernel allows for AI contributions.

I am trying NetBSD first and see how I get on. I am under no illusions that it will be harder than continuing with debian and its derivatives.

I have been using debian based OSs since I installed eeebuntu on my first eeepc in 2009, and every device I own currently uses one or another flavour of the debian family.

It will take a while to lose the "sudo apt update" muscle memory.

docs.kernel.org/process/coding-assistants.html

@tschenkel Good question - anyone?

I *was* planning to move to Linux - prob MintLinux - but now I'm not sure as it seems AI, (well, LLMs anyway), is going to turn up embedded in all Linux based OS's.

My problem is that I don't know much about Linux other than the little I learned about OG Unix - my intention was to learn as I used it.

Now, it feels like checkmate 😔, so any advice on where I can turn would really be appreciated...please?!!!

@Oyu_Fka
I don't think the problem is that LLMs are going to be embedded in Linux. At least not in the same inescapable way they are in Windows and OSX - they will be available, but mostly not installed by default, and if they are (some distros may include them, but Mint likely won't) they will be ee easy to disable.

The problem in this thread is that many Linux distros had "no AI code in core elements" policies, which are now eroded.

@tschenkel Thanks for the reassurance! Tbh, it's early days yet...I have to buy a second hand laptop first, lol!

I'll carry on reading up on Mint in that case - it was the fact, as you say, that AI is being allowed into the cores of some distributions that bothered me, particularly with open source products...🤔maybe I'm over thinking, (again), lol!

@Oyu_Fka
If you want to get away from forced LLM integrations, Linux is a good start. You'll notice a massive change from Windows or even OSX.

Mint is a good distro to start with.

@tschenkel My initial interest in moving to Mint, (or Linux generally), was privacy, but I grew concerned due to the increasing integration of LLMs which could compromise that, either through poor coding/design, or worse still, deliberately in the way google operates...or even other actors with even more nefarious purposes.

Anywayz, I'll put my cynicism away and get back to Mint 👍!

@ohmu @tschenkel Thank you - windowsXP was the last OS I used in any meaningful way, lol, which is encouraging 👍!

Yes, I've decided to try my luck with Mint - I could do with a challenge, lol!

@fschaap
bsd doesn’t support a host of hardware drivers, we cannot run away from this. maybe in the short term but we cannot keep hopping distros and systems to avoid this. I don’t want to say the ai sloppers won but at least for now our entire field is under slop occupation unless we actually do something about it.

hooligans don’t just get bored and leave, especially if you don’t start forcing them out..

@b3lt3r
NetBSD is not really harder than Linux when you know how to use the basic system (shell, vi, emacs, etc).

And your notice regarding the Linux Kernel and genAI can also be transferred to FreeBSD. This is also mainly a kernel only, most of the programs use nearly identical source code as used on Linux.

So to my example on the other toot, just imagine Wayland or X11 having genAI inside, would mean you could not use any OS for a workstation anymore.

@feff Yes, and perhaps I should have been "on this" sooner.

I could do nothing, adopt a watching brief, and pretend that that makes me neutral on the topic.

And doing nothing is definitely superficially attractive, because the doing something meaningful is a huge amount of work.

I'm pretty sure that even our doorbell runs Linux.

@freshstart

> And that would be responsible use AI, right?

What criteria are you using for this assessment? That would seem to be important here, as to whether your conclusion is really just reinforcing your choice to use genAI because you would find it helpful to do so, or if it is a considered, objective stance.

@freshstart
I would think that just based on license concerns alone would necessitate a policy of "No source-code or binary files may be generated wholly or in part by an LLM."

If _you_ didn't write it, or if you don't have a license to relicense it, you can't in good order publish it under an open-source license. (Or, for that matter, a closed-source license.)

@freshstart Why would
it be ‘hypocritical’ as a developer? Also as a developer, I don’t find it at all unreasonable to reject all those uses. (Personally I either find them unreliable, or at best not useful enough to justify the harms.)

So it seems like it’s about personal ethics and tradeoffs rather than any objective evaluation (for both of us).

@neil

@freshstart I wouldn't say yes to that. The possible utility of AI doesn't change the underlying ethics or environmental costs.

This reasoning is exactly what leads to shit like Operation Paperclip. It's okay if we let these people get away with their horrifying crimes as long as they help us with rockets, right?

Right?

You might not like it, but ignoring ethics because the tool is useful is this exact reasoning. Is it on the same scale? I dunno, maybe not yet, but it's the pure utilitarianism path, so, you know… be careful what you're stepping in.

@growse

I think that I see that in the same way as I see other questions of provenance.

For instance, if I promise that I have the permission to contribute a patch, but I do not, that might be hard to detect, but is nevertheless important.

I agree with that framing.

The provenance thing provokes the same question for me about where the line is though.

A hidden actor that writes the code directly that I claim? Probably against the policy.

Someone/thing that coaches me through writing the code? Someone who writes a blog/book that inspires me to create the code? Less clear.

What about a machine that takes the voice output of someone unable to type and turns that into code? Is that acceptable?

I don't think "no LLMs / no AI" is anything like precise enough to be useful. It's a good statement of principle, but a wholly inneffective policy imo. I have no good answer or suggestion though :(

@growse

I don't think that most of these are new issues, though - and some seem (to me, anyway) somewhat contrived/forced in the context of the issue in discussion (speech to text in particular).

That said, yes, clarity is welcome.

take your point on contrived. My experience is that the credibility of these things live and die with the edge cases. Today's hypothetical is tomorrow's case study etc. etc.

We will discover and figure this out over time, I'm sure of it! 🙂

@davidgerard i disagree only with "at best". there is a _chance_ it will be used to block some or all slop code on grounds which a techboy who thinks they are "non–political" could not argue with.

on the other hand, this is broadly the same policy as adopted by the kernel. ***the kernel***.

if i want to be absolutist about this, time to switch to Haiku. i can't do that.

edit: i certainly would think well of anyone who can. i can't.

@loogpy

"just" does a lot of heavy lifting here.

@Dragon The BSDs haven’t yet allowed LLM use for core contributions, but several accept it in upstream packages. E.g, OpenZFS has some commits “co-authored by” Claude, and FreeBSD includes OpenZFS. It’s not FreeBSD code, but the point of the BSDs is they ship a cohesive system, so it’s fair to assign them responsibility for what they ship.

OpenBSD’s stance is purely about code quality, and they haven’t seen LLM-generated code which meets their bar. They don’t tell people not to try, though.

NetBSD has a strong anti-LLM stance centered on the fact LLM-generated code can’t be BSD-licensed.

@AlienJay it's not Linux, but #NetBSD has a policy forbidding what they call "tainted code": netbsd.org/developers/commit-guidelines.html

#Gentoo has a similar position.

Both allow individual exceptions to be made after careful review of all concerns.