While he is correct from a purely technical perspective, a lot of the backlash against other people using AI (LLMs generally) on collaborative work is due to how it is currently being implemented in a way that harms the environment, drives up hardware prices, sucks up mass amounts of data it shouldn’t, breaks laws, negatively impacts those lower on the social latter , and a bunch of stuff outside of how it impacts code.
Like my limited experience with LLMs has been bad because it is frequently wrong, but I hate it because it is shoved into everything and has all the negative impacts I listed above. I wouldn’t hate it and discourage others from using it if it was just new tech that didn’t work.
I find it does well for:
- Reformatting data / text
- Finding new movies or music from a list of your likes
- Finding web info like google used to in 2010 before ad based content over swept all the searches
Other than those its a pain in the ass, and I keep having to disabling it or unistall products to make my windows work machine usable…so much garbage packed into every app now.
Finding web info like google used to in 2010 before ad based content over swept all the searches
LLMs are only useful because Google has deliberately made their search engine worse. Since I pay for Kagi, I don’t need to consult a chatbot anymore, I can just search again like I did before the AI hype.
So we technically don’t need LLMs for search. But we are nudged extremely hard into that direction.
I like the chatbot kagi has in their pay tier but not enough to pay for the biggest tier. Its been useful for searches like “I want the best GPU that will fit in 220mm” but aside from that I’m happy enough with their search to not bother.
Surely you can find the right GPU by just going to a site like pcpartpicker, filtering by length, and sorting by review score.
You’ve never used an LLM against an MCP server with a properly built API. The ad hoc shit you can do is amazing.
Imagine you are managing 1,500 endpoints. A mix of Windows (both clients and servers), Mac, and Linux. Now imagine that one day there’s a CVSS 10.0 in Eclipse Temurin 17 and you suddenly need to know everywhere it’s installed and which precise 17.x version.
You could spend all damn day fussing with SQL -or- you could fire up $LLM and tell it “Give me a report showing all installed instances of Temurin below version 17.3. Sorted by endpoint type and operating system”.
Suddenly want to know how many windows workstations haven’t rebooted in more than 14 days? You can do that and get an answer in not much more time than it takes you to type the question.
Suddenly need to change a setting in 15 different Microsoft Tenancies? Snap, it’s done.
LLMs are NOT magic bullets but they are useful for WAY more than what you listed. You just don’t know that because you don’t do that kind of work.
Do you know who can also do all that? A properly trained ma. Maybe learn your craft mate.
If you start with LLM before knowing your shit, you will be worthless in 10 years.
I don’t care that it can do that, not until they fix all the negatives it has.
Our IT guy does this stuff, but there’s training involving to get it to do that. I’m holding out for Zorin Grid which is supposed to do all that from a singled overview dashboard
I’m pretty sure he already handed out a ban for someone who submitted a slop PR.
He’s been pretty clear that he’s only interested in clean code additions. It doesn’t really matter how you achieved your PR so much so thst it’s quality code with a useful purpose for the kernel.
“AI is a tool” is a thought-terminating cliche. The phrase is intended to foreclose other ways of framing the discussion.
Yeah, it ignores the biggest problems a lot of people have with AI. It’s not the slop code, it’s that it was generated at a ridiculous cost to the environment and the population around the datacenter where it was trained.
Its both low quality code and massive energy costs. Its a loose loose situation. If AI disappeared tomorrow, nothing would change, besides tech bros actually have to do a honest days work.
Techbros NEVER have done a honest days work and NEVER will.
I used to see anti-AI people as myopic and foolish, but seeing how it’s being developed and implemented is really eye-opening. They’re pushing an unfinished, unrefined technology that often adds to the ecological crisis and might be poisoning people’s drinking water, so they can roll the dice in terms of what it’ll do to the economy while using it to spy on everyone. The public face of all this is these companies telling us that we should be terrified of what it’ll do to the job market, but we need to invest in it big time, right now. I apologize to anyone I spoke down to. I get it. It’s not a way forward, but I get it.
The LLMs are actually very refined and perhaps even finished.
It’s just not meant to be general intelligence.
Disaster capitalists can’t be trusted to mix the vinegar for washing windows, much less a technology that can be used to undermine labour itself.
the mailing list entry that phoronix links to: https://lore.kernel.org/linux-media/CAHk-=wi4zC+Ze8e+p3tMv8TtG_80KzsZ1syL9anBtmEh5Z40vg@mail.gmail.com/
"
On Tue, 14 Jul 2026 at 19:01, Roman Gushchin roman.gushchin@linux.dev wrote:
I think it makes the point of sashiko - helping maintainers - unachievable. If the point to not use
LLMs in general, let’s discuss this, not how to make each use case more complex.It seems like [1] expresses a very anti-LLM position in general
Yes.
And no, that’s not the position of the Linux kernel.
I realize that some people really dislike AI, but this is an area
where I’m willing to absolutely put my foot down as the top-level
maintainer.Linux is not one of those anti-AI projects, and if somebody has issues
with that, they can do the open-source thing and fork it.Or just walk away.
AI is a tool, just like other tools we use. And it’s clearly a useful one.
It may not have been that “clearly” even just a year ago, but it’s no
longer in question today.There are other questions around AI (like what the economy of it will
actually look like in the end), but “is it useful” is no longer one of
those questions. Anybody who doubts that clearly hasn’t actually used
it.Yes, it can also be a somewhat painful tool, both for maintainer
workloads and just from a “it keeps finding embarrassing bugs”
standpoint.But the solution is not to put your head in the sand and sing “La La
La, I can’t hear you” at the top of your voice like some people seem
to do.The solution is to make sure those LLM tools help maintainers
instead of just causing them pain. There’s no question on that side.We’re not forcing anybody to use it, but I will very loudly ignore
people who try to argue against other people from using it.And no, AI isn’t perfect. But Christ, anybody who points to the
problems at AI had better be looking in the mirror and pointing at
themselves at the same time.Because it’s not like natural intelligence is always all that great either.
The kernel project has been and will continue to be about the technology.
Sure, the social angle of working on open source is important and
often a very motivating part of the project, but in the end that’s a
side benefit, not the point of the project.This is NOT some kind of “social warrior” project, never has been,
and never will be.In the kernel community we do open source because it results in better
technology, not because of religious reasons.And so we make decisions primarily based on technical merit. Not fear
of new tools.Linus"
Thank you for providing the full quote of what Linus actually said so folks have context, regardless of how people feel about the position he’s taking its always really helpful to have primary sources ❤️
It’s not fear over new tools, it’s frustration with how it’s wielded and how inappropriately, irresponsibly, unethically, and greedily it’s pushed.
AI, itself is not the problem. The people who use AI, and the people who steal and trample and bludgeon to get it and while using it are the problem.
To disregard this entire aspect of the conversation is… Alarming and disappointing.
I believe we should all be pushing for ethical and responsible AI use and development, and what constitutes that and why. If not, the only resulting conclusion should and will be, simply: “Fuck AI”, which is where we are, now.
What if we rush slop generators with ineffective safeguards to market instead?
And no, AI isn’t perfect. But Christ, anybody who points to the problems at AI had better be looking in the mirror and pointing at themselves at the same time.
No one is saying humans are perfect either. But Christ, the scale is clearly not the same. Before, an average bad dev had a limit blast radius. Now they can push a million lines of code and write a convincing description of why this is actually correct with little effort.
later
(Oops. Wait. You’re right to push back. This is totally wrong. This should have never happend.)
Blast radius you say say? Claude is that you?
LOL! Damn, I’m getting brain rot from work… 😭
I also don’t use entire cities worth of power and clean water.
AI is not the problem. It’s the people and companies behind the popular public AI services that are the problem.
ML algorithms are fine. It’s the capitalist use cases that are shit and bad for everyone.
And all the stolen and continued stealing and privacy issues and biases and all of the other ethical issues… But yeah.
This is the correct outlook on it.
Academically, artificial neural network algorithms are cool as fuck, even given their shortcomings.
Capitalist greed has weaponized them and done horrible atrocious things to “improve” them (stealing works they have no rights to in order to profit off of them, replacing humans who want to work, ruining the arts, etc.)
It’s doing some pretty awesome things in the medical field as well, where experts can review it’s output and make decisions based on everything else they know and are experts in
It’s weird, like if experts feed AI things they know to be true, correct it when it gets it wrong, rinse and repeat. Then only use that AI for the things they trained it on, it seems to be pretty good. Whereas AI fed the entirety of human shit and corrected by middle managers, who are only experts in inflating their own ego, tends to output a lot of self-inflating bullshit.
i think a yet unanswered question in this example is, how will the next generation of experts (doctors) be able to validate the outputs of AI when every new trainee doctor has come to rely on the AI tools so much that they no longer need to put in the work to develop the very skills necessary to validate AI output?
so far in human history, most experts in their fields have arrived there through a challenging process of many hours of effort and many repetitions. medicine is a “practice” that you should theoretically get better at over time. same as carpentry or sculpting or playing violin.
you can’t become an expert carpenter by watching videos of houses getting built. you have to actually do the work of building houses. over and over again. until you know what a well made house looks like, inside and out. experience and intuition play key roles in this process.
similarly, you can’t become a master sculptor by 3D printing a million different objects. or a talented pianist by clicking on a soundboard. people may appreciate and enjoy all the stuff you are producing, but is that the same as being a trusted “expert”?
a tool speeds things up and removes barriers to entry for non experts. very cool, the world needs fewer gatekeepers for sure. but the tool doesn’t necessarily teach you the specialized knowledge intrinsic to being an expert. but AI is very different from most other tools because it doesn’t help you like a hammer helps a carpenter or how a metronome helps a pianist. AI mostly just gives you a result. it bypasses the entire process. sure, arriving quickly at a result is wonderful if you only care about the product, but if the people entrusted to ensure the quality and validity of the results have only ever seen it done ‘the easy way’ I’m not sure if something important is being lost or diminished.
not trying to be clever or circular, just genuinely concerned that de-skilling is a likely consequence of long term (multi generational) AI usage. and the generation of experts (whose expertise actual trained the tools) will be retired or dead so we can’t ask them anymore.
I hate to admit it, but he’s right. I’m no fan of using AI either, but there are times and places where it can actually be a useful tool when used correctly. What’s really important is the quality of the contributions made to the project, and the usefulness should be judged on its merits alone of those individual contributions, not the “icky“ feeling people get from the idea of using AI at all.
Torvalds is clearly not here defending “vibe coding”, and he’s perfectly happy for people to not use AI at all. But when AI, as a tool, is used to responsibly, it can clearly be quite useful. And for the coders who can use it in a responsible and useful way to make meaningful contributions, I also think they should be permitted to.
He mentions the economic and environmental aspects of AI but I think dismisses them a little too quickly. While AI as a tool can clearly provide many benefits, the external costs (i.e. costs not directly born by the project but incurred by others/society as a result of use of the tool) should not be ignored.
And I don’t see mention of the seeming expropriation of works without compensation by the AI developers and the as yet unresolved IP aspects of using AI. These are part of the big picture economic impacts but also have moral and possibly legal implications.
In short, I think there are reasons not to treat AI as merely a tool.
Yeah he’s had a very neutral take on it. Like it’s a thing and it does good work some of the time but ultimately code not written by a person is harder to maintain. At least when a person writes code they typically have a strong understanding of what they wrote and can accurately communicate that to others.
As a developer I can say that plenty of developers have trouble accurately communicating how code they just wrote works.
That said its still infinitly being able to ask them questions on it then trying to parse AI code.
Hell, I’ve written code on Friday and by Monday it makes zero sense to me.
Who wrote this ridiculous, nonfunctional, unreadable spaghetti code?
Oh, I did?
… It’s not so bad.
Not only trouble communicating but often knowing what the code they just wrote actually does. I will ever remember one of the first times I wrote a thorough set of tests for what I thought was a correct and simple little bit of code: it wasn’t doing what I thought, until many iterations later.
My experience thus far is not all sessions even with the same AI are equal. There’s still that issue of erroneous errors and of course differing levels of quality of output each and every time it’s asked to do something. Like sometimes it’s bang on amazing and other times it’s straight up gaslighting you over statements that are blatantly false.
Like sometimes it’s bang on amazing and other times it’s straight up gaslighting you over statements that are blatantly false.
Well yeah, they’ve got no understanding of anything, it has no ability to detect if it’s right or wrong, it’s all just text prediction.
To temporarily drop into a philosophical debate, all humans are doing is seemingly action potentials in neurons which form models and do predictions not too dissimilar but a lot better and of course self aware (I’m assuming other people are self aware)
That said I’m surprised current models are even as good as they are, most people take the progress since 2017 for granted, then again the AI hype train has over promised like mad and when the thing works it can be flipping amazing, it’s just not consistently that good and failures are unpredictable. I honestly didn’t think we would be here until 2050 to 2080.
Alright philosophy aside, humans are easier to reason with so far, particularly when questioning.
Well written response, don’t know why you are downvoted. Makes me think back to a philosophy class where I questions if we are more than a machine that is able to product certain formulations or responses based on how our brains are set and the relevant input. For instance based on context it’s very easy for me to guess what my partner is thinking. For myself, where do my thoughts come from? Can I think a truly random thought? I’ve no idea and I think not. How is that different than ai? Because I have built neuron connects forming this framework that is my mind and allow my ideas to be formed unlike just randomly guessing. On the flip side, as an academic am I not just making an educated guess every time I think?
It is difficult to get a man to understand something, when his salary depends upon his not understanding it. - Upton Sinclair
Bullseye for Linus. Not really surprising given his pedigree.
The reaction here is a surprise however. Maybe it just took the right person to state it eloquently, or maybe the worst-of-the-worst on Lemmy haven’t read the article yet. I’m curious to see the overall reaction to this by the most incendiary of individuals; if they bother responding at all.
Removed by mod
The fuck?
"This is NOT some kind of “social warrior” project, never has been, and never will be. "
Respect for Linus gone… He’s right about AI, but all other open source community driven IT is always a social warrior project against the Corporate psychos that constantly tries to herd us in and suck us off the wrong way… F’ Linus for not rejecting those parasites strongly…
FLOSS/FOSS and anti-capitalist community based IT forever !
trying to claim someone else’s hard work as your own again?
typical ML
remember to smurf as your .world account next time
Past the whole thing about theft of code, theft of wages and theft of rights, my next problem with AI is that it’s viral, in even worse terms than say the GPL is viral.
Once you accept slop submissions into a project, you have the problem that someone would have to check it. But why spend your own mental and emotional energy, better dedicated to dealing with and interacting with fellow humans, to trudge through a mass of hallucinated, gaslighting crap? Better set up some AI that checks the submitted code.
And now you have two problems.
Fortunately, lists like Open Slopware exist.
Linus is severely mistaken that Linux is not a “social warrior” project, but I guess that’s mostly because he’s the face the community has to try and “sanitize” for corporate the posture that people like RMS have. Gone are the times where he would openly flip Nvidia the bird. Which is not bad, but it’s unfortunately ill-timed: if history has taught us anything useful this decade is that, right now, bootlicking to corporate and its trends such as “uniformity” and “neutrality” is not the win you think it is.
Oh that’s a neat novel idea. That it’s viral. I know about the concept, but that’s a good development in terminology for it.
Yup! Honestly we should publicly treat and sham AI usage that way when we interact with developer-oriented ticketing systems.
Psychologically, what do you think is happening to a lot of the people here who seem to be failing to see the issue with Torvalds’s statements, seem to be even defending it, or even seem to be further obfuscating the particular discussion as if agenda driven? Because it is understanding and addressing these perspectives, which are going to move the needle on this.
As much as AI code is systemically viral, so too, are ideas and rhetoric on its responsible, considerate, and ethical future, if it is to have one.
AI, as it stands in the form we know it now, has no responsible, considerate or ethical future. So long as there are pedophile oligarchies, any such technology can not feasible exist in an ethical manner.
Paraphrasing as the Star Trek fans say: first you get Socialism / Communism, and from there you get the matter replicators.
Linus is severely mistaken that Linux is not a “social warrior” project,
What the fuck are you talking about?
FOSS is inherently political, it’s about whether or not you should have complete control of something you own. That’s a political stance.
Everything anyone cares about is political.
I’d say FOSS is not more political than paying taxes. FOSS is just another way of using the copyright law enforcement without which FOSS cannot exist. If no copyright law enforcement is in place, then the GPL license wouldn’t have power. So FOSS is just another way a law abiding citizen may use their rights/entitlement and perform their obligation with copyright law like how they have right/entitlement and obligation from taxes.
I’d say FOSS is not more political than paying taxes.
Which already means a lot: not paying rent for the right to exist as a natural person in an environment managed by what is essentially a rent-seeker who got there beforehand and accrued more power, is already quite a political decision. And I’ve not even gotten to the political point of if the resources acquired via taxing are used correctly or for the benefit of that local society.
FOSS is not more political than paying taxes.
If we’re including the other option, not paying taxes, then I agree. The choice of whether or not to use FOSS is as political as the choice of whether or not to pay taxes.
Coming back here from another discussion, I do find it funny now that a legal entity can choose to not pay taxes but a human cannot avoid that in any reasonable way
I mean you can if you’re rich enough. But yeah.
FOSS cannot exist under capitalism, as it is quite literally the definition of Marxist Praxis. Linux is quite literally the single most wide spread and successful example of communist production in world history. It is, by its very existence, a “social warrior” project.
See that’s what’s confusing about it. He’s saying one thing and doing another.
Why? The proof that linux exists under capitalism already proves your point does not hold
Linux, in fact, does not exist under capitalism. The Linux Foundation does not own capital. It does not pay taxes. Value is neither rented nor extracted from the workers who fully retain all ownership of the labor value of the product of Linux. The same way a commune can exist despite capitalism also existing, it does not mean that commune exists under capitalism.
Hmmm, I guess it depends on what you mean by capital. Because The Linux Foundation certainly has assets that they could donate to other FOSS project
The actual definition of capital. By the people that invented the term. A resource that is rented to another in exchange for part or all of the value generated by that resource. The Linux Foundation, being a non-profit, does not own any capital whatsoever, otherwise they would lose their non-profit status.
Then why did almost all of FOSS predominantly start in capitalist, free market economy countries?
Why did communism?
Russia: a tsarist autocracy.
China: an imperial monarchy.
Vietnam, Laos, Cambodia, North Korea: colonial or agrarian societies.
Cuba: Batista’s corrupt military dictatorship.
Has communism actually ever risen from a free market capitalist country? The other way happened a lot. Marx was definitely wrong about this one.
As for FOSS, I believe the key is personal freedom, not the economic system that much. Linus just did what he wanted to do and he didn’t need any permission.
edit Downmodded and no replies. You guys usually have a great answer to everything, but I guess this one was too tough.
The GPL fanatics just don’t want their real leader (Richard Stallman) in the spotlight.
