"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
„The guy who runs the kernel couldn’t get his head around Python.“ What is that supposed to mean? Do they want to make the point, that Linus Torvalds is too stupid to learn python and therefore uses AI to do it? Because independent of one might think about the use of AI in the kernel, I don’t think that this is the case…
I think there are way better arguments against the use of AI in the kernel than personal attacks against Thorvalds.
AI imo is good at finding potential issues or places to optimize but it shouldn’t be writing code on the kernel. It’s up to the human to investigate and adjust, not AI slop.
We just need more volunteers? Isn’t it just that the development is on such few people?
That is one of many issues.
Put a pin in socio-economic-environmental issues for a second.
AI is good at finding bugs/vulnerabilities. If you're using it for peer programming, the output is 100% what you're comfortable with, while you're still directing what goes where and how. That's right up Torvald's alley. Not every piece of code we write is a masterpiece of complicated art. Sometimes, when we are writing a complicated piece, we are unable to see all the ramifications of each piece in a way that AI appears to be able to see.
From this standpoint, even if we had a LOT more people working on the kernel. We'd amplify problems, especially if the people we threw on the project were more free-hands than artists.
Unpin that socio-economic-env stuff now.
Until tokens cost enough that it's not reasonable, or the AI providers make things dumb enough to save money/meet law/eco demands, most code at most places will likely go through at least a post-cleanup/analysis phase. Everything would be labeled slop.
I don't think, long term, that we're going to have AI writing everything at the level we are even now, but it's not impossible. That's not to say I advocate it or push it forward, but there may simply be no way to stop it if the economics / laws allow it to continue.
Coming in here and seeing all the bans...
I know there are two sides to every story. So one shouldn't just judge from the number of bans, comments upvote ratio, and definitely not by also reading the modlog reasoning...
It's like half the thread belongs in one of the power_tripping_bastard/mod_abuse communities
Linus Torvalds has been AI curious for a while. He went on Linus Sebastian’s YouTube channel, Linus Tech Tips, in November last year. He was clearly well into the chatbot already: [YouTube]
That’s reality. Deal with it. That genie is out of the bottle. You’re not getting it back. And you’re not getting it back whether you are a photographer who’s out of work. Those things are gone, right? Because you can fake pictures so much better now. Or you’re a programmer that has to learn to deal with a new reality.
I think actually on the programming side, we’re in a fairly good spot. I really think that AI will be a tool and it will make people more productive. I think that vibe coding is great for getting into programming. I think it’s going to be a horrible thing to maintain.
Uhh...
ITT: Hyperbole
I mean, we're in the middle of a cold civil war across the world right now, with fascism and puritanism on the STRONG rise. And that's not hyperbole.
I'm currently looking into forking the Linux kernel. Worst thing that can happen is that it fails, or I have to change professions.
Maybe I’ll give BSD another go
Bummer. I was hoping this was a case where someone created their own open-source model, trained on their own work and not Internet scraping or destroying books or whatever, and did the thing on their own local infrastructure. There's nothing unethical about that.
Seems pretty clear from his statements that he's just slop 'coding' with frontier models.
Apparently we can all look forward to things that always worked to suddenly be buggy as hell, just as it is now with Windows.
Even if you train your own model, its still worse than a human doing human work every single time and runs huge risks to a project like this.
had all this happened 2 years ago i don’t think i would’ve bothered switching away from windows tbh
i wonder how many people now are thinking the same thing…
We're now thinking of switching to BSD. But I'm hoping enough kernel maintainers and others will move to a fork of the last pre-LLM-touched kernel(s) ASAP.
I don't mind it IF the programmer using it does it while understanding what is churned out and how it works, ergo, if he is skilled enough to use his AI only as a tool to save time.
If he has no expertise and the code is completely unverified and uncontrolled, then it's slop.