Until Claude can diagnose and treat Kent Overstreet for whatever neurospicy killed bcachefs, I will never trust it blindly. It just does things faster. Sometimes it breaks things faster, other times is finds things ready to be broken faster.
Lawyer-turned-software-developer reporting in! (American, to be clear…)
My concern with AI generated code has to do with software licensing. Mainly…. What is the license of AI generated code??? What if it regurgitates entire sections of proprietary code? Unless I’ve totally missed some legal/tech news, these questions are FAR from settled.
I feel like I’m taking crazy pills sometimes, but that seems like something the Linux kernel community should be at least mildly curious about….
Look, to be fair, maybe they are! I don’t follow the kernel mailing list; I don’t know what they talk about. I’m just a humble backend developer/infrastructure guy.
But I know it worries me… personally, I mostly use LLMs in my own work via asking questions about code I’ve already written. Then I look at suggestions it’s made, or potential issues it points out, and decide what to do accordingly. But that’s just me!
If it regurgitates proprietary code the model should not have had access to for training data… well…
they should change the name to LinAIux
I like the spirit, but this is a pretty hamfisted pun. Doesn’t Lainux make more sense?
apparently many people’s stance on ai is “ai is bad unless my idol uses it”. there is not intellectual discipline, suddenly they are willing to ignore all the other problems with ai: jod lose, excessive energy use, training on stolen data, preying on vulnerable people, exacerbate wealth inequality and concentration of power. and let’s not forget the ones that apply specifically to Linus: diskilling and ai psychosis.
“b-but linus reviews his code!” 🤦
Here’s mine: the world would be a much, much better place if LLMs did not exist. Also, it would be stupid–today–to not use LLMs to find flaws in the Linux kernel. The only other option is to get your machines compromised.
Using LLMs for security auditing? Ok, but people are mad that LLM-assisted code, tons of it, is now getting into the kernel and Linus is no longer slamming the brakes…
The only other option is to use something inconvenient like OpenBSD, constantly audited for three decades, LLMs could find only very few obscure bugs.
OpenBSD, constantly audited for three decades
Oh buddy… Did you not hear that not only did OpenBSD start using AI for audits but that it actually found a handful of serious issues that decades of human audits have missed?
There is a difference between an AI doing a security review (which they have become quite accomplished) and generating code that gets committed into the codebase.
Again, how does one verify that the code it has generated isn’t some verbatim chunk of copyrighted code that it happened to be trained on, and it turned out to be the best fit? There’s no guarantee that a model will not do this, even if instructed.
Also, content generated by AI models has already been deemed uncopyrightable, and this includes code. So are some larrtsmof the Linux kernel now able to be copied freely without GPL restrictions? How can you license code or executables compiled from said code if you do not hold the copyright on it?
And conversely, many people seem to have a “the moment someone says something nice about AI I dismiss them as either evil or an ignoramus” stance.
Many people on Lemmy will start to crash out the moment you suggest AI is anything other than pure evil
It’s not the algorithms people oppose, it’s the way they are trained within a capitalist system (infringement of many copyrights without compensation), they way they are being marketed as a living mind that can basically befriend us and take a mental load off of us (they can’t and it’s problematic to humans who allow them to try), and they way they’re being used an excuse to buy up the wntire world’s capacity for computing so that rich tech oligarchs can force us all onto the cloud services for all computing in the future.
I think if at the very least they weren’t bulldozing people’a home and poisoning their drinking water in order to forever capture all technology and then rent it back to us we might have a more favorable view of the possibilities, but so far Big Tech has done almost nothing but try to find ways to make individuals and enterprise addicted to their Mechanical Turk in order to take advantage of us.
I haven’t been keeping track in any formal sense, but it feels like it’s gradually getting a bit better here. There’s still a very strong anti-AI current running through most of the Threadiverse but I’ve been getting upvotes along with the downvotes and there are a few dedicated AI-related communities too. Maybe as the population grows it’s just naturally getting more diverse.
As long as nobody is being like a personalized salesman of AI and getting it in everyone’s face, I think it’s okay. I still find AI as a useful tool. But I also don’t see AI as an everything-solve-all kind of tool. It has its limits and I’ve seen them by experience.
I’ve agreed with Linus on a lot of things over the years. This is not one of them.
But I hear what you’re saying, and we’re not talking about me.
It’s like there’s nuance that deserves to be acknowledged or something. Weird.
Nope.
The crucial fact here is: like it or not, Linus decides what Linux is. As we’ve done for 30+ years, we decide whether we accept his project as-is. Otherwise, we move on.
It doesn’t get more complicated.
No one decides what linux is. Linux is literally open source. You have no idea what open source means if you think open source means one person owns Linux.
Linus has decided what Linux is & isn’t since forever. It has his fucking name on it for a reason. Linus decided Linux means AI code so Linux has AI code now. No way around it. We can only pick until a specific version of the kernel when no LLM was used but after that the future Linux Kernelf will forever have LLM code in them.
No distro is going to stop using new kernels, everyone downstream is essentially powerless on this decision. The Linux Foundation is HUGE compared to 90% of the distro’s workforce (volunteers). Only IBM & Canonical surpass it & they are NOT going to maintain the next Linux fork for the next 30 years.
Likuc maens AI code now and for the foreseeable future. To my demise.
Just because it’s open source it doesn’t mean there’s not a person or group of people deciding what happens to the ‘main’/‘upstream’ version of it.
Sure you can fork it, and remove all AI contributions, but who will use your fork?
When people say “the Linux kernel”, they mean the version maintained by Linus Torvalds. They might mean a fork for a specific SoC or manufacturer or distro, but those are all just the Linux Kernel maintained by Linus Torvalds plus a few patch sets. There’s no considerable divergence.
Linus’ stance was initially “no AI written code”. And he does have final say in what goes into the repo he maintains. Now he’s changed that tune, he allows AI written code, and no matter how many forks you make to strip it out… what matters is Linus’ upstream.
Sure you can fork it, and remove all AI contributions, but who will use your fork?
This actually should be done. Let there be options. But the human fork won’t magically return you to the previous world of no-AI. AI will still audit it and find vulnerabilities humans can’t. Time will tell which version prevails, but I think at best the human fork will have only a niche use. Perhaps in education.
Your read on that was very different than mine.
Linus controls Linux. Anyone who disagrees with how he runs it is free to fork it because the source is open.
But good luck maintaining a Linux kernel fork for any length of time unless you’ve got lots of contributors lined up and a solid maintenance structure ready to deploy. It worked with LibreOffice and OpenJDK, but it’s failed with far more projects than it’s succeeded with.
I don’t like him and I don’t like AI.
Have you seen this? https://codeberg.org/ethical-foss/ai-free-foss-alternatives
Fucking love that they listed pretty much every single programming language that is used today, leaving only esoteric shit.
On a personal level I get it… but as a professional, good fucking luck trying to find a job without using any of those languages, not to mention the plethora tools.
I’m rather glad I’ve gotten out of tech.
I expect AI is going to come for every role there is and turn it into being a prompt monkey.
However, I can always turn a wrench!
Ai isn’t going away… It’s like a nuclear arms race. The government won’t let it go away… They need it and will do anything locally and globally to solidify ai (and digitizing life) as the modern foundation of governing.
If people don’t want ai, they should be completely boycotting their nations. Votes don’t reach this level of need. “Need” as in the government would even sacrifices its own people to reach the goal. For all we know it’s the whole reason America had trump as president… So they could distract and implement the social and psychological strategy of easing into the upcoming modern foundation.
It’s not just “the government won’t let it go away.” Users won’t let it go away.
I find AI to be incredibly useful for running my coding agent, for example. I run it myself on a local machine. Every AI company on Earth could go bankrupt tomorrow and I’d keep on chugging away with it. Lots of other enthusiasts and smaller companies making use of it would do likewise.
as long as it doesn’t start pretending to be my friend 🐙
I am sure out of all projects Linux has the expertise and guardrails to ensure LLM generated code works as expected
Exactly! It’s a tool that creates slop with an untrained hand. With a trained professional hand who reviews the work it’s just a helper.
Huzzah! Saurman it’s finally using the seeing stone! He’s such a wise old wizard. He’ll be able to see the future and help us choose the right path. This could never go wrong or cause any ethics issue or anything!
“model” means “mode”, not LLM here:
He therefore hopes that the unusually large number of commits “is a sign that the AI fixes are becoming more part of the ‘normal development’ than a flood of fixes that go into the whole stabilization phase."
“I think that’s a much healthier model,” he added, before hedging with the observation that “people are expecting 7.4 to be the next LTS [long-term support] release, which is another independent reason why some releases tend larger than others.”
Is he okay? He looks so frail
“Don’t you think he looks tired?”
I’d guess Ozempic, most of the folks my age that take it get frail looking for a while.
ActionRetro has the same look right now I guess they could be sick but…
Its probably mostly because he’s in his 50s and does a desk job for a career.
For anybody confuse about why Torvalds is suddenly such a slopper it’s because AI companies pumped millions into the Linux Foundation which gives him a paycheck.
Is that a fact or just a suspicion? Because I don’t see you providing any sources for that.
Don’t get me wrong I suspect so myself, but I wouldn’t make such a claim without backing it up.“SAN FRANCISCO – March 17, 2026 – The Linux Foundation, the nonprofit organization enabling mass innovation through open source, today announced $12.5 million in total grants from Anthropic, AWS, GitHub, Google, Google DeepMind, Microsoft, and OpenAI to strengthen the security of the open source software ecosystem. The funding will be managed by Alpha-Omega and the Open Source Security Foundation (OpenSSF), trusted security initiatives within the Linux Foundation, to develop long-term, sustainable security solutions that support open source communities worldwide.”
Immediately after this announcement Linus was suddenly a lot less critical of AI.
what makes what the article describes slop?
Way too many people just equate any form of LLM/agentic/AI with pure slop.
As a tech lead who has to review a lot of code: those who did shit without LLM do shit with it. Those who are good without, are good with it.
I’ve found that the quality of code produced by my team has dropped significantly since my org pushed everyone to use LLMs. “Let me throw an agent at it” usually means “here’s another headache of a PR for me to review.” Engineers with decades of experience are writing such utter fucking crap now, and it doesn’t matter how many docs I write for their fuckass agents to ingest, the quality doesn’t seem to be getting better. They can make their agent loop on it for functionality/bug and hallucination squashing, but that usually just means I get a bloated fucking mess to review. It’s miserable and it makes me want to go shovel actual hog shit.
Sincerely,
a very sad and tired senior dev who used to enjoy his job
As a senior dev… AI actually improved the code we wrote in my last job. Mind you the codebase was in a shite position (tons of technical debt, plague-like avoidance of inline comments because “all of that knowledge is on Jira/Notion” except it wasn’t; lots of legacy code, undocumented weird behaviours, everyone was doing their own coding style without regards to the rest of the codebase and so on).
With controlled AI contributions, we’ve managed to:
- refactor large chunks of code where the overall high level logic didn’t need to change, but the underlying tooling did (e.g. lots of UI layers got rewritten from a very early attempt at a standardised MVVM approach to a well defined MVI one, we’ve also managed to eliminate the remaining RxJava/RxKotlin legacy flows to replace them with Coroutines)
- UI layers got refactored in no time because the AI agent was able to convert them appropriately (and even fixed a handful of UI bugs)
- a handful of core monoliths (yes I know… I hated working with them as well, but we never had the chance to properly refactor) got split up into more manageable, vertical-specific domain controllers without breaking
- the elimination of a legacy overarching component that was estimated at 6 months, got reduced to 6 weeks
- unit test coverage skyrocketed because even if the code was hand-written, an agent would provide coverage, based not just on implementation but on intention (and adoption of TDD/BDD finally started spreading).
Meanwhile code quality improved because we’ve had well established patterns and styles defined in the agent instructions, and even the anti-documentation “pragmatists” got shut up when they finally understood that the AI works better if there’s fucking comments to explain what’s going on
Being “good with it” doesn’t negate the overwhelming harms it causes. Those harms are part of the slop: environmental slop, work environment slop, societal degradation slop.
Throwing around the word slop just devalues it.
Slop has a meaning and it does NOT equal harm.
We’re harmmaxxing!
The invention of electricity also did a tremendous amount of harm. Should we have also just tried to put it back in its box?
