Everyone’s focusing on the AI policy (perhaps because Leah for some reason said this was the most important part of the release), but the UEFI/EDK integration and Chromebook support is a massive step forward. Suddenly Libreboot has multiplied the devices available by many times, and future devices will be much easier to add. This is huge.
They REALLY don't like AI, to the point of "wish I didn't need to rely on GCC and GRUB because they have AI in them" levels of hate AI. See this further down the linked page:
>The fork in question, of U-Boot, is heavily AI vibe-coded and thus useless to Libreboot. This is not mere conjecture; in said aforementioned U-Boot mailing list thread, the author stated thus. We will not, in any way, encourage or facilitate use of AI. It has been made clear that several projects Libreboot uses also have some AI-assisted commits; GCC, GRUB, etc. This is not our fault, but it is my intention that I and others will audit these projects in the future, de-slopping them and possibly even forking them entirely for that very reason.
> U-Boot upstream itself is staunchly anti-AI, at least at the time of this release, and much discussion of the above was had on their mailing list.
I saw this as well, which is funny because I've sent plenty of patches written using AI assistance to upstream U-Boot. It was my understanding that U-Boot was mostly following kernel conventions in general as many U-Boot drivers share code(in addition to the actual subsystem maintainers themselves) with their kernel equivalents.
I have learned to treat statements like that as a general red flag.
Someone who acts like a radical on practical issues will almost certainly act like a radical in other areas as well - including areas that matter to me.
There are debates to be had on every technology, but foaming-at-the-mouth kind of talk is a sure sign of a maintainer (and by extension a project) I want nothing to do with.
I respect projects with a no AI policy, but if that policy starts with the words “fuck no”, I’m out of there.
It's possible this type of "radical" behavior is why this project exists in the first place. I'm willing to bet if the Libreboot maintainer(s) were not "radical" they'd just be using a Mac and be content with a closed source bootloader.
Not to say this "radical" behavior is helpful, and I understand why people don't like it.
I don't really see it as a question of "separating art from the artist". That's usually invoked when other aspects of the personal life of the artist has the potential to distract from the art itself.
The parent comment is closer to saying that Picasso wouldn't have been Picasso if in an alternate universe he had avoided controversial/politically charged subjects like the Spanish Civil War and instead appealed to the lowest common denominator for a mainstream commercial audience to sell postcards.
AI in its current form is a commercial product created from stolen labor.
I think there are extremely legitimate reasons to be critical of that part at the very least, and I don't think it's a particularly "radical" stance within Free Software, which is fundamentally conscientious of such questions.
(I don't know whether the Libreboot authors have this particular point in mind.)
Has anyone who contributed to producing the wealth of human knowledge, art, code, and literature making up the training data been compensated in any way?
Well since the early 18th Century (in a limited way) copyright rights have been given to creators - the overwhelming majority of countries are signatories to international treaties (for the last 100 years) that give creators control of created artistic works (which has been deemed to include program code).
Exploitation of those rights is down to individuals, only since Google came on the scene has any company been able to get away with broad ranging infringement of these rights without censure.
As for your cheque, you have the right to sell your work. No-one is obliged to buy it, of course. You also, under copyright, have the right to make a license to exchange use of your works for binding the users to conditions of release of their derivative/associated works ([F]OSS licensing).
You still have the right to sell it, and if ML models have been trained on those works you have the right to seek compensation.
That already didn't happen, regardless of AI. The thing about freely sharing knowledge is that people take that knowledge and use it to build further knowledge that helps you.
You might say that the fact LLMs are being provided as services by big techs invalidate that. But some people taking the provided knowledge for themselves has also always happened. More importantly, though, is that big tech provided LLMs aren't the only way to access the technology. Open weights and local models exist.
Casting this as “knowledge” betrays a very fundamental lack of knowledge about how LLMs work, or a desire to mislead for whatever reason.
They are not databases of facts. Even if they were, that would still constitute a very serious violation. They aren’t, and they’re not particularly good at serving that function either.
No, LLMs are capable of reproducing things (text, art, other media) based on their training data. Consider being an artist who produced something original that somehow made it into that training set - your chances of actually making a living are now vastly smaller.
The result is the same as if there was no copyright at all: Big companies will just exploit whoever, no art would be produced.
This is a huge issue, and there needs to be a reckoning.
How do you think LLMs work? Because you seem to think they're a big compressed folder of every content they were trained upon, thus a big permanent copyright violation.
You're probably also capable of reproducting some art off the top of your head, your existence doesn't somehow violate copyright.
Besides, anything they're reproducing is the work of very successful artists whose work is widespread. Anyone who had their worked looked at once during training isn't being affected anymore than some artist who never had anything of theirs looked at.
Also, copyright benefits big companies much more than it benefits artists, by the way. Not being able to take any art for free and requiring to pay is a great trade-off for getting a monopoly over anything they pay for. Art would still be made regardless, just like it was before and just like it is right now even when, outside of these big corporations who mostly respect it, copyright is constantly infringed upon.
As I understand it, the argument is about defending intellectual property. Not that strange to hear from the Free Software community, where many people have strong opinions about stuff like licensing. The "Free Imformation" (pirate) community does not fully overlap with the Free Software community.
One thing is that most if not all models more likely than not were trained on leaked Microsoft and other proprietary code sets.
For projects like WINE or ReactOS that absolutely have to enforce hard firewalls between people with "insider knowledge" to avoid contamination and resulting legal issues, this alone makes the use of AI for coding highly problematic.
If you believe that, then why aren't LLMs free to use for every person on Earth?
If you want anybody to keep making anything, including programmers, artists, and writers, you're going to have to pay them somehow.
Anthropic and OpenAI would not have a product at all without the contributions of thousands of non-consenting humans. Yet they somehow have the gall to charge customers a premium.
Anthropic and OpenAI aren't the only ways to use LLMs. There are plenty of LLMs where you can download the weights and run them locally. There are even others that you can train yourself (see the LLM360 family).
They refuse to release the models, the training method and, training data. They don't even want you to ask it too many questions lest you teach another ai. Very "open".
Since that's your position -- have you ever produced proprietary software in exchange for money? How do you intend to make up for your unethical actions, if you have?
And, how do you envision the white collar economy, which is entirely built off of selling knowledge, will operate if you get your way?
> And, how do you envision the white collar economy, which is entirely built off of selling knowledge, will operate if you get your way?
It will crash and burn anyway. That kind of economy made sense before the Internet, before mobile phones made it possible to carry the entire world's knowledge in your pocket, only impeded by your ability to efficiently query the database and apply the concepts.
Now with AI? It's even worse. AI not just can efficiently query its own knowledge dataset it was trained upon, but it can also search other databases, apply the knowledge to the input and to put the cherry on the icing, it can do so when operated in natural language. And we're not going to be able to put the genie back into the bottle, not with people being used (and able) to run open weight models on local hardware. OpenAI et al don't have any actual moat behind them, they are just making stuff a bit more convenient and are for now the cheaper way to access compute resources.
For Western countries, the white collar economy is going down the drain, black collar (mining and other forms of dirty resource extraction) is already gone off to places that "compete" by using slave labor, and what remains of blue collar now that industry went the same way as black collar is heavily under threat because so many white collar people are fleeing towards the trades.
It's unclear from your comment whether you think this is good or bad. You certainly seem to think it is inevitable.
I just want to remind you that nothing here is inevitable - someone is making it happen, and it's only possible because we're not enforcing any of the laws we have about it for some reason, such as copyright. Well, the US isn't enforcing it. The EU is scrambling to come up with reasonable regulations. God knows what China is doing.
I agree that AI, like many other things, would have been an unambiguously great invention in a post-capitalist society, but that's not where we are. They are unbelievably powerful, impossible to control, and fully owned by powerful corporations, so the future looks pretty dystopic for humans who love freedom, art, and science.
But I hope you realize that we have the chance to fight back and get this right. But it does require that you do something.
> It's unclear from your comment whether you think this is good or bad. You certainly seem to think it is inevitable.
I think it is bad, but the consequences are manageable.
> someone is making it happen
We're past that point. Open weights plus local inference is already good enough for an awful lot of use cases. Even if we'd say we ban all AI training, we yeet China and the US from the Internet... it's too late.
> But I hope you realize that we have the chance to fight back and get this right.
The best chance we have is mitigate the consequences. That means first and foremost UBI, backed by actually useful corporate taxation. We need to prepare our economies for the shift towards an economy that requires fewer and fewer people to actually do work. For the moment, the only thing that is relatively safe from both automation and scale efficiencies is some trades (plumbers and electricians), some sections of healthcare and some sections of black collar work. Everything else is under massive threat from AI, automation (have you looked at precision farming?!) or both, and politicians seem like a deer frozen in the headlights.
This is the classic doomer take. Beware of dramatic but simple answers. UBI doesn’t really seem like a sustainable way forward, and at the same time there’s so much we don’t know about the limits.
Don’t play half-court tennis. Other people can read and write too, and social shifts are simultaneous slow, but then sudden.
For example, I’m not seeing a lot of trust in AI as a technology for things that are actually important. It takes one wrong diagnosis from an expensive AI doctor to erode all trust and make everyone prefer a human expert.
When your project is literally writing a BIOS/ UEFI alternative that exists for the purpose of making base runtime systems open and verifiable and to improve the user’s ability to trust in their system, an opposition to unverified machine generated code seems not only justifiable but sensible.
If Leah were not taking steps to mitigate generated code in the project, the project would no longer meet my needs as a Libreboot user.
> Someone who acts like a radical on practical issues will almost certainly act like a radical in other areas as well - including areas that matter to me.
Not using LLMs is a practical choice. The tech doesn't work, and if you let it run amok in your software, your software becomes crap. I try to avoid projects that use LLMs for development as much as I can, because those projects have a much higher likelihood of being a buggy mess.
I find it hard to square treating that like such a red flag while supporting free open source software more generally. The entire ecosystem is rife with ethics/morals driving the initiative. There is a lot of flag planting and red line drawing.
People say AI or not, if the thing is good then it should be used. But if a group actively rejects it (which is valid even if I’m not that militant myself) it compromises the project suddenly?
Personally I treat whinging about other people's ideals as a red flag.
It's libreboot. The entire point of its creation was to take coreboot and rip out the non-libre blobs. Its author is trans (IIRC), which also carries its own very pointed perspective on the world.
Some of it goes too far in my own opinion - I have no problem trusting AMD's final microcode for the Opteron 63xx any differently than the baked-in mask for the 62xx. But regardless, I respect someone who is pushing the whole envelope as they see it. At the very least their effort creates a lot of things I do care about!
Rereading this I realize I might be in the running for the RMS award for technically correct social obtuseness so let me just say that my intent was not to denigrate trans at all, just saying that being trans is itself intrinsically political (as in you're going to be continually clashing with many people's default assumptions, even before the manufactured "culture war" throws gas on the fire). And then furthermore being openly trans is going to impart those politics to the project. Which is most certainly her right, as it is her project! In fact I would say that trying to push a standard of you can contribute your code but stfu about your other ideals is basically cruel.
Well, yes, this is a project to create a fully libre, GPL'd boot chain free of binary blobs. The few users I know stick with aging ThinkPads. I'm not sure what you expected.
Likewise, it if claims something is “vibe coded” when it clearly is not (such as U-Boot).
Agentic AI has allowed 86Box to make massive strides forward. This includes a substantial contribution to emulate the Intel Inboard 386 that was created by someone who describes himself as a non-programmer. The community is now bigger and stronger, although we’ve had some aches and pains with slop AI generated PRs.
The same goes for Gravity Linux which is booting in my M4 Macs with working GPU and most other peripherals. What used to take years now takes weeks, so we can do much more thorough testing and ultimately have a better end product.
And what happens when the AI bubble pops and the various publishers and IP rights holders come around with their hands out claiming 86Box stole their code ala' SCO?
AI code is legally nebulous and relying on it for a serious project has major risks. Linux has already fended off SCO and has deep pockets to boot. Not all open source projects do.
More projects should be thinking less about the 'moral' issues (yes, they exist) and more about the nuts and bolts province and liability issues of using unverifiable code.
Yes, Linux uses it -because they're big and professional and have allies like IBM and Microsoft (ironies of ironies) on their side.
> And what happens when [..] the various publishers and IP rights holders come around with their hands out claiming 86Box stole their code ala' SCO?
It won't, for a very simple reason. The same reason why the world isn't in the middle of a nuclear winter right now. Mutually assured destruction.
Every single proprietary software vendor right now is making heavy use of AI code generation, and we're already at a point of no return where pretty much every single commercially relevant codebase out there is contaminated. Do you really think companies will go on a suing spree where by doing so they will open up themselves to being sued for exactly the same reason?
>It won't, for a very simple reason. The same reason why the world isn't in the middle of a nuclear winter right now. Mutually assured destruction.
An interesting take, which I honestly hadn't considered.
The problem remains that it is still a cudgel which larger companies can still use it to shut down smaller companies/free alternatives. I'm thinking a similar situation to the abuse of the DMCA -but we'll find out.
Yes, anti fascism is a radical position these days.
Already in 2022, before the rise of these LLMs, there was a book "Resisting AI: An Anti-fascist Approach to Artificial Intelligence" by Dan McQuillan.
And then:
> The rapid development of digital technologies, including applications of artificial intelligence (AI), is having a significant impact on the social, economic, and political life. From the way individuals communicate and form relationships to the methods by which political discourse and economic transactions occur, digital technologies increasingly mediate our lives and societies. But while presented as milestones in innovation and progress, these technological transformations have also raised concerns about democracy (Nemitz 2018; Kreps and Kriner 2023; Jungherr 2023; Coeckelbergh 2024) and seem to introduce and support mechanisms of control, forms of organization, and ideological patterns that bear striking resemblances to historical fascist phenomena.
> being plausible but slightly wrong and un-auditable-at scale-is the killer feature of LLMs, not a bug that will ever be meaningfully addressed, and this combination of properties makes it an essentially fascist technology. By “fascist” in this context, I mean that it is well suited to centralizing authority, eliminating checks on that authority and advancing an anti-science agenda.
I know this won't be popular with this crowd but most of you have become supporters of the rising fascism in the United States by not resisting these LLMs.
> Yes, anti fascism is a radical position these days.
Given that these groups are basically composed of communists, if not tankies, that's not surprising. Specially when some of those served as sabotours for the USSR (notorious for its respect for human dignity).
This is an incredible piece of anti-intellectualism. Its arguments aren't compelling either: AI might flood with innacurate studies that are difficult to check. Rather than improving said check as to reduce the reliance on trust and improve science as a whole; like the good reactionary the author is, they proppose to ban this particular technology. AI also centralizes authority... because? There is also an irony in them complaining about this theoretical centralization while also trying to preserve the inherent trust-based system in academia, so I suspect they aren't actually committed to proper intellectualism. There is more to complain about this post, but, then, it would hijack this whole comment.
> Marc Andreessen have called one of the two authors of the very Fascist Manifesto his patron saint. Search for Marinetti on [...]
Is it surprising that the guy who founded Futurism, a rather radical and aggressive celebration of innovation and progress, would be mentioned in a movement about technological progress?
> A very influential blogger, Curtis Yarvin who have started Dark Enlightenment, the philosophy behind it all regularly posts straight nazi shit. He also has influenced JD Vance and Peter Thiel who have been instrumental in the creation of OpenAI.
JD Vance hasn't been instrumental to OpenAI's creation, and Peter Thiel is just one of its original investors. Not to mention, NRx is a mess of movement composed of contradictory ideas.
I have noticed you have been using "these LLMs". While I do agree that distancing from OpenAI's models is good (even if it's just on the basis of decentralization), the anti-AI movement either wants to monopolize AI via regulation or ban it outright.
This is just as bad as blanket hating/rejecting AI without ever trying to use or understand it. I don’t care what tools a craftsman uses so long as the job is done correctly/well. If you can somehow prove that AI is necessary for this that’d be impressive.
Seems more like clean room RE to me, or general free software stuff. You can't contribute proprietary code to free software, if someone doesn't agree to a license change you have to remove/rewrite their contributions, if you've seen the Windows source code you aren't allowed to contribute to ReactOS (IIRC), etc. LLMs are black boxes with dubious training data and aren't really "you" writing the code directly, so whether you can then license that is questionable. This is especially important for a project like LibreBoot that is trying to liberate machines. The only "easy" solution I see is to abolish copyright, which would be fine. As it is a lot of software is being tainted and the whole world of free software is in danger. It's a shame people gloss over this because they think LLMs are cool or help productivity or whatever. Everything's been irradiated now. Now if someone wants to guarantee a given project had no AI contributions they have to fork a years-old version and likely lose a lot of good non-LLM code in the process. What a waste of time and resources that is. Setting us way back. This needed to be taken way more seriously on day 0.
That's not what it implies to me. And let's not pretend that a person with this approach and this manifesto is going to publish their patches to deslop slopped upstream projects.
contributing to projects that have slop in them has legal liability. you'd have to essentially hardfork to remove the slop from git history and such.
for my part, when i do work on this, what i will do is this: re-use existing nuke.list mechanism in lbmk (used already to delete dubious files, e.g. code files with unclear licensing in u-boot) to delete the slopped files. then provide cleaned versions of the same files, with necessary patching re-written as necessary, and files that use it patched to include e.g. foo.deslop.c instead of foo.c - then git history is removed inside tarballs. filter out the changelogs for projects accordingly.
this is the most reasonable way to do it on projects with very little slop. with projects that become almost entirely slop, a hard fork becomes necessary.
So what OS are you building on and running on the machine? https://codeberg.org/ethical-foss/open-slopware#kernels doesn't seem to list anything I would consider a good alternative to Linux, although illumos or 9front becoming the preferred build environment for libreboot would make my inner nerd happy.
currently debian linux, though i have porting lbmk to openbsd on TODO. with a few small exceptions (e.g. recent tmux patch) they're generally quite hostile to use of AI.
yes that is why i mention openbsd. i haven't ported lbmk to it yet, was going to months ago. i'll start up that effort again at some point. future releases will one day be built with openbsd, and when that day comes, my reasons to use linux will drop to near zero. i'm an avid fan and user of openbsd - i even maintain a port for it (librewolf web browser)
> think for yourself. AI will only ruin your mind and deskill you. here are some articles:
> think AI makes you super hacker? think again: https://arxiv.org/abs/2507.09089 ["Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity"]
> consider that the infrastructure for AI is economically unsustainable; when the bubble bursts, and it will, we will suddenly be left with millions of people thoroughly deskilled who will be useless. lots of reeducation needed, and lots of people will just be left behind.
> consider that it is controlled by only a minority of very wealthy individuals; with it, they have unprecedented control over human thought through manipulation of information, and control over what information is presented through filtering (what they call training).
> AI is a threat to the health and democracy of our entire world.
> libreboot will not facilitate its use. AI is the most dangerous technology ever invented by human beings.
This one especially really just shows they are cherry picking things which support a particular position regardless of the relevance, the difference between Early-2025 AI and Late-2026 AI is absolutely massive. Early 2026 it was reported by Greg Kroah-Hartman that AI bug reports rapidly went from obviously wrong or low quality to being real reports.[0] The kernel now has AI tooling heavily integrated into patch review workflows as well with quite good results.[1]
> the difference between Early-2025 AI and Late-2026 AI is absolutely massive
If you only accept papers that were published in reputable venues and analyze LLMs that are less than a month old, you are left with zero useful papers (since writing a useful paper takes longer than a month). It is impossible to have an intellectual discussion if all scientific material is disregarded.
> If you only accept papers that were published in reputable venues and analyze LLMs that are less than a month old, you are left with zero useful papers (since writing a useful paper takes longer than a month).
Why would one look only at formal published scientific papers? Especially if the only publications are known to be quite outdated and highly likely to be no longer particularly relevant anymore.
> It is impossible to have an intellectual discussion if all scientific material is disregarded.
One can certainly have an intellectual discussion even for areas without relevant formal published scientific papers.
Certainly one can do so if the subject matter is in the area of your own expertise, that can even lead to scientific papers being written.
I'm an experienced open source developer leveraging AI tools for linux kernel, u-boot and buildroot contributions in addition to reverse engineering, so clearly this subject is well within my area of expertise regardless of what some outdated published research says.
I probably wouldn't even disagree with that paper's assessment at the time, however the situation is very different now. I have noticed that there is an issue with many experienced open source developers simply having not figured out how to properly leverage AI for open source projects they work on. Probably one reason for this is that productive workflows with AI are often very different from previous workflows and it takes some practice to figure out how to best use AI tools, these best practice can also change over time.
Unfortunately it is. Thought the whole thing was a scam until six months ago. I think pretty highly of my development skills and have to be objective and admit I am outclassed in debugging and operational tasks now (and it isn't close). Writing non-script code isn't quite there yet (they are bad at abstraction) but I can't confidently predict how much longer that will be the case.
> when the bubble bursts, and it will, we will suddenly be left with millions of people thoroughly deskilled who will be useless. lots of reeducation needed, and lots of people will just be left behind.
what do you mean by this, that costs will become too high and people won't afford it? hardware won't go up in smoke "when the bubble bursts"
People won't be left without AI/LLMs. They'll be left with less value in their investments. The tech is here to stay, the market discovering the actual commercial value of these products won't change that.
How terrible. I'm sitting on multiple AI reverse engineered firmwares that I'm not sure what to do with. I suppose they'll never make it to upstream now. My only option is to ask AI to integrate this stuff.
This conflict has come up a lot in open source lately and the answer has been to just create your own repo and space where you can dump unlimited ai output without needing anyone’s permission. If users find it useful they can use it, and there’s no conflict with maintainers who don’t want it.
Time will tell if these new AI allowed projects like artcraft will prove more popular than fully human curated ones.
The whole point of open source is to get this stuff upstream so that a rich software commons is created.
I'm not sure I want to participate in this space anymore if everyone is just going to have their own isolated forks. I might as well keep my own private version of everything and never deal with maintainers ever again.
There has always been peoples personal forks and random one man hobby projects on Github. I think the Artcraft project is a good example of what's to come. Gimp would have never accepted thousands of commits to convert it in to a replica of Photoshop, and they shouldn't have to. The people who want to go full speed AI programming can run their own project and don't have to bother the Gimp maintainers.
Is it really out of the question to stop using AI and continue contributing upstream to free software projects? It was just a few years ago this was the norm.
Yeah, I saw this "open-slopware" list listing many battle-tested web servers as "permissive AI policy", while one of the alternatives (not all though, some inherit battle-testing) isn't really battle-tested. In fact, I saw entangled codebase and not-so-good coding practices.
This list seems...odd. For instance, the list decries the permissive AI policy of the Linux kernel, of Fedora, of Bazzite/Bluefin/etc...and then includes Secureblue as an alternative...?
Secureblue is a neat project, but as you say, this would make more sense as a generic discovery list. If you want to avoid projects with a permissive AI policy, you should probably take into account that Secureblue is based on bluebuild (from the Bazzite team), Fedora, Linux, etc etc etc.
not that im particularly pro-ai, but it seems a bit silly to rely on a paper about early-2025 ai when people are using late-2026 ai, given how different they are.
The interesting thing about that paper, I would have said, was the _perception_ gap; people believed that it was making them more productive even when it was doing the opposite. That effect presumably hasn’t gone away.
> how are they different? did the algorithms upon which they operate suddenly change?
Just like most technologies AI algorithms and tools get continuously refined/improved over time.
One thing modern AI has gotten especially good at in the last year is reverse engineering, it went from struggling and frequently making mistakes when analyzing binaries even with say a specialized Ghidra MCP integration to being able to very quickly and accurately analyze even heavily stripped BROM binaries with almost no user guidance or specialized environments. Binaries that would previously take days or weeks to analyze can be analyzed by a modern AI in minutes to hours.
The difference is massive and obvious to anyone who has been working with these tools over these time periods.
Presumably yes. It is safe to assume that algorithms running the model changed, the training procedures changed, and the scale changed. Unfortunately the AI companies refuse to disclose the details of any of this, which is a harm to human knowledge (hopefully a temporary one). Yet the results speak for themselves.
For me it's only been in the last month with Opus 5.5 that I've been seeing code quality I actually respect (though this is with substantial human effort in the loop). Though this experience is from work where I don't have access to non-Anthropic models, nor Fable.
i do not believe you are asking that question in good faith.
and, if you are the leah who wrote the policy, i am positive this would be a completely fruitless discussion as you are obviously entrenched in your beliefs. i am not under the delusion that anything i could say, show, or prove to you would change your mind.
calling anyone with a different view from you a bot does nothing except make you look immature, and just proves my point: you aren't engaging in good faith, and you don't want to actually have a conversation. i'm happy we both don't need to waste time on one.
good luck with your project. i'll mark this one down as "like the art, dislike the artist"
It takes time to research and publish. It'll always be some slightly old model. Is this argument just a permanent get out of jail free card? Add 10 to each year, will we be seeing that said in the future?
i don't think models will continue to improve at the rate they have been, so eventually it will be less of an issue. there are also questions to answer and studies to be done that aren't as sensitive about the specific model.
however, productivity is sensitive to the model, and models have significantly improved in the last 1.5-2 years. sometimes things move quickly and research gets outdated. ai is not the first field where this has happened.
the research is still valuable, it just has to be acknowledged that it is specific to a point in time. using it for an argument 2 years later just shows that the leah is being intellectually dishonest, and it weakens the overall message.
There is a respectable way to approach this and this.
The right way is something like, we've been getting spammed with a bunch of PR's from AI that we find difficult to review (because AI allows the average bob who isn't technically minded to make a PR). We have therefore decided not to allow AI submissions.
---
I think the people who have some moral / ideological issue and bring that into projects kind of feel like clowns to me. I agree with Linus that the point is to make something that people find useful, and obviously you want it to be maintainable.
Saying "it makes you dumber etc" is not related to the project, with all due respect it's non of your fucking business how I learn / retain skills.
I get it, but everything is ideological -- whether you want to hear about it or not.
In a world where software exerts considerable control and influence over peoples' lives, your ability to insulate yourself from that influence is a priveledge. Software IS ideological. Free software was created so people can have the freedom to share and collaborate on their own terms. There is an extent to which it gets extreme, but without the GPL we would live in an even more monopolistic "free" market hell world than we already do. That's a moral obligation that spawned a practical one. If no one ever took a moral stance because you find it annoying, the world would be an objectively worse place.
I always find these sorts of responses fascinating. “Oh no people are being mean about our lovely robots!!” Like, people are entitled to dislike and be sceptical of LLMs.
Yeah, the (poisoned) AGENTS.md file reads lots like "f*ck AI", which feels like anti-AI dogmatism for me. If I really wanted to use this project, I would use Coreboot instead or fork it to remove proprietary blobs.
I can't use critical software that rejects AI to this degree for privacy reasons.
You can't expect to sufficiently protect your users privacy when not using AI for finding vulnerabilities.
Same thing with Asahi Linux with their highly anti-AI stance. I had a bit of hope when I saw someone was making an LLM assisted fork but it's not the same support.
While MacOS is likely becoming much more secure, Linux distros that reject AI would fall even further behind, if that's possible.
Ai and privacy in the same sentence is wierd. You supported your argument, I guess, but unless you are using local ai, the developer would be sacrificing their own privacy.
Something about having a bot controlled by a third party comb your entire codebase and make changes on your behalf strikes me as not the most privacy-friendly approach to software. I know your question wasn't being asked in good faith, but I've answered it honestly.
The end result is open source so bots are combing through it anyway. I’d personally want to sandbox the agent with only access to the code being worked on and none of my personal data, but open source code isn’t my personal data.
>Privacy is the assurance that your data is only seen by the parties you intend to view it.
I’ve been assuming we’re using open source software. When developing open source, you’re putting your code out publicly. Your code is not private data.
As far as the agent acting on your computer, you can easily use Claude Code on the web on Anthropic’s VMs, or just run it in something like UTM or another user on your laptop.
So I don’t see how this is a big loss of privacy. By the definition, you’re intending Anthropic to see your code just as anyone else can.
Ways you could use AI in a non-privacy respecting way would be like letting it read all your personal notes while assuming the AI company will not share them or look at them. I don’t see code as sensitive like this.
Your prompts, pattern of speech, usage patterns, when you work, stuff you say to the LLM besides coding related. Those would be a fair argument for compromising the privacy of the developer. But as long as you consider what you say to be public, then you're intending that data to be out their, so your privacy isn't being compromised. You could make the same argument about developing in the open on GitHub. Anyone can learn about your usage patterns, time zone, etc. It is also fair to say you're giving that AI company more power over you by giving them exclusively that information, while on GitHub it would be information given to everyone. .
You may not, but that entirely comes down to who you're willing to trust. Its certainly not unreasonable for an individual to handle information deemed too sensitive to be allowed to be exfiltrated to a third party. Your use case is not the same as the entire world's.
The comment I replied to was saying that "the developer sacrificing their own privacy". I don't think AI for coding significantly invades developers personal privacy.
The Assisted-by: tag is no longer requirement because LLM assistance is the default.
Robot friends enable you to get much more done faster and more accurately, that if you call yourself an engineer you're a damn fool not to make use of them.
Unfortunately this is bad for one of my projects, a slop detector extension... it assumes that AGENTS.md be short in size to say stuff like "You may view and use this code but not edit or commit anything"... this rant is long enough that it classifies as slop itself!
I find this hilarious since I'm actively having my clankers work on porting EDK2 to an SoC that's used in a Chromebook, though not the ones mentioned here.
Are you publishing anywhere that I can follow? In particular, it would be really helpful if you could publish the prompts and workflow, so that the rest of us can more quickly use it to replicate your work on other machines:)
(Signed, the owner of several Chromebooks in need of better firmware)
I think it's better that some people have strong opinions. For example, the world would suck if everyone was rms, but he has nevertheless had a positive impact on the world imo.
RMS is extremely radical but he's also reasonable and consistent unlike these guys. There's also GNU Guix which recently banned AI-assisted contributions [1] due to copyright infringement concerns and licensing doubts. This is an entirely reasonable take for a hardcore free software codebase, it's literally built around this political concept.
On the contrast, libreboot's stuff reads like a parody or a shitpost. It's hard to believe they're being entirely serious.
The libreboot maintainer has always seemed irrational, but then again, a rational person would have just left their proprietary bios alone and we would never get projects like this.
i don't live in a free country. the UK has enacted several laws impeding fundamental civil liberties in recent years. rolling back on several human rights laws, now they're even removing the right to trial by jury, etc. it's quite troubling actually.
I'd rather have the project that works, rather than one that has the correct political stance.
You can look at what being achieved right now with AI, GoW3 (an absolute master piece BTW) on the PS4 Remaster can be played on Linux at 120FPS thanks to some decompilation/reverse engineering framework with AI.
I am generally also fed up of politics (whatever they are) forced into projects.
Also AI doesn't make you stupid. It is bullshit, he has cherry picked information to sooth his own biases.
It is bizarre having conversations with some people. There is a reverse engineering framework that had God of War 3 remaster running natively on Linux at 120 fps and all this tech is AI-assisted. A friend of mine was demonstrating it running to me last night, and then people like yourself tell me AI can't do anything when you can download and run it yourself.
I just had it write a bunch of script hooks for gamemoderun and at work I had it do probably a week's worth of work in an afternoon earlier today. It also smoke-tested everything for me.
It is utterly bizarre, and I think you are just denying reality.
You didn't get my point. It might very well be that the only project that lets you actually freely boot a computer in 5 years will be this one, specifically because they did not let an AI touch it.
In that regard, it will be the only project, like you put it, "that works".
Like I said, I don't necessary subscribe to their view, but the diversity of opinion matters, it helps with minor things like, oh, survival.
What you are saying is utter nonsense. I really don't know where to start.
Why would AI matter? If you say legality, I really don't care that much about copyright law (I run a torrent box 24/7), and I would just git clone the repo and put it on tor. That is what people have been doing with plenty of projects that are more legally dubious.
Also, diversity of opinion doesn't necessarily help with survival either. I swear people just say things because other people have said it.
But they're doing it under the banner of the libre software movement, whose raison d'être is free (as in liberty) software. Having political strings attached is antithetical to the movement.
Libre software is a political movement. And it makes sense for libre contributors to be highly skeptical of a tool that may or may not violate copyright, profits from the work of other libre codebases and is controlled by tech giants.
I think there is a deeper point regarding software freedom specifically - if a person is unable to understand the source code of something, then they are less able to modify it. It's why the definition of libre software centers around source code, even though modifying binaries is entirely possible. Despite being able to have a libre vs non-libre disassembler (analogous to proprietary vs open weight AI models), needing to use a tool at all diminishes your own ability to make changes to the code.
Now having said that, AI also increases software freedom in that it enables many people who aren't otherwise able to understand codebases (or even write simple code) to create/modify their own software. When we're talking about AI in the context of libre software, this is the constructive aspect we should be focusing on! Well defined plugin APIs that we know people are going to lay slop upon, etc.
But still, this doesn't have much bearing on a well-established fundamental systems project that needs to remain generally concise and human-understandable.
> Having political strings attached is antithetical to the movement.
I believe you may have a rather narrow notion of what politics are.
In my book, politics is the art of making choices when your ability to predict the future is limited, and therefore the willingness, courage and fortitude to make choices under uncertainty.
Under that definition, they're making a choice, one that materially and obviously can't be backed by sound reasoning or logic (because predicting the future is hard), that's politics.
Agree in general, though with Libreboot the pedantry is more or less the point of the project. You fork a project and put 'libre' in the name, I personally see that as a statement of dedication and expect you to stick to those guns. The experiment isnt much of an experiment otherwise.
When the alternative is being surrounded by rabid AI enthusiasts who absolutely lose their shit whenever anyone says "no" to them, being left behind is a great choice.
I can't be the only one who thinks the way they handled that is incredibly disrespectful right?
It's their own business to disallow AI contributions.
But instead of adding that to the Code Of Conduct they chose to knowingly misuse a file to target a group of people they don't like.
If I want to use AI on my copy of that repo, I have to delete that file and consider if the developer has added any more of what is (by intent) basically malware to the source code. Sure the file is just an unhinged rant, but now I have to wonder if they attempted to put some prompt injection in the repo.
“If I want to crawl a website with a robots file that says ‘fuck off’, I have to ignore the robots file. If I want to use AGPL code as part of a commercial web service without creating my source, I have to ignore the license.”
Like, c’mon now. This is far from the weirdest open source restriction.
I don't see how any of those things are similar, I think you've missed my point. Those hypothetical authors are using robots and license files correctly.
Would you be OK if a repo had Visual Studio files that were designed to crash visual studio if they are used, just because the maintainer hates Microsoft? That's what this feels like, a deliberate malicious abuse of the intent that adding agents.md would otherwise imply.
Linus's take is just the same nonsensical false equivalence I've seen a million times. He sure has fallen a long way since his glory days.
Someone should ask him if he's also planning on no longer reviewing patches to the kernel and just accepting them unconditionally, since AI is no different from a compiler after all, and nobody reviews compiler outputs.
Yes, and that take is bad, because AI is not a tool in the traditional sense. In the software world, tools are generally expected to be deterministic. Would Linus use a compiler that is known to have a random chance to output the wrong instructions? Or an editor that has a random chance to swap something you typed for something else?
> compiler outputs are tested by the compiler projects + the automated testing infra
And AI outputs are tested by the AI companies and the various testing suites that are run on the kernel. If they're really the same like Linus seems to believe, there's no reason to ever read AI generated patches before merging them.
You say "faith-based political filter," I say "standing on principles." Your valuation of outcomes and outputs over inputs and intent is also a faith-based political filter, in that they make the value judgement that outputs are the only thing that matter when shipping a product or managing a project.
I'm surprised Libreboot was able to add all the ChromeOS devices at once, Just adding the 2 ports for Heads hasn't been easy, although I'm not an expert when it comes to Coreboot.
Anyway, I love seeing more support for Chromebooks, they are great devices. Funny enough my Acer Chromebook (Kano) is now more open and more secure than my Framework 13s.
It’s great to see this happening at the same time—Heads has some important features that Libreboot lacks.
It may be harder to port it if those features require neutered (rather than simply removed) blobs. Maybe that’s why the Libreboot port was easier? Not sure, I’m aware of the benefits that Heads provides but I don’t know much about the internals.
It's refreshing to see serious open-source projects adopting a strict policy on LLM-assisted contributions.
Besides considering frontier LLMs a serious threat to humanity, I find LLM-assisted programming simply...not fun! I tolerate only AI-assisted bug reports, as they are genuinely useful.
I guess I'm AI centrist lol but to me, I can see a way around the 'threat' to humanity issue simply by going local and avoiding the use of datacenter ai. That might reduce the quality of code but running it on your own computer is a lot better -at least IMO.
My issues, personally, are the IP rights issues. What was this shit trained on? When the bubble pops (as I've said a lot in this thread -I know) who's going to shake me down?
> Imported all 233 currently supported Chromebooks from the MrChromebox coreboot distro. Baytrail (Kip) tested - others need testing. Source-only in this release (must build from source, no binaries provided).
Nice. I think I loaded MrChromebox onto my Google Pixel some years ago. I should see if it can get Libreboot now.
This cross-industry corporate naming trend, where each company picks a word or two and then creates a convoluted web of products with overlapping functionality and intended audiences, and expects us to, like, follow along... This trend needs to end. Microsoft Copilot 365 for Azure, meet Gemini Pixel Colab for business, and kindly ritually immolate each other.
>"Libreboot is free, open source boot firmware replacing proprietary BIOS/UEFI firmware on supported Intel/AMD and ARM64 hardware." [...] MrChromebox supports nearly every Chromebook..."
This is great news! Finally an open-source UEFI replacement which can work on nearly every Chromebook, including used/old Chromebooks!
Now, I haven't reviewed the source... it may still contain binary blobs for various drivers; I don't know. If it doesn't, then that's absolutely awesome! But if it does, then remember that having some of the UEFI firmware code open-source is a huge step forward, and it's only a matter of time before those closed-source binary blobs are reverse engineered (possibly by AI) and/or rewritten from scratch entirely, as open-source codebases...
So, a good step forward for open-source, right-to-repair, understandability of the boot chain, the Chromebook hardware ecosystem (used and new), and tranparency in general!
If you want something with more power look into the Dell Latitude 5400 Chromebook Enterprise or the Dell latitude 5300 2-in-1 Chromebook Enterprise. Both of these devices have DDR4 user replaceable memory.
There is also the Framework 13 Chromebook edition, but that seems almost impossible to find now-a-days.
there is upstream code but it's not well-integrated, and google's tree is largely optimised for their own payload called Deptchcharge. mrchromebox tweaks coreboot for use with edk2 so you can use normal operating systems, and the modifications that they make are extensive, e.g. mechanism in EDK2 to update EC firmware at boot time. mrchromebox is, for our purposes, a coreboot distro or indeed its own upstream. libreboot takes mrchromebox and makes several of its own modifications. "just use coreboot" is easier said than done :)
I realize that, but what's the selling point of Libreboot on Chromebooks when Coreboot is already supported, and it merely merged in support from Coreboot?
Assurance that binary blobs and Intel ME support has been properly removed. It’s possible to do this with Coreboot if you know what you’re doing, but this is an off the shelf configuration.
Nothing has changed about modern Intel CPUs requiring binary blobs. Libreboot is no longer blobless. Canoeboot is blobless. Libreboot is like decaf. It's not caffeine free, it's decreased caffeine.
You’re correct, I guess haven’t been following this over the last few years. It looks like blobs are now merely minimized and ME is neutered with me_cleaner. So the only difference versus stock coreboot is that these minimizations are now the default configurations.
Old Libreboot is now apparently Canoeboot, same dev but different project name.
Neat project. I find their new anti-AI policy to be interesting, but not at all useful to anyone. There are no objective arguments. No evidence of poor code quality in late 2026. No examples of why AI can't help the project. Just a rant.
I am very glad to see people with opinions that flow into what they do, but the question becomes: are you trying to make the best software, or make the most idealistic software? There's a big difference.
I am not at all advocating the use of straight AI slop, but with a policy like that you're opening yourself up for missed bugs, security issues, delayed features, and other things. Especially if you move away from software just because an AI wrote one line of code in it. You can still review the code. If it does the thing right, then use it!
Everyone’s focusing on the AI policy (perhaps because Leah for some reason said this was the most important part of the release), but the UEFI/EDK integration and Chromebook support is a massive step forward. Suddenly Libreboot has multiplied the devices available by many times, and future devices will be much easier to add. This is huge.
some of the comments about ai policy are probably bots too. it's common these days to see political discourse spammed by bots on projects.
True. Also, if this is really Leah, congrats on the release and thanks for all your hard work on this over the years.
yes i am leah rowe
Thank you for your service ma'am. I hold you and your efforts in highest respects.
[flagged]
I will finally be able to use my old chromebook for something.
FYI https://docs.mrchromebox.tech/ in case you weren't aware of it.
They don't seem to like AI.
https://codeberg.org/libreboot/lbmk/src/branch/master/AGENTS...
They REALLY don't like AI, to the point of "wish I didn't need to rely on GCC and GRUB because they have AI in them" levels of hate AI. See this further down the linked page:
>The fork in question, of U-Boot, is heavily AI vibe-coded and thus useless to Libreboot. This is not mere conjecture; in said aforementioned U-Boot mailing list thread, the author stated thus. We will not, in any way, encourage or facilitate use of AI. It has been made clear that several projects Libreboot uses also have some AI-assisted commits; GCC, GRUB, etc. This is not our fault, but it is my intention that I and others will audit these projects in the future, de-slopping them and possibly even forking them entirely for that very reason.
> U-Boot upstream itself is staunchly anti-AI, at least at the time of this release, and much discussion of the above was had on their mailing list.
I saw this as well, which is funny because I've sent plenty of patches written using AI assistance to upstream U-Boot. It was my understanding that U-Boot was mostly following kernel conventions in general as many U-Boot drivers share code(in addition to the actual subsystem maintainers themselves) with their kernel equivalents.
I have learned to treat statements like that as a general red flag.
Someone who acts like a radical on practical issues will almost certainly act like a radical in other areas as well - including areas that matter to me.
There are debates to be had on every technology, but foaming-at-the-mouth kind of talk is a sure sign of a maintainer (and by extension a project) I want nothing to do with.
I respect projects with a no AI policy, but if that policy starts with the words “fuck no”, I’m out of there.
It's possible this type of "radical" behavior is why this project exists in the first place. I'm willing to bet if the Libreboot maintainer(s) were not "radical" they'd just be using a Mac and be content with a closed source bootloader.
Not to say this "radical" behavior is helpful, and I understand why people don't like it.
The timeless debate of separating art from the artist.
In another project (literally: "Untitled") of theirs they are explicit about motivation:
> This is to fight against the trend of crappy github-powered websites that plague the internet.
I don't really see it as a question of "separating art from the artist". That's usually invoked when other aspects of the personal life of the artist has the potential to distract from the art itself.
The parent comment is closer to saying that Picasso wouldn't have been Picasso if in an alternate universe he had avoided controversial/politically charged subjects like the Spanish Civil War and instead appealed to the lowest common denominator for a mainstream commercial audience to sell postcards.
AI in its current form is a commercial product created from stolen labor.
I think there are extremely legitimate reasons to be critical of that part at the very least, and I don't think it's a particularly "radical" stance within Free Software, which is fundamentally conscientious of such questions.
(I don't know whether the Libreboot authors have this particular point in mind.)
All things are built on knowledge that came before it. In what way is AI built on stolen labor?
Has anyone who contributed to producing the wealth of human knowledge, art, code, and literature making up the training data been compensated in any way?
I'm certainly still waiting for my check.
> been compensated in any way?
Yes: https://www.wired.com/story/anthropic-settlement-lawsuit-cop...
Well since the early 18th Century (in a limited way) copyright rights have been given to creators - the overwhelming majority of countries are signatories to international treaties (for the last 100 years) that give creators control of created artistic works (which has been deemed to include program code).
Exploitation of those rights is down to individuals, only since Google came on the scene has any company been able to get away with broad ranging infringement of these rights without censure.
As for your cheque, you have the right to sell your work. No-one is obliged to buy it, of course. You also, under copyright, have the right to make a license to exchange use of your works for binding the users to conditions of release of their derivative/associated works ([F]OSS licensing).
You still have the right to sell it, and if ML models have been trained on those works you have the right to seek compensation.
That already didn't happen, regardless of AI. The thing about freely sharing knowledge is that people take that knowledge and use it to build further knowledge that helps you.
You might say that the fact LLMs are being provided as services by big techs invalidate that. But some people taking the provided knowledge for themselves has also always happened. More importantly, though, is that big tech provided LLMs aren't the only way to access the technology. Open weights and local models exist.
Casting this as “knowledge” betrays a very fundamental lack of knowledge about how LLMs work, or a desire to mislead for whatever reason.
They are not databases of facts. Even if they were, that would still constitute a very serious violation. They aren’t, and they’re not particularly good at serving that function either.
No, LLMs are capable of reproducing things (text, art, other media) based on their training data. Consider being an artist who produced something original that somehow made it into that training set - your chances of actually making a living are now vastly smaller.
The result is the same as if there was no copyright at all: Big companies will just exploit whoever, no art would be produced.
This is a huge issue, and there needs to be a reckoning.
How do you think LLMs work? Because you seem to think they're a big compressed folder of every content they were trained upon, thus a big permanent copyright violation.
You're probably also capable of reproducting some art off the top of your head, your existence doesn't somehow violate copyright.
Besides, anything they're reproducing is the work of very successful artists whose work is widespread. Anyone who had their worked looked at once during training isn't being affected anymore than some artist who never had anything of theirs looked at.
Also, copyright benefits big companies much more than it benefits artists, by the way. Not being able to take any art for free and requiring to pay is a great trade-off for getting a monopoly over anything they pay for. Art would still be made regardless, just like it was before and just like it is right now even when, outside of these big corporations who mostly respect it, copyright is constantly infringed upon.
As I understand it, the argument is about defending intellectual property. Not that strange to hear from the Free Software community, where many people have strong opinions about stuff like licensing. The "Free Imformation" (pirate) community does not fully overlap with the Free Software community.
Well, it was literally stolen in some cases, for starters.
How do you steal knowledge? Knowledge shouldn't be locked up.
One thing is that most if not all models more likely than not were trained on leaked Microsoft and other proprietary code sets.
For projects like WINE or ReactOS that absolutely have to enforce hard firewalls between people with "insider knowledge" to avoid contamination and resulting legal issues, this alone makes the use of AI for coding highly problematic.
I hope everything you make is public domain, then. Otherwise that would be hypocritical.
As it should. Copyright shouldn't exist in the first place.
Most of it is...
If you believe that, then why aren't LLMs free to use for every person on Earth?
If you want anybody to keep making anything, including programmers, artists, and writers, you're going to have to pay them somehow.
Anthropic and OpenAI would not have a product at all without the contributions of thousands of non-consenting humans. Yet they somehow have the gall to charge customers a premium.
Anthropic and OpenAI aren't the only ways to use LLMs. There are plenty of LLMs where you can download the weights and run them locally. There are even others that you can train yourself (see the LLM360 family).
I'm all for it, free AI allotment and universal income is the future.
It’s not a future that’s currently in the cards, no.
It certainly is. People will put up with scraps only for so long before the system comes down.
You're paying for compute?
They refuse to release the models, the training method and, training data. They don't even want you to ask it too many questions lest you teach another ai. Very "open".
Since that's your position -- have you ever produced proprietary software in exchange for money? How do you intend to make up for your unethical actions, if you have?
And, how do you envision the white collar economy, which is entirely built off of selling knowledge, will operate if you get your way?
> And, how do you envision the white collar economy, which is entirely built off of selling knowledge, will operate if you get your way?
It will crash and burn anyway. That kind of economy made sense before the Internet, before mobile phones made it possible to carry the entire world's knowledge in your pocket, only impeded by your ability to efficiently query the database and apply the concepts.
Now with AI? It's even worse. AI not just can efficiently query its own knowledge dataset it was trained upon, but it can also search other databases, apply the knowledge to the input and to put the cherry on the icing, it can do so when operated in natural language. And we're not going to be able to put the genie back into the bottle, not with people being used (and able) to run open weight models on local hardware. OpenAI et al don't have any actual moat behind them, they are just making stuff a bit more convenient and are for now the cheaper way to access compute resources.
For Western countries, the white collar economy is going down the drain, black collar (mining and other forms of dirty resource extraction) is already gone off to places that "compete" by using slave labor, and what remains of blue collar now that industry went the same way as black collar is heavily under threat because so many white collar people are fleeing towards the trades.
> It will crash and burn anyway.
It's unclear from your comment whether you think this is good or bad. You certainly seem to think it is inevitable.
I just want to remind you that nothing here is inevitable - someone is making it happen, and it's only possible because we're not enforcing any of the laws we have about it for some reason, such as copyright. Well, the US isn't enforcing it. The EU is scrambling to come up with reasonable regulations. God knows what China is doing.
I agree that AI, like many other things, would have been an unambiguously great invention in a post-capitalist society, but that's not where we are. They are unbelievably powerful, impossible to control, and fully owned by powerful corporations, so the future looks pretty dystopic for humans who love freedom, art, and science.
But I hope you realize that we have the chance to fight back and get this right. But it does require that you do something.
> It's unclear from your comment whether you think this is good or bad. You certainly seem to think it is inevitable.
I think it is bad, but the consequences are manageable.
> someone is making it happen
We're past that point. Open weights plus local inference is already good enough for an awful lot of use cases. Even if we'd say we ban all AI training, we yeet China and the US from the Internet... it's too late.
> But I hope you realize that we have the chance to fight back and get this right.
The best chance we have is mitigate the consequences. That means first and foremost UBI, backed by actually useful corporate taxation. We need to prepare our economies for the shift towards an economy that requires fewer and fewer people to actually do work. For the moment, the only thing that is relatively safe from both automation and scale efficiencies is some trades (plumbers and electricians), some sections of healthcare and some sections of black collar work. Everything else is under massive threat from AI, automation (have you looked at precision farming?!) or both, and politicians seem like a deer frozen in the headlights.
This is the classic doomer take. Beware of dramatic but simple answers. UBI doesn’t really seem like a sustainable way forward, and at the same time there’s so much we don’t know about the limits.
Don’t play half-court tennis. Other people can read and write too, and social shifts are simultaneous slow, but then sudden.
For example, I’m not seeing a lot of trust in AI as a technology for things that are actually important. It takes one wrong diagnosis from an expensive AI doctor to erode all trust and make everyone prefer a human expert.
It is inevitable. You can ban it but that will only drive it underground. The technology is too good, it will never go away.
What an unbelievably ironic username.
When your project is literally writing a BIOS/ UEFI alternative that exists for the purpose of making base runtime systems open and verifiable and to improve the user’s ability to trust in their system, an opposition to unverified machine generated code seems not only justifiable but sensible. If Leah were not taking steps to mitigate generated code in the project, the project would no longer meet my needs as a Libreboot user.
> Someone who acts like a radical on practical issues will almost certainly act like a radical in other areas as well - including areas that matter to me.
Not using LLMs is a practical choice. The tech doesn't work, and if you let it run amok in your software, your software becomes crap. I try to avoid projects that use LLMs for development as much as I can, because those projects have a much higher likelihood of being a buggy mess.
I find it hard to square treating that like such a red flag while supporting free open source software more generally. The entire ecosystem is rife with ethics/morals driving the initiative. There is a lot of flag planting and red line drawing.
People say AI or not, if the thing is good then it should be used. But if a group actively rejects it (which is valid even if I’m not that militant myself) it compromises the project suddenly?
Personally I treat whinging about other people's ideals as a red flag.
It's libreboot. The entire point of its creation was to take coreboot and rip out the non-libre blobs. Its author is trans (IIRC), which also carries its own very pointed perspective on the world.
Some of it goes too far in my own opinion - I have no problem trusting AMD's final microcode for the Opteron 63xx any differently than the baked-in mask for the 62xx. But regardless, I respect someone who is pushing the whole envelope as they see it. At the very least their effort creates a lot of things I do care about!
Rereading this I realize I might be in the running for the RMS award for technically correct social obtuseness so let me just say that my intent was not to denigrate trans at all, just saying that being trans is itself intrinsically political (as in you're going to be continually clashing with many people's default assumptions, even before the manufactured "culture war" throws gas on the fire). And then furthermore being openly trans is going to impart those politics to the project. Which is most certainly her right, as it is her project! In fact I would say that trying to push a standard of you can contribute your code but stfu about your other ideals is basically cruel.
Well, yes, this is a project to create a fully libre, GPL'd boot chain free of binary blobs. The few users I know stick with aging ThinkPads. I'm not sure what you expected.
You realize that it is their "radicalism" that gives us the Libreboot project to begin with, right?
Hah! Glad I'm not the only one who has this theory https://news.ycombinator.com/item?id=50104114
You make it sound like it's required to have nice, free things.
it absolutely is. appeasement never worked. i put my foot to the floor and that's it. without me, there would be no libreboot at all.
Richard Stallman was a noted centrist for decades after all
"I dunno about this GNU project, man. They seem like a bunch of radicals. Why do they have to impugn all proprietary software like that?"
Yeah its quite the contrast to the recent QGIS core "no AI" policy that very pointedly had no emotive language at all.
We've reached the godwin point quite quickly.
Likewise, it if claims something is “vibe coded” when it clearly is not (such as U-Boot).
Agentic AI has allowed 86Box to make massive strides forward. This includes a substantial contribution to emulate the Intel Inboard 386 that was created by someone who describes himself as a non-programmer. The community is now bigger and stronger, although we’ve had some aches and pains with slop AI generated PRs.
The same goes for Gravity Linux which is booting in my M4 Macs with working GPU and most other peripherals. What used to take years now takes weeks, so we can do much more thorough testing and ultimately have a better end product.
And what happens when the AI bubble pops and the various publishers and IP rights holders come around with their hands out claiming 86Box stole their code ala' SCO?
AI code is legally nebulous and relying on it for a serious project has major risks. Linux has already fended off SCO and has deep pockets to boot. Not all open source projects do.
More projects should be thinking less about the 'moral' issues (yes, they exist) and more about the nuts and bolts province and liability issues of using unverifiable code.
Yes, Linux uses it -because they're big and professional and have allies like IBM and Microsoft (ironies of ironies) on their side.
YOU do not.
> And what happens when [..] the various publishers and IP rights holders come around with their hands out claiming 86Box stole their code ala' SCO?
It won't, for a very simple reason. The same reason why the world isn't in the middle of a nuclear winter right now. Mutually assured destruction.
Every single proprietary software vendor right now is making heavy use of AI code generation, and we're already at a point of no return where pretty much every single commercially relevant codebase out there is contaminated. Do you really think companies will go on a suing spree where by doing so they will open up themselves to being sued for exactly the same reason?
>It won't, for a very simple reason. The same reason why the world isn't in the middle of a nuclear winter right now. Mutually assured destruction.
An interesting take, which I honestly hadn't considered.
The problem remains that it is still a cudgel which larger companies can still use it to shut down smaller companies/free alternatives. I'm thinking a similar situation to the abuse of the DMCA -but we'll find out.
Yes, anti fascism is a radical position these days.
Already in 2022, before the rise of these LLMs, there was a book "Resisting AI: An Anti-fascist Approach to Artificial Intelligence" by Dan McQuillan.
And then:
> The rapid development of digital technologies, including applications of artificial intelligence (AI), is having a significant impact on the social, economic, and political life. From the way individuals communicate and form relationships to the methods by which political discourse and economic transactions occur, digital technologies increasingly mediate our lives and societies. But while presented as milestones in innovation and progress, these technological transformations have also raised concerns about democracy (Nemitz 2018; Kreps and Kriner 2023; Jungherr 2023; Coeckelbergh 2024) and seem to introduce and support mechanisms of control, forms of organization, and ideological patterns that bear striking resemblances to historical fascist phenomena.
https://link.springer.com/article/10.1007/s00146-026-02862-9
Further,
> being plausible but slightly wrong and un-auditable-at scale-is the killer feature of LLMs, not a bug that will ever be meaningfully addressed, and this combination of properties makes it an essentially fascist technology. By “fascist” in this context, I mean that it is well suited to centralizing authority, eliminating checks on that authority and advancing an anti-science agenda.
https://blog.bgcarlisle.com/2025/05/16/a-plausible-scalable-...
Marc Andreessen have called one of the two authors of the very Fascist Manifesto his patron saint. Search for Marinetti on https://a16z.com/the-techno-optimist-manifesto/
A very influential blogger, Curtis Yarvin who have started Dark Enlightenment, the philosophy behind it all regularly posts straight nazi shit. He also has influenced JD Vance and Peter Thiel who have been instrumental in the creation of OpenAI. https://www.wired.com/story/book-excerpt-the-optimist-open-a... More on Dark Enlightenment https://www.cccb.org/en/w/articles/nick-land-or-the-origins-... and https://qz.com/1007144/the-neo-fascist-philosophy-that-under... "a few consistent themes, circling around technology, warfare, feudalism, corporate power, and racism".
I know this won't be popular with this crowd but most of you have become supporters of the rising fascism in the United States by not resisting these LLMs.
US Vice Presidents have a history of inventing important new computer technologies
> Yes, anti fascism is a radical position these days.
Given that these groups are basically composed of communists, if not tankies, that's not surprising. Specially when some of those served as sabotours for the USSR (notorious for its respect for human dignity).
> https://blog.bgcarlisle.com/2025/05/16/a-plausible-scalable-...
This is an incredible piece of anti-intellectualism. Its arguments aren't compelling either: AI might flood with innacurate studies that are difficult to check. Rather than improving said check as to reduce the reliance on trust and improve science as a whole; like the good reactionary the author is, they proppose to ban this particular technology. AI also centralizes authority... because? There is also an irony in them complaining about this theoretical centralization while also trying to preserve the inherent trust-based system in academia, so I suspect they aren't actually committed to proper intellectualism. There is more to complain about this post, but, then, it would hijack this whole comment.
> Marc Andreessen have called one of the two authors of the very Fascist Manifesto his patron saint. Search for Marinetti on [...]
Is it surprising that the guy who founded Futurism, a rather radical and aggressive celebration of innovation and progress, would be mentioned in a movement about technological progress?
> A very influential blogger, Curtis Yarvin who have started Dark Enlightenment, the philosophy behind it all regularly posts straight nazi shit. He also has influenced JD Vance and Peter Thiel who have been instrumental in the creation of OpenAI.
JD Vance hasn't been instrumental to OpenAI's creation, and Peter Thiel is just one of its original investors. Not to mention, NRx is a mess of movement composed of contradictory ideas.
I have noticed you have been using "these LLMs". While I do agree that distancing from OpenAI's models is good (even if it's just on the basis of decentralization), the anti-AI movement either wants to monopolize AI via regulation or ban it outright.
I wouldn't trust their products simply because AI as a tool works and their refusal to use it means their software will be worse off for it.
This is just as bad as blanket hating/rejecting AI without ever trying to use or understand it. I don’t care what tools a craftsman uses so long as the job is done correctly/well. If you can somehow prove that AI is necessary for this that’d be impressive.
You just have to use it for a week to realize it should be a standard part of your tool set like your ide or your debugger.
Nah. It’s not that essential (potentially “yet”)
> de-slopping them and possibly even forking
Or, hear me out, they could instead just improve and contribute the code they think is suboptimal
I put "de-slopping" in the same bucket as "rolling your own crypto" and "let's rewrite done software in Rust just because".
Seems more like clean room RE to me, or general free software stuff. You can't contribute proprietary code to free software, if someone doesn't agree to a license change you have to remove/rewrite their contributions, if you've seen the Windows source code you aren't allowed to contribute to ReactOS (IIRC), etc. LLMs are black boxes with dubious training data and aren't really "you" writing the code directly, so whether you can then license that is questionable. This is especially important for a project like LibreBoot that is trying to liberate machines. The only "easy" solution I see is to abolish copyright, which would be fine. As it is a lot of software is being tainted and the whole world of free software is in danger. It's a shame people gloss over this because they think LLMs are cool or help productivity or whatever. Everything's been irradiated now. Now if someone wants to guarantee a given project had no AI contributions they have to fork a years-old version and likely lose a lot of good non-LLM code in the process. What a waste of time and resources that is. Setting us way back. This needed to be taken way more seriously on day 0.
That's in the previous half of the sentence you removed:
> ...it is my intention that I and others will audit these projects in the future, de-slopping them...
That's not what it implies to me. And let's not pretend that a person with this approach and this manifesto is going to publish their patches to deslop slopped upstream projects.
contributing to projects that have slop in them has legal liability. you'd have to essentially hardfork to remove the slop from git history and such.
for my part, when i do work on this, what i will do is this: re-use existing nuke.list mechanism in lbmk (used already to delete dubious files, e.g. code files with unclear licensing in u-boot) to delete the slopped files. then provide cleaned versions of the same files, with necessary patching re-written as necessary, and files that use it patched to include e.g. foo.deslop.c instead of foo.c - then git history is removed inside tarballs. filter out the changelogs for projects accordingly.
this is the most reasonable way to do it on projects with very little slop. with projects that become almost entirely slop, a hard fork becomes necessary.
So what OS are you building on and running on the machine? https://codeberg.org/ethical-foss/open-slopware#kernels doesn't seem to list anything I would consider a good alternative to Linux, although illumos or 9front becoming the preferred build environment for libreboot would make my inner nerd happy.
currently debian linux, though i have porting lbmk to openbsd on TODO. with a few small exceptions (e.g. recent tmux patch) they're generally quite hostile to use of AI.
the kernel is like 20% AI code now
yes that is why i mention openbsd. i haven't ported lbmk to it yet, was going to months ago. i'll start up that effort again at some point. future releases will one day be built with openbsd, and when that day comes, my reasons to use linux will drop to near zero. i'm an avid fan and user of openbsd - i even maintain a port for it (librewolf web browser)
It's likely significant percentages of the debian userspace is written using some form of AI assistance as well.
How do you use the Linux kernel?
> think for yourself. AI will only ruin your mind and deskill you. here are some articles:
> think AI makes you super hacker? think again: https://arxiv.org/abs/2507.09089 ["Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity"]
> https://blog.codeberg.org/protecting-our-floss-commons-from-...
> consider that the infrastructure for AI is economically unsustainable; when the bubble bursts, and it will, we will suddenly be left with millions of people thoroughly deskilled who will be useless. lots of reeducation needed, and lots of people will just be left behind.
> consider that it is controlled by only a minority of very wealthy individuals; with it, they have unprecedented control over human thought through manipulation of information, and control over what information is presented through filtering (what they call training).
> AI is a threat to the health and democracy of our entire world.
> libreboot will not facilitate its use. AI is the most dangerous technology ever invented by human beings.
...
They have plenty of reasons not to.
> Early-2025 AI
This one especially really just shows they are cherry picking things which support a particular position regardless of the relevance, the difference between Early-2025 AI and Late-2026 AI is absolutely massive. Early 2026 it was reported by Greg Kroah-Hartman that AI bug reports rapidly went from obviously wrong or low quality to being real reports.[0] The kernel now has AI tooling heavily integrated into patch review workflows as well with quite good results.[1]
> They have plenty of reasons not to.
Not really any good reasons however.
[0] https://www.theregister.com/software/2026/03/26/linux-kernel...
[1] https://github.com/sashiko-dev/sashiko
> the difference between Early-2025 AI and Late-2026 AI is absolutely massive
If you only accept papers that were published in reputable venues and analyze LLMs that are less than a month old, you are left with zero useful papers (since writing a useful paper takes longer than a month). It is impossible to have an intellectual discussion if all scientific material is disregarded.
> If you only accept papers that were published in reputable venues and analyze LLMs that are less than a month old, you are left with zero useful papers (since writing a useful paper takes longer than a month).
Why would one look only at formal published scientific papers? Especially if the only publications are known to be quite outdated and highly likely to be no longer particularly relevant anymore.
> It is impossible to have an intellectual discussion if all scientific material is disregarded.
One can certainly have an intellectual discussion even for areas without relevant formal published scientific papers.
Certainly one can do so if the subject matter is in the area of your own expertise, that can even lead to scientific papers being written.
I'm an experienced open source developer leveraging AI tools for linux kernel, u-boot and buildroot contributions in addition to reverse engineering, so clearly this subject is well within my area of expertise regardless of what some outdated published research says.
I probably wouldn't even disagree with that paper's assessment at the time, however the situation is very different now. I have noticed that there is an issue with many experienced open source developers simply having not figured out how to properly leverage AI for open source projects they work on. Probably one reason for this is that productive workflows with AI are often very different from previous workflows and it takes some practice to figure out how to best use AI tools, these best practice can also change over time.
> the difference between Early-2025 AI and Late-2026 AI is absolutely massive
No. No it is not, despite how many times AI proponents like to claim it is.
Unfortunately it is. Thought the whole thing was a scam until six months ago. I think pretty highly of my development skills and have to be objective and admit I am outclassed in debugging and operational tasks now (and it isn't close). Writing non-script code isn't quite there yet (they are bad at abstraction) but I can't confidently predict how much longer that will be the case.
> No. No it is not, despite how many times AI proponents like to claim it is.
Why do you think there isn't a huge difference? I've used AI tools over this time period and the difference is very obvious.
> when the bubble bursts, and it will, we will suddenly be left with millions of people thoroughly deskilled who will be useless. lots of reeducation needed, and lots of people will just be left behind.
what do you mean by this, that costs will become too high and people won't afford it? hardware won't go up in smoke "when the bubble bursts"
People won't be left without AI/LLMs. They'll be left with less value in their investments. The tech is here to stay, the market discovering the actual commercial value of these products won't change that.
Uboot on x86 is like a snowplow on a hatchback. Sure, it can work, but why?
Grub is more extensively used and has a lot of features. Or boot Linux on UEFI using a UKI executable without any bootloader.
How terrible. I'm sitting on multiple AI reverse engineered firmwares that I'm not sure what to do with. I suppose they'll never make it to upstream now. My only option is to ask AI to integrate this stuff.
This conflict has come up a lot in open source lately and the answer has been to just create your own repo and space where you can dump unlimited ai output without needing anyone’s permission. If users find it useful they can use it, and there’s no conflict with maintainers who don’t want it.
Time will tell if these new AI allowed projects like artcraft will prove more popular than fully human curated ones.
The whole point of open source is to get this stuff upstream so that a rich software commons is created.
I'm not sure I want to participate in this space anymore if everyone is just going to have their own isolated forks. I might as well keep my own private version of everything and never deal with maintainers ever again.
There has always been peoples personal forks and random one man hobby projects on Github. I think the Artcraft project is a good example of what's to come. Gimp would have never accepted thousands of commits to convert it in to a replica of Photoshop, and they shouldn't have to. The people who want to go full speed AI programming can run their own project and don't have to bother the Gimp maintainers.
Is it really out of the question to stop using AI and continue contributing upstream to free software projects? It was just a few years ago this was the norm.
That assumes I'd have reverse engineered the firmwares without AI. The truth is I simply don't have enough free time for this.
Some of that reads really funny.
Also
> https://codeberg.org/ethical-foss/open-slopware
Feels more like a software discovery list than an indictment. Thanks for the categorization.
> software discovery list
You're not kidding, I feel like this is better than most of those awesome-X lists that were popular for a while.
Yeah, I saw this "open-slopware" list listing many battle-tested web servers as "permissive AI policy", while one of the alternatives (not all though, some inherit battle-testing) isn't really battle-tested. In fact, I saw entangled codebase and not-so-good coding practices.
This list seems...odd. For instance, the list decries the permissive AI policy of the Linux kernel, of Fedora, of Bazzite/Bluefin/etc...and then includes Secureblue as an alternative...?
Secureblue is a neat project, but as you say, this would make more sense as a generic discovery list. If you want to avoid projects with a permissive AI policy, you should probably take into account that Secureblue is based on bluebuild (from the Bazzite team), Fedora, Linux, etc etc etc.
It is indeed a list, though not everyone will use it the same way!
But will you still feel this way once you try to actually use that software?
They list Firefox, Chromium/Blink, Ladybird and WebKit on their no-no list.
Most people use browsers without major incidents.
Considering how popular many of these projects are, I (and likely many others on HN) already do!
So yes, yes we will.
The funniest section there is https://codeberg.org/ethical-foss/open-slopware#desktop-brow..., which includes... all mainstream browsers? lmao
you arent kidding!
>"think AI makes you super hacker? think again: https://arxiv.org/abs/2507.09089"
not that im particularly pro-ai, but it seems a bit silly to rely on a paper about early-2025 ai when people are using late-2026 ai, given how different they are.
The interesting thing about that paper, I would have said, was the _perception_ gap; people believed that it was making them more productive even when it was doing the opposite. That effect presumably hasn’t gone away.
how are they different? did the algorithms upon which they operate suddenly change?
> how are they different? did the algorithms upon which they operate suddenly change?
Just like most technologies AI algorithms and tools get continuously refined/improved over time.
One thing modern AI has gotten especially good at in the last year is reverse engineering, it went from struggling and frequently making mistakes when analyzing binaries even with say a specialized Ghidra MCP integration to being able to very quickly and accurately analyze even heavily stripped BROM binaries with almost no user guidance or specialized environments. Binaries that would previously take days or weeks to analyze can be analyzed by a modern AI in minutes to hours.
The difference is massive and obvious to anyone who has been working with these tools over these time periods.
Presumably yes. It is safe to assume that algorithms running the model changed, the training procedures changed, and the scale changed. Unfortunately the AI companies refuse to disclose the details of any of this, which is a harm to human knowledge (hopefully a temporary one). Yet the results speak for themselves.
For me it's only been in the last month with Opus 5.5 that I've been seeing code quality I actually respect (though this is with substantial human effort in the loop). Though this experience is from work where I don't have access to non-Anthropic models, nor Fable.
Exactly, same with personal transportation. They still need to be fed oats, right?
>how are they different?
i do not believe you are asking that question in good faith.
and, if you are the leah who wrote the policy, i am positive this would be a completely fruitless discussion as you are obviously entrenched in your beliefs. i am not under the delusion that anything i could say, show, or prove to you would change your mind.
bots are unworthy of any faith.
calling anyone with a different view from you a bot does nothing except make you look immature, and just proves my point: you aren't engaging in good faith, and you don't want to actually have a conversation. i'm happy we both don't need to waste time on one.
good luck with your project. i'll mark this one down as "like the art, dislike the artist"
It takes time to research and publish. It'll always be some slightly old model. Is this argument just a permanent get out of jail free card? Add 10 to each year, will we be seeing that said in the future?
i don't think models will continue to improve at the rate they have been, so eventually it will be less of an issue. there are also questions to answer and studies to be done that aren't as sensitive about the specific model.
however, productivity is sensitive to the model, and models have significantly improved in the last 1.5-2 years. sometimes things move quickly and research gets outdated. ai is not the first field where this has happened.
the research is still valuable, it just has to be acknowledged that it is specific to a point in time. using it for an argument 2 years later just shows that the leah is being intellectually dishonest, and it weakens the overall message.
Okay? Their repo, their rules.
Fork it and use Ai if you have a problem.
There is a respectable way to approach this and this.
The right way is something like, we've been getting spammed with a bunch of PR's from AI that we find difficult to review (because AI allows the average bob who isn't technically minded to make a PR). We have therefore decided not to allow AI submissions.
---
I think the people who have some moral / ideological issue and bring that into projects kind of feel like clowns to me. I agree with Linus that the point is to make something that people find useful, and obviously you want it to be maintainable.
Saying "it makes you dumber etc" is not related to the project, with all due respect it's non of your fucking business how I learn / retain skills.
Free Software was a moral & ideological issue to Stallman.
I feel people have the right to bring any sort of issue into the projects they make. Nobody is forced to participate.
I get it, but everything is ideological -- whether you want to hear about it or not.
In a world where software exerts considerable control and influence over peoples' lives, your ability to insulate yourself from that influence is a priveledge. Software IS ideological. Free software was created so people can have the freedom to share and collaborate on their own terms. There is an extent to which it gets extreme, but without the GPL we would live in an even more monopolistic "free" market hell world than we already do. That's a moral obligation that spawned a practical one. If no one ever took a moral stance because you find it annoying, the world would be an objectively worse place.
I always find these sorts of responses fascinating. “Oh no people are being mean about our lovely robots!!” Like, people are entitled to dislike and be sceptical of LLMs.
No one's claiming they don't have a right to their opinion, But just as they have a right to their opinion, I have the right to say it's silly.
I am not claiming that they must stop following their silly ideas. I am just saying they are silly.
[dead]
Yeah, the (poisoned) AGENTS.md file reads lots like "f*ck AI", which feels like anti-AI dogmatism for me. If I really wanted to use this project, I would use Coreboot instead or fork it to remove proprietary blobs.
I can't use critical software that rejects AI to this degree for privacy reasons.
You can't expect to sufficiently protect your users privacy when not using AI for finding vulnerabilities.
Same thing with Asahi Linux with their highly anti-AI stance. I had a bit of hope when I saw someone was making an LLM assisted fork but it's not the same support.
While MacOS is likely becoming much more secure, Linux distros that reject AI would fall even further behind, if that's possible.
Edit: added critical
Ai and privacy in the same sentence is wierd. You supported your argument, I guess, but unless you are using local ai, the developer would be sacrificing their own privacy.
>the developer would be sacrificing their own privacy
How so?
Something about having a bot controlled by a third party comb your entire codebase and make changes on your behalf strikes me as not the most privacy-friendly approach to software. I know your question wasn't being asked in good faith, but I've answered it honestly.
The end result is open source so bots are combing through it anyway. I’d personally want to sandbox the agent with only access to the code being worked on and none of my personal data, but open source code isn’t my personal data.
It was asked in good faith.
Going by this definition [1]:
>Privacy is the assurance that your data is only seen by the parties you intend to view it.
I’ve been assuming we’re using open source software. When developing open source, you’re putting your code out publicly. Your code is not private data.
As far as the agent acting on your computer, you can easily use Claude Code on the web on Anthropic’s VMs, or just run it in something like UTM or another user on your laptop.
So I don’t see how this is a big loss of privacy. By the definition, you’re intending Anthropic to see your code just as anyone else can.
Ways you could use AI in a non-privacy respecting way would be like letting it read all your personal notes while assuming the AI company will not share them or look at them. I don’t see code as sensitive like this.
Your prompts, pattern of speech, usage patterns, when you work, stuff you say to the LLM besides coding related. Those would be a fair argument for compromising the privacy of the developer. But as long as you consider what you say to be public, then you're intending that data to be out their, so your privacy isn't being compromised. You could make the same argument about developing in the open on GitHub. Anyone can learn about your usage patterns, time zone, etc. It is also fair to say you're giving that AI company more power over you by giving them exclusively that information, while on GitHub it would be information given to everyone. .
[1] https://www.privacyguides.org/en/basics/why-privacy-matters/
You may not, but that entirely comes down to who you're willing to trust. Its certainly not unreasonable for an individual to handle information deemed too sensitive to be allowed to be exfiltrated to a third party. Your use case is not the same as the entire world's.
The comment I replied to was saying that "the developer sacrificing their own privacy". I don't think AI for coding significantly invades developers personal privacy.
The current hope is to eventually upstream the “fork” (Gravity) into the Linux kernel. The Asahi project can continue to run in parallel.
The Linux kernel extensively uses AI now; most lkml submissions have an Assisted-by: LLM tag.
The Assisted-by: tag is no longer requirement because LLM assistance is the default.
Robot friends enable you to get much more done faster and more accurately, that if you call yourself an engineer you're a damn fool not to make use of them.
> The Assisted-by: tag is no longer requirement
Do you have a link? Because this contradicts the kernel documentation.
https://docs.kernel.org/process/coding-assistants.html
Outdated. It was agreed at the Kernel Developers Summit in Prague to deprecate the tag, and Linus is on board: https://www.xda-developers.com/linus-calls-banning-ai-genera...
good!
Unfortunately this is bad for one of my projects, a slop detector extension... it assumes that AGENTS.md be short in size to say stuff like "You may view and use this code but not edit or commit anything"... this rant is long enough that it classifies as slop itself!
Huh, good to see sane people still exist. Good on them.
[flagged]
We really don’t need to fit everything into simplistic US political memes.
I find this hilarious since I'm actively having my clankers work on porting EDK2 to an SoC that's used in a Chromebook, though not the ones mentioned here.
Are you publishing anywhere that I can follow? In particular, it would be really helpful if you could publish the prompts and workflow, so that the rest of us can more quickly use it to replicate your work on other machines:)
(Signed, the owner of several Chromebooks in need of better firmware)
Cool, that's why open source projects get left behind. The attitude
I think it's better that some people have strong opinions. For example, the world would suck if everyone was rms, but he has nevertheless had a positive impact on the world imo.
RMS is extremely radical but he's also reasonable and consistent unlike these guys. There's also GNU Guix which recently banned AI-assisted contributions [1] due to copyright infringement concerns and licensing doubts. This is an entirely reasonable take for a hardcore free software codebase, it's literally built around this political concept.
On the contrast, libreboot's stuff reads like a parody or a shitpost. It's hard to believe they're being entirely serious.
[1] https://codeberg.org/guix/guix/pulls/11407
The libreboot maintainer has always seemed irrational, but then again, a rational person would have just left their proprietary bios alone and we would never get projects like this.
> Cool, that's why open source projects get left behind. The attitude
They're making a political choice. It's a free country, and more power to them.
It's not a choice that I necessarily agree with, but one that may - who knows - prove prescient 5 years down the road.
No one can predict the future, that's what politics are partly about and the reason why diversity of opinion of choice matters.
Do you have a crystal ball?
i don't live in a free country. the UK has enacted several laws impeding fundamental civil liberties in recent years. rolling back on several human rights laws, now they're even removing the right to trial by jury, etc. it's quite troubling actually.
I'd rather have the project that works, rather than one that has the correct political stance.
You can look at what being achieved right now with AI, GoW3 (an absolute master piece BTW) on the PS4 Remaster can be played on Linux at 120FPS thanks to some decompilation/reverse engineering framework with AI.
I am generally also fed up of politics (whatever they are) forced into projects.
Also AI doesn't make you stupid. It is bullshit, he has cherry picked information to sooth his own biases.
> I'd rather have the project that works
Then more often than not, you want the project that eschews LLMs. They can't program worth a damn.
I find that they program very well.
It is bizarre having conversations with some people. There is a reverse engineering framework that had God of War 3 remaster running natively on Linux at 120 fps and all this tech is AI-assisted. A friend of mine was demonstrating it running to me last night, and then people like yourself tell me AI can't do anything when you can download and run it yourself.
I just had it write a bunch of script hooks for gamemoderun and at work I had it do probably a week's worth of work in an afternoon earlier today. It also smoke-tested everything for me.
It is utterly bizarre, and I think you are just denying reality.
> I'd rather have the project that works,
You didn't get my point. It might very well be that the only project that lets you actually freely boot a computer in 5 years will be this one, specifically because they did not let an AI touch it.
In that regard, it will be the only project, like you put it, "that works".
Like I said, I don't necessary subscribe to their view, but the diversity of opinion matters, it helps with minor things like, oh, survival.
What you are saying is utter nonsense. I really don't know where to start.
Why would AI matter? If you say legality, I really don't care that much about copyright law (I run a torrent box 24/7), and I would just git clone the repo and put it on tor. That is what people have been doing with plenty of projects that are more legally dubious.
Also, diversity of opinion doesn't necessarily help with survival either. I swear people just say things because other people have said it.
But they're doing it under the banner of the libre software movement, whose raison d'être is free (as in liberty) software. Having political strings attached is antithetical to the movement.
Libre software is a political movement. And it makes sense for libre contributors to be highly skeptical of a tool that may or may not violate copyright, profits from the work of other libre codebases and is controlled by tech giants.
I think there is a deeper point regarding software freedom specifically - if a person is unable to understand the source code of something, then they are less able to modify it. It's why the definition of libre software centers around source code, even though modifying binaries is entirely possible. Despite being able to have a libre vs non-libre disassembler (analogous to proprietary vs open weight AI models), needing to use a tool at all diminishes your own ability to make changes to the code.
Now having said that, AI also increases software freedom in that it enables many people who aren't otherwise able to understand codebases (or even write simple code) to create/modify their own software. When we're talking about AI in the context of libre software, this is the constructive aspect we should be focusing on! Well defined plugin APIs that we know people are going to lay slop upon, etc.
But still, this doesn't have much bearing on a well-established fundamental systems project that needs to remain generally concise and human-understandable.
> Having political strings attached is antithetical to the movement.
I believe you may have a rather narrow notion of what politics are.
In my book, politics is the art of making choices when your ability to predict the future is limited, and therefore the willingness, courage and fortitude to make choices under uncertainty.
Under that definition, they're making a choice, one that materially and obviously can't be backed by sound reasoning or logic (because predicting the future is hard), that's politics.
The lack of political literacy and self-awareness in this thread is embarrassing.
Someone can always fork it if they find themselves sufficiently outraged.
Agree in general, though with Libreboot the pedantry is more or less the point of the project. You fork a project and put 'libre' in the name, I personally see that as a statement of dedication and expect you to stick to those guns. The experiment isnt much of an experiment otherwise.
Coreboot still exists.
Libreboot is not a fork of Coreboot.
It's a distribution. That is, they build and distribute firmware images. Coreboot just ships source.
Coreboot's docs lists the variety of distributions: https://doc.coreboot.org/distributions.html
Well, maybe you can bring it into the future with your favorite LLM!
When the alternative is being surrounded by rabid AI enthusiasts who absolutely lose their shit whenever anyone says "no" to them, being left behind is a great choice.
And yet people who are training models need more and more open source software to feed the beast. Quite the conundrum.
I can't be the only one who thinks the way they handled that is incredibly disrespectful right?
It's their own business to disallow AI contributions. But instead of adding that to the Code Of Conduct they chose to knowingly misuse a file to target a group of people they don't like.
If I want to use AI on my copy of that repo, I have to delete that file and consider if the developer has added any more of what is (by intent) basically malware to the source code. Sure the file is just an unhinged rant, but now I have to wonder if they attempted to put some prompt injection in the repo.
“If I want to crawl a website with a robots file that says ‘fuck off’, I have to ignore the robots file. If I want to use AGPL code as part of a commercial web service without creating my source, I have to ignore the license.”
Like, c’mon now. This is far from the weirdest open source restriction.
I don't see how any of those things are similar, I think you've missed my point. Those hypothetical authors are using robots and license files correctly.
Would you be OK if a repo had Visual Studio files that were designed to crash visual studio if they are used, just because the maintainer hates Microsoft? That's what this feels like, a deliberate malicious abuse of the intent that adding agents.md would otherwise imply.
Anti-AI is rationally and logically incorrect, because it focuses on inputs, rather than outputs.
The Linux kernel has always been focused on outcomes and outputs, which is the right focus. Here's Linus: https://x.com/LundukeJournal/status/2110524360377458786
Outputs: Code quality and long term maintenance matters. Outputs: Serving users with correctness, security and performance matters.
Focus on "purity of input" is a faith-based political filter, not engineering.
Linus's take is just the same nonsensical false equivalence I've seen a million times. He sure has fallen a long way since his glory days.
Someone should ask him if he's also planning on no longer reviewing patches to the kernel and just accepting them unconditionally, since AI is no different from a compiler after all, and nobody reviews compiler outputs.
>Linus's take is just the same nonsensical false equivalence I've seen a million times. He sure has fallen a long way since his glory days.
After the SCO mess, I'm very shocked he didn't reject AI out of hand simply for the provenance and liability issues.
the take is don't ban tools, and technically the compiler outputs are tested by the compiler projects + the automated testing infra
> the take is don't ban tools
Yes, and that take is bad, because AI is not a tool in the traditional sense. In the software world, tools are generally expected to be deterministic. Would Linus use a compiler that is known to have a random chance to output the wrong instructions? Or an editor that has a random chance to swap something you typed for something else?
> compiler outputs are tested by the compiler projects + the automated testing infra
And AI outputs are tested by the AI companies and the various testing suites that are run on the kernel. If they're really the same like Linus seems to believe, there's no reason to ever read AI generated patches before merging them.
> tools are generally expected to be deterministic
Says who?
You say "faith-based political filter," I say "standing on principles." Your valuation of outcomes and outputs over inputs and intent is also a faith-based political filter, in that they make the value judgement that outputs are the only thing that matter when shipping a product or managing a project.
Wow, this is awesome! This relates to work I've been doing with the [Heads](https://osresearch.net/) project. Specifically I got a PR merged for the first ChromeOS Port using MrChromebox's code as an upstream dependency: https://github.com/linuxboot/heads/pull/2133
I'm now working on the 2nd ChromeOS device in this PR right now: https://github.com/linuxboot/heads/pull/2225 And there is a 3rd port by someone else: https://github.com/linuxboot/heads/pull/2219
I'm surprised Libreboot was able to add all the ChromeOS devices at once, Just adding the 2 ports for Heads hasn't been easy, although I'm not an expert when it comes to Coreboot.
Anyway, I love seeing more support for Chromebooks, they are great devices. Funny enough my Acer Chromebook (Kano) is now more open and more secure than my Framework 13s.
It’s great to see this happening at the same time—Heads has some important features that Libreboot lacks.
It may be harder to port it if those features require neutered (rather than simply removed) blobs. Maybe that’s why the Libreboot port was easier? Not sure, I’m aware of the benefits that Heads provides but I don’t know much about the internals.
Either way, thanks for your work on this.
It's refreshing to see serious open-source projects adopting a strict policy on LLM-assisted contributions.
Besides considering frontier LLMs a serious threat to humanity, I find LLM-assisted programming simply...not fun! I tolerate only AI-assisted bug reports, as they are genuinely useful.
I guess I'm AI centrist lol but to me, I can see a way around the 'threat' to humanity issue simply by going local and avoiding the use of datacenter ai. That might reduce the quality of code but running it on your own computer is a lot better -at least IMO.
My issues, personally, are the IP rights issues. What was this shit trained on? When the bubble pops (as I've said a lot in this thread -I know) who's going to shake me down?
> Imported all 233 currently supported Chromebooks from the MrChromebox coreboot distro. Baytrail (Kip) tested - others need testing. Source-only in this release (must build from source, no binaries provided).
Nice. I think I loaded MrChromebox onto my Google Pixel some years ago. I should see if it can get Libreboot now.
Getting Libreboot running on a Pixel would be a pretty wild second life for a phone that was never meant to be a Chromebook.
Wrong Pixel.
https://en.wikipedia.org/wiki/Chromebook_Pixel
Which is sad; a Pixel phone using libreboot + EDK2 to offer a "normal" boot path to Linux would be sweet.
Also the Pixelbook :)
This cross-industry corporate naming trend, where each company picks a word or two and then creates a convoluted web of products with overlapping functionality and intended audiences, and expects us to, like, follow along... This trend needs to end. Microsoft Copilot 365 for Azure, meet Gemini Pixel Colab for business, and kindly ritually immolate each other.
>"Libreboot is free, open source boot firmware replacing proprietary BIOS/UEFI firmware on supported Intel/AMD and ARM64 hardware." [...] MrChromebox supports nearly every Chromebook..."
This is great news! Finally an open-source UEFI replacement which can work on nearly every Chromebook, including used/old Chromebooks!
Now, I haven't reviewed the source... it may still contain binary blobs for various drivers; I don't know. If it doesn't, then that's absolutely awesome! But if it does, then remember that having some of the UEFI firmware code open-source is a huge step forward, and it's only a matter of time before those closed-source binary blobs are reverse engineered (possibly by AI) and/or rewritten from scratch entirely, as open-source codebases...
So, a good step forward for open-source, right-to-repair, understandability of the boot chain, the Chromebook hardware ecosystem (used and new), and tranparency in general!
If it runs on Chromebooks does that mean Chromebooks can be made into generic UEFI laptops?
This is interesting given Googlebooks actually use the same boot process as Chromebooks, they just chain into aboot.
Yes. Even before this you could use MrChromebox's UEFI Coreboot on a lot of ChromeOS devices. You can run Linux or even Windows on them.
For Intel/AMD Chromebooks, yes.
I'm on a libreboot-only infrastructure and can recommend. It's easy to do and you dont need to keep track of how to configure coreboot.
It just works and that's great.
Hopefully, this will end up in its foundation or even the linux one, with most motherboard manufacturers behind it.
But... it should be binary specifications with "human-like" assembly implementations since you have to remove the "compiler" issue from the SDK.
Nice! I took a 250eu chromebook, turned it into my Linux laptop perfect little machine. A bit slow but 4 cores + 8gb ram will do.
If you want something with more power look into the Dell Latitude 5400 Chromebook Enterprise or the Dell latitude 5300 2-in-1 Chromebook Enterprise. Both of these devices have DDR4 user replaceable memory.
There is also the Framework 13 Chromebook edition, but that seems almost impossible to find now-a-days.
Oehhhhhhhhhhh Might check that out.
Impressive - and I have several "Edgar" class Acer chromebooks with full aluminum cases and 1920x1080p sharp screens - will be testing this out.
I thought Chromebooks natively came with Coreboot support.
there is upstream code but it's not well-integrated, and google's tree is largely optimised for their own payload called Deptchcharge. mrchromebox tweaks coreboot for use with edk2 so you can use normal operating systems, and the modifications that they make are extensive, e.g. mechanism in EDK2 to update EC firmware at boot time. mrchromebox is, for our purposes, a coreboot distro or indeed its own upstream. libreboot takes mrchromebox and makes several of its own modifications. "just use coreboot" is easier said than done :)
This is Libreboot, not Coreboot. Libreboot removes binary blobs and Intel ME. It’s a freedom-oriented distribution of Coreboot.
I realize that, but what's the selling point of Libreboot on Chromebooks when Coreboot is already supported, and it merely merged in support from Coreboot?
Assurance that binary blobs and Intel ME support has been properly removed. It’s possible to do this with Coreboot if you know what you’re doing, but this is an off the shelf configuration.
Nothing has changed about modern Intel CPUs requiring binary blobs. Libreboot is no longer blobless. Canoeboot is blobless. Libreboot is like decaf. It's not caffeine free, it's decreased caffeine.
You’re correct, I guess haven’t been following this over the last few years. It looks like blobs are now merely minimized and ME is neutered with me_cleaner. So the only difference versus stock coreboot is that these minimizations are now the default configurations.
Old Libreboot is now apparently Canoeboot, same dev but different project name.
Neat project. I find their new anti-AI policy to be interesting, but not at all useful to anyone. There are no objective arguments. No evidence of poor code quality in late 2026. No examples of why AI can't help the project. Just a rant.
I am very glad to see people with opinions that flow into what they do, but the question becomes: are you trying to make the best software, or make the most idealistic software? There's a big difference.
I am not at all advocating the use of straight AI slop, but with a policy like that you're opening yourself up for missed bugs, security issues, delayed features, and other things. Especially if you move away from software just because an AI wrote one line of code in it. You can still review the code. If it does the thing right, then use it!
[dead]