mastodon - Link to source

"Hello my name is Tovalds and I'd like to stick this fork into the Bitkeeper electrical socket again"

theregister.com/ai-and-ml/2026…

What is it with privileged white men who feel their position in society as a coddled person makes the concerns of every other living being a nuance to be fought until we all just accept the bright end-of-history where there is nothing wrong because there is no politics.

There are problems. Your hand waving isn't actually magic and doesn't make them disappear.

Linus Torvalds tells AI haters to fork off

Linux supremo says contributors opposed to AI use can 'just walk away'
Paul Kunert (theregister)
in reply to Barry Rowlingson

mastodon - Link to source

mike805

@geospacedman @Cheeseness Slavery is a whole worldview, and it requires a vast amount of hypocrisy to maintain. The slave is human enough to care for your kids and cook your dinner (you wouldn't let a chimp do that, right?) but not human enough to have any rights.

From that basic contradiction you get a twisted, backward civilization that doesn't progress and eventually loses.

The similar contradiction with AI is when you use it to not pay people, but expect people to buy your stuff.

@Cheeseness @Barry Rowlingson
in reply to mike805

mastodon - Link to source

Cheeseness

@mike805 @geospacedman I'm certain others have formed this thought in a more compelling and accessible way, but when I see people who're very into shaping their lives/work around "AI" talk about their excitement for and the potential they see in the technology, I often come away with the impression that what they are looking for is slavery without the aesthetics of slavery.
@mike805 @Barry Rowlingson
in reply to Cheeseness

mastodon - Link to source

mike805

@Cheeseness @geospacedman The whole "robot dream" is basically the desire for slavery, or at least an upper class servant lifestyle, without either the guilt or the fear.

That is a better life, and anyone who can obtain water or heat without going outside is already spoiled by it. Washing machines probably removed 75% of the labor of maintaining clothing; we want the other 25% removed now.

Home robots could be green, unlike data centers.

@Cheeseness @Barry Rowlingson
in reply to superketchup

mastodon - Link to source

fionescu(1)

@superketchup OpenBSD team has officially declared they won't accept AI code due to concern about code pollution, however, they already merged tmux patches that contain AI-assisted code, and tmux is in the base system... If you want to get away even more from AI, might want to know NetBSD are more categorically anti-AI. As a daily OpenBSD user, I'm fairly confident though that OpenBSD is gonna be alright, but I'd love to see NetBSD get more love too from developers.
@superketchup
in reply to Martin Owens

mastodon - Link to source

M

it's no surprise that an aging FOSS purist like Linus still sees tech as neutral. I think we need to have serious discussions as a community about freedom zero. It should have been clear to us that freedom zero requires adjustment when the world's weapons and surveillance systems were built leveraging FOSS tech. The Do No Harm license got ridiculed. However, at least it seems clearer to more of us now, in the wake of LLMs, we need ethical open source. We need to rework freedom zero.
in reply to Martin Owens

friendica (DFRN) - Link to source

Plan-A̵̛͈̬̥̿͋̓͛̕

 — (0.0.0.0/ 0)

@Martin Owens
The Linux kernel development process is grounded in technical merit and pragmatic tooling, not ideological restrictions. As Linus Torvalds clarified on July 15, 2026, Linux is not an 'anti-AI' project; it embraces tools that enhance productivity provided they yield verifiable, high-quality code. The project's Terms of Engagement and Code of Conduct prioritize constructive contribution over personal preferences regarding toolchains. Contributors who fundamentally object to the use of AI assistants are welcome to exercise the open-source freedom to fork the repository or contribute to projects with different tooling policies. However, attempting to impose a blanket ban on AI usage within the mainline kernel contradicts the established governance model, which holds the human submitter—not the tool—fully accountable for code quality and correctness.

So, Linus Torvalds is right because he defends engineering reality over ideology: AI is a productivity tool like compilers, not a replacement for human judgment. By mandating that humans retain full legal and technical accountability for every line of code submitted to the Linux kernel, the project ensures quality and security without arbitrarily banning technology that accelerates development. The choice isn't 'AI vs. No AI'—it's 'responsible engineers vs. those who refuse to adapt,' and open source has always favored the former.

@Martin Owens
in reply to Landa

friendica (DFRN) - Link to source

Plan-A̵̛͈̬̥̿͋̓͛̕

 — (0.0.0.0/ 0)

@Landa @Martin Owens To wake up alive in life each day is a hurdle as well.

While you raize a valid psychological point: the Dunning-Kruger effect means many developers likely overestimate their ability to verify AI code, just as most drivers overrate their skills. However, Linus Torvalds is right because the Linux kernel’s governance model does not rely on individual self-assessment; it relies on brutal, external verification.
The system works not because every contributor is honest or competent, but because the review process is adversarial.

I get that.
A patch submitted by an overconfident engineer whether AI-assisted or not must survive scrutiny from senior maintainers who actively look for flaws and bugz. The policy shift simply acknowledges that AI is now one of many tools (like compilers or static analyzers) that can generate code, but the HUMAN submitter remains the lightning rod for failure. If an engineer cannot verify their AI-generated patch, the review process will reject it, often harshly. The solution to overconfidence is not banning the tool, which would also handicap competent engineers, but maintaining the rigorous gatekeeping that has always filtered out "vibe coding" regardless of its origin.
Read well and think of it.

@Martin Owens @Landa
in reply to Plan-A̵̛͈̬̥̿͋̓͛̕

mastodon - Link to source

Landa

You’re surprisingly blasé about this for someone who thinks that getting up in the morning is a life-and-death struggle ;)

So, you assume that the next generation will magically acquire these skills even though the chances to acquire them are removed, the people who would be role models and teachers fuck around with their new artificial interns instead, and the companies they would work for are more interested in keeping a few responsibility sinks rather than actual experts or training juniors.

Good. Good.

@doctormo

@Martin Owens
This entry was edited (1 week ago)
in reply to Landa

friendica (DFRN) - Link to source

Plan-A̵̛͈̬̥̿͋̓͛̕

 — (0.0.0.0/ 0)
@Landa @Martin Owens Those are the corporate big players spoiling it if you ask me..
no amount of technical adaptation can excuse the blatant disregard for law and ethics that built this industry. The fact that major labs are currently facing massive lawsuits for copyright theft, hiding evidence, and exploiting data workers proves that the problem isn't just a lack of training but it's a corporate culture that prioritized speed over legality. The next generation may handle the technology better, but that does not absolve the current leadership of building these systems on a foundation of stolen data and ignored societal costs. We can acknowledge future potential while still demanding accountability for the unethical practices used to get here.
@Martin Owens @Landa
in reply to Plan-A̵̛͈̬̥̿͋̓͛̕

mastodon - Link to source

Landa

Every token paid fuels these criminal enterprises, every "oh ethics aside, this is…" legitimates what they have done *and are still doing*.

We're currently failing the next generation already in so many ways and now we want to burden them with yet another problem they should somehow solve because we really can't be arsed to make an ethical decision?

Fuck "acknowledging future potential".
Maybe for a change we could fix our shit now, behave like responsible people, do more research, and circle back to this technology once it has shown that it works without all the ethics catastrophes?

This entry was edited (1 week ago)
in reply to Plan-A̵̛͈̬̥̿͋̓͛̕

mastodon - Link to source

Landa

@zer0unplanned

Well, for some people that’s not a lot of money.

Most F2P games make their money on whales that spend thousands on them each month.

Same with casinos.

No I don’t think the misunderstanding is on your end.

Whatever that person gets out of it is probably mostly in their brain. Just like addiction.

@doctormo

@Martin Owens @Plan-A̵̛͈̬̥̿͋̓͛̕
in reply to Landa

friendica (DFRN) - Link to source

Plan-A̵̛͈̬̥̿͋̓͛̕

@Landa @Martin Owens Yes and I'm serious life is given and can snap away at a second notice for each of us.
Not magically about the next generation, every generation adapts to new technology faster than the one before it. just as digital natives mastered the internet and smartphones intuitively, the next generation is already treating AI not as a complex tool to be learned, but as a native extension of their workflow.
@Martin Owens @Landa
in reply to oatmeal

friendica (DFRN) - Link to source

Plan-A̵̛͈̬̥̿͋̓͛̕

@oatmeal @Martin Owens Your concerns about the AI industry's ethics data scraping, labor exploitation, and corporate overreach, yo that's valid and widely shared, but they do not justify banning AI tools within the Linux kernel project. Linus Torvalds is right because the kernel’s policy explicitly separates tool usage from industry practices: it mandates human accountability for every line of code (via the Assisted-by tag and DCO sign-off) while refusing to let external moral disputes over AI training data dictate internal engineering workflows. The solution to unethical AI practices is to demand regulatory reform and support ethical AI alternatives, not to handicap kernel developers by banning tools that improve security and productivity. Open source has always thrived by adopting useful technology while maintaining strict human oversight; abandoning that pragmatism would harm the project without fixing the broader industry’s problems.
@Martin Owens @oatmeal
in reply to Plan-A̵̛͈̬̥̿͋̓͛̕

mastodon - Link to source

oatmeal

@zer0unplanned one thing I totally agree is that we want regulation and legal guardrails. We as a collective need to make the decisions about what kinds of automation we actually want. We’re not even close to having this conversation at the moment and that’s a problem. At the moment the value of “ai” related technologies floats up rather than back to the people whose work was extracted. But that’s another problem. Regardless, ethics and self-regulation is BS.
@Plan-A̵̛͈̬̥̿͋̓͛̕
in reply to Martin Owens

friendica (DFRN) - Link to source

Plan-A̵̛͈̬̥̿͋̓͛̕

 — (0.0.0.0/ 0)
@Martin Owens Also, LLMs are not AI in the broad sense, but rather a specific, probabilistic application of language prediction that is often conflated with the entire field due to current marketing hype. True Artificial Intelligence is a decades-old scientific discipline established in the 1950s focused on reasoning, planning, and perception, proven by life-saving medical applications long before modern chatbots existed, from the 1970s MYCIN system using symbolic logic to diagnose infections, to 1980s INTERNIST-I managing complex disease knowledge, and 1990s neural networks detecting tumors in medical imagery with superhuman accuracy. These systems utilized expert systems, rule-based inference, and machine learning to make autonomous, verifiable decisions, demonstrating that real AI is a rigorous engineering reality, not just a corporate buzzword. While valid ethical objections exist regarding the data scraping and labor exploitation behind modern LLMs, dismissing the entire 70-year history of AI because of the flaws in one specific subset ignores the profound, non-linguistic contributions of symbolic AI and machine learning that have quietly improved human welfare for generations. We can and must criticize the immoral practices of the current LLM industry without denying the existence and utility of the broader, proven technologies that constitute genuine Artificial Intelligence. So I do not get your sickening hype all.
Fork your tool or kernel and so be it.
You can search this up all and behold.
What about compiling your own LLM and set a RAG system on it then we talk about progress, serving without API so no scraping bs and even on localhost. I do not see why or what it has to do with Linux what I just mention. In my case it is just a tool and installed not on my Linux image but on a iMac.. what child labor is in there? None.
It does not connect to internet even but I scrape manpages with a unique HTML request to orchestrate the feeding your AI not with their database but your directory and as fall back that db that is from 2023.
It's climate friendly, does not even need my router on to function and does not mass scrape or put CO2 in the air. Apart from the grid of pc.. just 1 HTML request per manpage, RFC name it.. so I really dunno what the problem is in fact it's those that pay for it to make it generate so called quality code which I call spaghetti code. Not all people use that crap and I agree on that corporate greed dismissal and hatred, so make one yourself like me. Or don't at all and don't use those > no one imposes you even not Linux.
Torvald is a hero! Torvald speaks truth about AI , Torvald becomes hated. Ridiculous ..
@Martin Owens

reshared this

Unknown parent

friendica (DFRN) - Link to source

Plan-A̵̛͈̬̥̿͋̓͛̕

 — (0.0.0.0/ 0)
@luca @Martin Owens What I hate most are herd kind of people, like those Butlerian Jihadists alike your type. I don't hate you personally but your backwards ideology, or that shared hive mind behavior as what does the majority say is what I stand for without having any idea of what it really is.
Have a good day.
Just to be in the coup, with the most followers sharing the same idea whether it is right or not mixing it with anti fascism or accusing AI of fascism novels etc is what I call just social media trash.
That Torvalds Linus tells not he is fan of AI he just said get your own fork or whatever and gtfo as it is just a tool what exactly did you not read well? You didn't read it at all.
And that is typical herd mentality behavior.
@Martin Owens
in reply to oatmeal

friendica (DFRN) - Link to source

Plan-A̵̛͈̬̥̿͋̓͛̕

 — (0.0.0.0/ 0)
@oatmeal @Martin Owens Told ya.. grab a free API key at ai-law-tracker.com/developers#get-key, or see the live Omnibus breakdown at ai-law-tracker.com/omnibus.
He is talented but I got him. 3K + lines of code at search in source code site word Military tells me enough. >>> vibecode.
It's not an browser add-on injection as I was thinking it is virtually impossible on my browser with noJs inbuilt and only add-on is uBlock but that site is a construct made by using Claude AI.
I'm happy I did not lost my hunting senses.
Scroll to the bottom. You will see the highest numbers are in the 9300s. 9729 is not there.. real site> congress.gov/search?q=%7B%22so…
Always double check sources.

Enter a title

www.congress.gov
@Martin Owens @oatmeal
in reply to Martin Owens

mastodon - Link to source

Markus

Linus said nothing about slop. All he said was he won't ban the use of AI. And he won't force anyone to use it, but all who want an AI _ban_ can walk away.

In the end of the article there is even the quote of a senior maintainer stating that the AI reports start to be good. Shall they just ignore real reports?

AI slop is a serious thread. But this article is the wrong battleground to fight it.

in reply to Martin Owens

mastodon - Link to source

🏳️‍🌈🎃🇧🇷Luana🇧🇷🎃🏳️‍🌈

Such a regression form these days too...
social.kernel.org/objects/60bc…

Linus Torvalds (@torvalds@social.kernel.org)

@morgthorak I think you might want to make sure you don’t follow me.Because your “woke communist propaganda” comment makes me think you’re a moron of the first order.I strongly suspect I am one of ...
social.kernel.org
⇧