Login
You're viewing the stopgenai.org public feed.
  • Oct 14, 2026, 12:31 PM

    jesus christ, Linux has had a 100% LLM takeover.

    this project is completely cooked.

    even after the AI bubble pop, it's not clear how to recover from this level of bullshit from this level of bullshitter.

    lwn.net/SubscriberLink/1099594

    > Linus Torvalds believes that blaming LLMs for the kernel community's problems is wrong, when the problem is LLM users who do not understand how to use them or how the kernel works. A properly trained LLM is a useful tool, but it takes a lot of time to reach that point.
    > Torvalds and others repeatedly said low-quality contributions from new developers should receive low priority, and that the community should stop trying to make contribution easy. Many developers said, repeatedly, that trying to mentor those developers is almost always a waste of time; they do not listen.
    > There is some concern about companies maintaining private kernels with LLM help, but there does not seem to be much to be done about it.
    > Kernel developers should not be required to use LLMs to work in the community, but those who do need to have enough resources available to generate acceptable results.
    The requirement to add Assisted-by tags to patches created with LLMs will be removed, but maintainers still want to know about how a patch was created.

    so how's it going already?

    > The development community is struggling to keep up with the flood of security patches, and will continue to for some time. Some of the most urgent fixes may bypass maintainers and go straight to Torvalds. Maintainers were cautioned against telling their employers about vulnerabilities before public disclosure happens.

    Really, I'm not out to suck as much ass as Microsoft. That will just be a completely unintentional side effect.

    💬 28🔄 31⭐ 40

Replies

  • 💬 4🔄 0⭐ 10
  • 💬 0🔄 1⭐ 3
  • 💬 0🔄 1⭐ 0
  • 💬 3🔄 8⭐ 10
  • 💬 0🔄 0⭐ 2
  • 💬 0🔄 0⭐ 0
  • Oct 14, 2026, 7:11 PM

    @Rob_T_Firefly @kawa @davidgerard

    Yes, but judging by community complaints Microsoft, Google, and Apple have worked hard to bring their end users a new Deskslop experience too.

    So if we want something that works, we just need to use software that predates 2024. Until we get hacked.

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:17 PM

    @kawa @davidgerard approaching Skype levels of lining up the kick for years and then showing up too drunk and flinging a shoe into the sky instead

    💬 1🔄 1⭐ 14
  • 💬 0🔄 0⭐ 0
  • Oct 14, 2026, 12:35 PM

    @davidgerard NetBSD is feeling really nice these days.

    Yes, the Nouveau is too old to recognize my graphics card, and the audio doesn't work for reasons I'm yet to investigage. It's not a setup everyone could accept. But for my needs, it's a relief to have a computer in my house that isn't subject to all this.

    💬 1🔄 1⭐ 11
  • 💬 0🔄 1⭐ 2
  • 💬 0🔄 0⭐ 1
  • Oct 14, 2026, 12:47 PM

    @davidgerard "the community should stop trying to make contribution easy." is an incredible statement. Yes, everyone was spoiled by the extremely friendly and welcoming kernel submission experience, jumping through 900 flaming hoops on the LKML to get screamed at.

    💬 2🔄 3⭐ 3
  • Oct 14, 2026, 1:44 PM

    @moralrecordings @davidgerard real strong “old men yell at those goddamn kids to get off their lawn” vibes from that whole thing.

    Open source projects do definitely need to think about succession and passing projects on to the next generation, but I’m pretty sure “fuck you all, piss off” isn’t going to go down in history as one of the better plans for this.

    💬 0🔄 2⭐ 10
  • 💬 0🔄 0⭐ 0
  • Oct 14, 2026, 8:01 PM

    1995: Golly I sure do admire this #Torvalds fellow

    2001: Oh

    2007: Ohhh

    2012: Dear me

    2018: So Linus is ... apologizing? And ... taking a leave of absence? In order to get ... therapy? Wow I genuinely admire him again!

    2020: I mean he could have just told us all to kiss his ass instead of going to therapy. And he still could at any time just say "everybody fuck off, you can fork it!" Except that's obviously totally impossible

    2026: Welp, good to know some things never really change

    @mirabilos @moralrecordings @wordshaper @davidgerard

    💬 0🔄 5⭐ 0
  • 💬 1🔄 1⭐ 0
  • 💬 1🔄 1⭐ 0
  • 💬 1🔄 0⭐ 0
  • Oct 15, 2026, 2:03 AM
    We had just started getting good Linux compatability for video games! 😭
    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 0🔄 1⭐ 0
  • Oct 14, 2026, 2:37 PM

    @davidgerard The Slopocalypse is going to result in generational damage. Like, "we forgot how to make concrete for 500 years*" damage.

    *yes I know that's not really a thing.

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 2:37 PM

    @davidgerard

    Linus Torvalds believes that blaming LLMs for the kernel community’s problems is wrong, when the problem is LLM users who do not understand how to use them or how the kernel works.

    I once told someone “it’s not that I dislike leaf blowers, it’s that I don’t think they should be gas-powered, people are usually using them to export a problem to somebody else instead of solving it, and people are constantly using them for weird things like drying out puddles instead of moving leaves into nice collectible piles.”

    My interlocutor responded: “Okay, so you dislike leaf blowers.” And she was right.

    I think about that conversation a lot.

    💬 0🔄 1⭐ 0
  • Oct 14, 2026, 2:40 PM

    @davidgerard this feels like they finally get to the point where the old boys won't need to onboard new developers, which is a perfect strategy for the next 3-8 years, and afterwards the old devs will literally start dying.

    and thus the field for a novel operating system opens.

    💬 1🔄 2⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:06 PM

    @davidgerard “the community should stop trying to make contribution easy.” Oh, were they trying? I couldn’t tell.

    So then, as a matter of policy, it is difficult to contribute as an attempt at discouraging low-effort patches. And they also do not want to teach new people how to contribute. But spam bots don’t get discouraged, so they’re still around. So now they complain about the patch spam and blame new contributors but see no problem with the system. Fascinating.

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:08 PM

    @davidgerard
    That entire discussion was one big long category error. A #LLM does not "write code" just like it does not "write a poem". It creates a MODEL of code, or a MODEL of a poem, which a human being then decides to USE as code or USE as a poem. Just like a urinal wasn't a piece of art until Marcel Duchamp displayed it as art, signed by R. Mutt 1917.
    People who "don't know how to use it" are confused by the difference between a map and the actual landscape.

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Oct 14, 2026, 6:47 PM

    @davidgerard @jaypatelani I've had experience running #NetBSD since about 2005, but of the current devices I own, none are going to have netbsd work very well on them. I suppose I could apply the 'I don't want AI to win, I just need Microsoft,Google,Apple, etc, to lose' approach.

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 6:54 PM

    @dougbarry @jaypatelani in the video i'm writing right now, i say Linux is least-worst, NetBSD is anti-slop but doesn't yet have the hardware support, i'm holding on until debian 13 EOL in 2028 and we'll see where we are

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:13 PM

    @davidgerard "the problem is not in LLM but in how people uses LLM" from tech people gives me the same vibes as "the problem is not in communism but how people implemented communism" from soviet people. you're doing jihad extremely wrong kinda argument !

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:24 PM

    @davidgerard Unfortunately not surprised. This was deemed to happen since Linus said the "it's just a tool" (and he keeps saying it **every damn time**).

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:31 PM

    @davidgerard : to summarize : "LLMs are useful if you know how to use them. We allow LLMs. In an unrelated note, we also have a problem of new contributors not being good enough and not willing to learn."

    How to explain your problem and its solution you don’t want to see in the same sentence…

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 3:39 PM

    @davidgerard I hate to even respond to this post, but it's just so misguided that I can't in good conscience ignore it. You are cherry-picking partial quotes and taking them out of context in what appears to be a deliberate effort to discredit what are actually carefully considered and nuanced positions. As you're characterizing things, you'd think the kernel maintainers have just opened the LLM floodgates, when in fact they very much have not, and slop submissions are not accepted.

    💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Oct 14, 2026, 6:50 PM

    @Uair at the moment i'm holding on to see how things are when debian 13 is EOL. but there are no good options at all right now. NetBSD is anti-chatbot, but doesn't have the broad hardware support.

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • Oct 14, 2026, 6:55 PM

    @FediThing these guys literally think chatbots are here to stay, despite the financial doomsday all the AI vendors are heading for very shortly

    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 7:17 PM

    @davidgerard

    But, he said, it took six months to set up the tool to not produce bad responses. If you are a competent developer, he said, an LLM makes a great tool, but only after that effort has been put in to set it up. He resisted suggestions to post his LLM-training materials, saying that each developer needs to go through that process separately.

    Linus has a sweet setup that solves all the flaws of LLMs. No you can't see it.

    The setup: Claude.md: "Never guess. Don't make mistakes."

    💬 0🔄 3⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Oct 14, 2026, 11:13 PM

    @cyberquixote @LabSpokane @davidgerard

    There's Rust (which unfortunately is a LLM positive programming language project).

    And then there's Zig, which last time I checked has a strict anti-LLM stance. Zig doesn't offer the same safety guarantees and compile-time validation that Rust does. What it is, is essentially "The C programming language done right". It minimizes the cases that are UB provides constructs the make things easy, that are hard in C.

    Writing kernels in Zig is quite pleasant.

    💬 1🔄 0⭐ 0
  • 💬 1🔄 0⭐ 0
  • Oct 15, 2026, 2:33 AM
    I'd like to like Zig. Really the challenge is cross platform self hosting. They were relying on LLVM for that, which is kinda sucky.
    💬 0🔄 0⭐ 0
  • Oct 14, 2026, 9:50 PM

    @davidgerard Maybe this is just me saying "but a dog can't play basketball" but how you can allow LLM contributions to an open-source project? The foundation of open source is proper credit, or, to look at it in a different light, a chain of evidence. A computer cannot be held accountable, therefore, a computer must never fork a software.

    💬 0🔄 0⭐ 0
  • 💬 0🔄 0⭐ 0