Discussion
My current goal (in life) is to have a computer system, whether it's my primary system or not, which contains no "AI code assistant" generated software whatsoever. If I can't have that I want to get as close as possible. My two paths are (1) Make a Redox+Servo system then start attacking Rust dependencies, and (2) "something with the BSDs?". Several BSDs are candidates.
This one has, as far as I know, the *clearest* maintainer anti-AI statement; maybe it's the new base.
@syllopsium "it's unlikely to be viable to keep it out of browsers."
servo has a no-ai-contributions policy. that's half the reason i'm talking about it tonight
@syllopsium and if amd becomes a problem i will create. my own. GPU
@mcc I have a problem with Alpine’s AI stance: in practice it’s no stranger than the NetBSD foundation “no tainted code, and we define AI generated to be tainted.”
I think it’s fine to pick Alpine because you want a Linux userland over a NetBSD kernel but I don’t think that that they realize how hard a job they have signed up for.
The foundation will apply the rule to the entire NetBSD userland but Alpine will have to evaluate dozens of suppliers and hope not to miss slop.
@mcc I’m following along as well, I’d like to get as smooth and uncompromised as workflow as possible… maybe if BSD has legs I’ll try and contribute, but I do love Linux
@cuirass_entertainment i don't even like linux but i believe switching to bsd will mean giving up certain entire classes of activity i currently do with my laptop. and linux was supposed to be my escape hatch for the classes of activity i can currently only do on Windows (which is going away)
@mcc Yeah, I mean BSD will kill my entire workflow but I haven’t had a need for windows in quite some time
@cuirass_entertainment Okay, well, I would like at some point to play a video game, or use a piece of computer hardware I purchased from a store, or run an art program that is not "Krita" (Krita is very nice but it is not intended for all workflows). So I still have my Windows machine as a big hardware dongle for these activities.
(Notes:
- The problem with the BSDs is while several of them really are "better" about AI than Linux, it's usually "by accident", or indirectly through other policies such as copyright— implying if the copyright status of LLM code were clarified they might. I want someone who is *against LLM tech itself*.
- I still think a "long LTS", where Debian is forked right before Claude 4.5 is released and maintained "forever", is a good third option. But this would take a large, coordinated project.)
In general, it's very plain that there are *both* (1) people who expect that the appropriate amount of AI code assistant use by the people who write their software is "literally none", and (2) software engineers who do not want to use AI code assistants in writing their code, but can no longer find jobs willing to employ them to do so. The trick we need to solve is to somehow connect (1) to (2). (Assuming it is still possible for humans to fund humans to do labor. That might not be true anymore)
@mcc I came to this conclusion and the solution was to, well, exit employment. I figure the only way to fulfill this, due to lack of many employers in this regard, is to employ yourself. It’s unlikely to work out, but I’m unwilling to be a collaborator in the multitude of ways tech is ruining everything at the moment. It is encouraging, however, to see so many people also sick of this situation, and not just passively accepting it as the new normal.
@mcc As others have said 1) Isn't actually the important part, the number of people in 1) who will pay for it is.
Don't want to pay for it? Well, you get what you're given then. Either pay, or code it yourself.
@syllopsium i'm already paying for servo, codeberg, and mastodon and i'm about to add redox. will other people pay for better software? i dunno, i'll see what i can talk them into. i bet a lot of people are paying for at least one streaming video service they've been meaning to cancel but haven't got around to yet
@mcc
I’m Rania from Gaza. 💔🇵🇸
The war destroyed our home and left my children and me struggling to survive without the basic necessities.
I’m asking for your help. Even a small donation or a share can help us get through another difficult day. 🙏🤍
🔗 Donation link in bio
@mcc that also presumes that (1) wants that in a way that actually survives contact with the bill. Consumers notoriously want alternatives to Walmart etc right up until they see the ticket price.
@donw i think the difference between software and walmart is people have the option of just not using software, whereas walmart usually sells staples people need to live
@mcc i am available to start on whatever no-AI FLOSS project wants my help as soon as someone offers me a salary for it.
To make matters worse - people lie.
I promise you people use AI and fail to disclose it.
The only way to insist on no AI and be reasonably sure it's true, today, is to write your own. More power to you. I for one support your effort.
But - I'm way too lazy for that. I'll have to judge the code I use the old fashioned way - by reputation and test coverage and audit results.
@tbortels @mcc I think you may have missed that people who are against LLM use are often against it on ethical grounds, not merely because of technical and legal issues. I suspect that collaboration is still possible, but will require paying attention to the ethics and track record of the people responsible. Which, really, was always the case, but there seemed to be fewer people needing active exclusion.
I suspect most people who are anti-AI are actually anti-AI-companies. LLM is just a technology. The very real evils arise from people who unethically create, promote, or profit from AI use - not from the AI itself.
To put it another way - the bad guy in "The Texas Chainsaw Massacre" wasn't the chainsaw, despite being in the title. It's the people using it, both to make money and to do ancillary things like enshittify their phone support.
The real issue I see is that AI use is being forced on people - you call your bank and have to fight the AI instead of their prior shitty voicemail robot. That and the religious - both the "no AI ever no matter what" and the "AI will save the world" camps are deeply deranged.
The reality is it's almost impossible to put the genie back in the bottle. But it *is* possible to work out the failure modes and reign in the abusers.
@tbortels @mathew I don't want to control it because I don't want to use it. Meanwhile if someone else is using it, then I don't want to give them money, and I don't want to use their software. It's like, imagine I don't want to eat human meat. "The food is made with human meat" wouldn't suddenly become okay just because a person other than me cooked the meal.
I have sympathy. It's a pickle - people are doing things you don't approve of and want nothing to do with.
I have the same problem truth be told. For me it's violence and racism. I don't want it and I don't approve and they do it anyway.
I'm not making fun. I'm dead serious. How do we convince people to not do things we think are bad?
I don't have a good answer.
@tbortels @mathew That is a very large question. It seems to me the first thing to do is to accept that the technology is unfit and problematic rather than just going around saying "this is a good thing, we just need to change who owns it". I don't know what the best second thing is but the second thing within my power is to stop rewarding people doing the problematic thing by giving them money, using their platforms, etc. It is that second thing that *broadly gestures upthread* this was about
I don't think I ever suggested ownership was the issue - but I can see how my phrasing might have implied that.
I was asking *if you were in charge* - you're the king, the decider, your will be done - what would you do? And specifically - would you impose that decision on others?
My take here is that the ills that arise from *any* technology - not just AIs - arise from human misuse.
Fire burns and kills people. But - to forbid it entirely would be bad. Chainsaws maim - but they're a powerful tool. And I suspect LLMs have a place. What we need is to stop the abuse.
(As the prices of rent and food go up, and the "% of wealth" numbers start to look stranger [1% of the population owns 50% of the wealth, that sort of thing], if trends continue unbounded eventually you hit a point where the *only* way to get money to live is to do something that makes an already-wealthy person happy. More and more industries change from making products to sell, over to some sort of "storytime" where a billionaire is delighted by the way you dance and gives you "investment".)
@mcc grognard capture but backwards. Napoleon capture
@mcc LOL "eventually"
@mcc Historically, there have always been boundaries to how long that sort of thing can go on.
Pictured: Historical boundaries
@mcc i keep talking about communes like a crazy person bc i would literally rather do hard labor like farming than survive this way. i’m probably too weak and sickly to actually do it successfully but it feels like it should be possible to like. have an economy that’s real
@mcc In that case, the only alternative will make rich people very, very unhappy.
@mcc I mean, isn’t that the principe of VC funding already? And it’s the biggest factor of economic growth right now 😬.
@mcc It's an interesting idea but utterly unsustainable.
Effort would be better spent to make NetBSD an first class citizen, but unfortunately there's a lot of effort required to do that, and at some point it's necessary to engage with the corporate interests driving Linux.
I also think it's unfortunately a losing battle. Even if a certain amount of the base OS is AI free, it's unlikely to be viable to keep it out of browsers.
Even if the attitude of 'some applications are OK' is taken there are other dependencies such as mesa (uses AI now I understand), and I would be amazed if amdgpu is or will stay AI free.