I think this is a reasonable decision (although maybe increasingly insufficient).
It doesn't really matter what your stance on AI is, the problem is the increased review burden on OSS maintainers.
In the past, the code itself was a sort of proof of effort - you would need to invest some time and effort on your PRs, otherwise they would be easily dismissed at a glance. That is no longer the case, as LLMs can quickly generate PRs that might look superficially correct. Effort can still have been out into those PRs, but there is no way to tell without spending time reviewing in more detail.
Policies like this help decrease that review burden, by outright rejecting what can be identified as LLM-generated code at a glance. That is probably a fair bit today, but it might get harder over time, though, so I suspect eventually we will see a shift towards more trust-based models, where you cannot submit PRs if you haven't been approved in advance somehow.
Even if we assume LLMs would consistently generate good enough quality code, code submitted by someone untrusted would still need detailed review for many reasons - so even in that case it would like be faster for the maintainers to just use the tools themselves, rather than reviewing someone else's use of the same tools.
For well-intended open source contributions using GenAI, my current rules of thumb are:
* Prefer an issue over a PR (after iterating on the issue, either you or the maintainer can use it as a prompt)
* Only open a PR if the review effort is less than the implementation effort.
Whether the latter is feasible depends on the project, but in one of the projects I'm involved in it's fairly obvious: it's a package manager where the work is typically verifying dependencies and constraints; links to upstream commits etc are a great shortcut for reviewers.
Unfortunately, LLMs generate useless word salad and nonsense even when working on issues text, you absolutely have to reword the writing from scratch otherwise it's just an annoyance and a complete waste of time. Even a good prompt doesn't help this all that much since it's just how the tool works under the hood: it doesn't have a goal of saying anything specific in the clearest possible way and inwardly rewording it until it does, it just writes stuff out that will hopefully end up seeming at least half-coherent. And their code is orders of magnitude worse than even their terrible English prose.
I don’t think you’re being serious. Claude and GPT regularly write programs that are way better than what I would’ve written. Maybe you haven’t used a decent harness or a model released in the last year? It’s usually verbose, whereas I would try the simplest thing that could possibly work. However, it can knock out but would have taken me multiple weekends in a few minutes. The value proposition isn’t even close.
It’s fine to write things by hand, in the same way that there’s nothing wrong with making your own clothing with a sewing machine when you could have bought the same thing for a small fraction of the value of your time. Or in the same fashion, spending a whole weekend, modeling and printing apart, you could’ve bought for a few dollars. I think we need to be honest about differentiating between the hobby value of writing programs versus the utility value of programs. Redox is a hobby project, and, while it’s very cool, I’m not sure it has a strong utility proposition. Demanding that code be handwritten makes sense to me for the maintainer because the whole thing is just for fun anyway. There isn’t an urgent need to RIIR Linux. I would not apply this approach to projects where solving the problem is more important than the joy of writing the solution.
> Claude and GPT regularly write programs that are way better than what I would’ve written
Is that really true? Like, if you took the time to plan it carefully, dot every i, cross every t?
The way I think of LLM's is as "median targeters" -- they reliably produce output at the centre of the bell curve from their training set. So if you're working in a language that you're unfamiliar with -- let's say I wanted to make a todo list in COBOL -- then LLM's can be a great help, because the median COBOL developer is better than I am. But for languages I'm actually versed in, the median is significantly worse than what I could produce.
So when I hear people say things like "the clanker produces better programs than me", what I hear is that you're worse than the median developer at producing programs by hand.
no. I'm a pretty skilled programmer and I definitely have to intervene and fix an architectural problem here and there, or gently chastise the LLM for doing something dumb. But there are also many cases where the LLM has seen something that i completely missed or just hammered away at a problem enough to get a solution that is correct that I would have just given up on earlier.
The clanker can produce better programs than me because it will just try shit that I would never have tried, and it can fail more times than I can in a given period of time. It has specific advantages over me.
A lot of computer users are domain experts in something like chemistry or physics or material science. Computing to them is just a tool in their field, e.g. simulating molecular dynamics, or radiation transfer. They dot every i and cross every t _in_their_competency_domain_, but the underlying code may be a horrible FORTRAN mess. LLMs potentially can help them write modern code using modern libraries and tooling.
My go-to analogy is assembly language programming: it used to be an essential skill, but now is essentially delegated to compilers outside of some limited specialized cases. I think LLMs will be seen as the compiler technology of the next wave of computing.
It can certainly be true for several reasons. Even in domains I'm familiar with, often making a change is costly in terms of coding effort.
For example just recently I updated a component in one of our modules. The work was fairly rote (in this project we are not allowed to use LLMs). While it was absolutely necessary to do the update here, it was beneficial to do it everywhere else. I didn't do it in other places because I couldn't justify spending the effort.
There are two sides to this - with LLMs, housekeeping becomes easy and effortless, but you often err on the side of verbosity because it costs nothing to write.
But much less thought goes into every line of code, and I often am kinda amazed that how compact and rudimentary the (hand-written) logic is behind some of our stuff that I thought would be some sort of magnum opus.
When in fact the opposite should be the case - every piece of functionality you don't need right now, will be trivial to generate in the future, so the principle of YAGNI applies even more.
I can agree with that. So essentially: "Claude and GPT regularly write programs that are way better than what I would’ve written given the amount of time I was willing to spend."
That's correct, because most of the cost of code is not the development but rather the subsequent maintenance, where AI can't help. Verbose, unchecked AI slop becomes a huge liability over time, you're vastly better off spending those few weekends rewriting it from scratch.
if it wasn't so maddening it would be funny when you literally have to tell it to slow down, focus and think. My tinfoil hat suggests this is intentional to make me treat it like a real, live junior dev!
I feel like every person stating things of this nature are literally not able to communicate effectively (though this is not a barrier anymore, you can get a dog to vibe code games with the right workflow, which to me seems like quite an intellectual thing to be able to do.
Despite that, you will make this argument when trying to use copilot to do something, the worst model in the entire industry.
If an AI can replace you at your job, you are not a very good programmer.
Copilot isn't a model. Currently it's giving me a choice of 15 different models. By all evidence, AI is nowhere close to replacing me, but to hear other people tell it, it is weeks or maybe months away.
> If an AI can replace you at your job, you are not a very good programmer.
Me and millions of other local yokel programmers who work in regional cities at small shops, in house at businesses, etc are absolutely COOKED. No I cant leet code, no I didnt go to MIT, no I dont know how O(n) is calculated when reading a function. I can scrap together a lot of useful business stuff but no I am not a very good programmer.
>no I dont know how O(n) is calculated when reading a function
This is really, honestly not hard. Spend a few minutes reading about this, or even better, ask a LLM to explain it to you and clear your misconceptions if regular blog posts don't do it for you. This is one of the concepts that sounds scarier than it is.
edit: To be clear there are tough academic cases where complexity is harder to compute, with weird functions in O(sqrt(n)) or O(log(log(n)) or worse, but most real world code complexity is really easy to tell at glance.
It does generate word salad (and usefulness depends on the person reading it). If both the writer and the reader share a common context, there's a lot that can be left out (the extreme version is military signal). An SOS over the radio says the same thing as "I'm in a dangerous situation, please help me if you can" but the former is way more efficient. LLMs tend to prefer the latter.
But all work isn't done by LLMs at the moment and we can't be sure that it will be so the question is ridiculous.
Maybe one day it will be.. And then people can reevaluate their stance then. Until that time, it's entirely reasonable to hold the position that you just don't
This is especially true with how LLM generated code may affect licensing and other things. There's a lot of unknowns there and it's entirely reasonable to not want to risk your projects license over some contributions.
I use them all the time at work because, rightly or wrongly, my company has decided that's the direction they want to go.
For open source, I'm not going to make that choice for them. If they explicitly allow for LLM generated code, then I'll use it, but if not I'm not going to assume that the project maintainers are willing to deal with the potential issues it creates.
For my own open source projects, I'm not interested in using LLM generated code. I mostly work on open source projects that I enjoy or in a specific area that I want to learn more about. The fact that it's functional software is great, but is only one of many goals of the project. AI generated code runs counter to all the other goals I have.
Basically all of my actual programming work has been done by LLMs since January. My team actually demoed a PoC last week to hook up Codex to our Slack channel to become our first level on-call, and in the case of a defect (e.g. a pagerduty alert, or a question that suggests something is broken), go debug, push a fix for review, and suggest any mitigations. Prior to that, I basically pushed for my team to do the same with copy/paste to a prompt so we could iterate on building its debugging skills.
People might still code by hand as a hobby, but I'd be surprised if nearly all professional coding isn't being done by LLMs within the next year or two. It's clear that doing it by hand would mostly be because you enjoy the process. I expect people that are more focused on the output will adopt LLMs for hobby work as well.
> It's clear that doing it by hand would mostly be because you enjoy the process.
This is gaslighting. We're only a few years into coding agents being a thing. Look at the history of human innovation and tell me that I'm unreasonable for suspecting that there is an iceberg worth of unmitigated externalities lurking beneath the surface that haven't yet been brought to light. In time they might. Like PFAS, ozone holes, global warming.
Ultimately you always have to trust people to be judicious, but that's why it doesn't make any changes itself. Only suggests mitigations (and my team knows what actions are safe, has context for recent changes, etc). It's not entirely a black box though. e.g. I've prompted it to collect and provide a concrete evidence chain (relevant commands+output, code paths) along with competing hypotheses as it works. Same as humans should be doing as they debug (e.g. don't just say "it's this"; paste your evidence as you go and be precise about what you know vs what you believe).
That's sounds like the perfect recipe for turning a small problem into a much larger one. 'on call' is where you want your quality people, not your silicon slop generator.
I say let people hold this stance. We, agentic coders, can easily enough fork their project and add whatever the features or refinements we wanted, and use that fork for ourselves, but also make it available for others in case other people want to use it for the extra features and polish as well. With AI, it's very easy to form a good architectural understanding of a large code base and figure out how to modify it in a sane, solid way that matches the existing patterns. And it's also very easy to resolve conflicts when you rebase your changes on top of whatever is new from upstream. So, maintaining a fork is really not that serious of and endeavor anymore. I'm actually maintaining a fork of Zed with several additional features (Claude Code style skills and slash commands, as well as a global agents.md file, instead of the annoying rules library system, which I removed, as well as the ability to choose models for sub-agents instead of always inheriting the model from the parent thread; and yes, master branch Zed has subagents! and another tool, jjdag)
That seems like a win-win in a sense: let the agentic coders do their thing, and the artisanal coders do their thing, and we'll see who wins in the long run.
>> but also make it available for others in case other people want to use it for the extra features and polish as well.
this feels like the place where your approach breaks down. I have had very poor results trying to build a foundation that CAN be polished, or where features don't quickly feel like a jenga tower. I'm wondering if the success we've seen is because AI is building on top of, or we're early days in "foundational" work? Is anyone aware of studies comparing longer term structural aspects? is it too early?
> We, agentic coders, can easily enough fork their project
And this is why eventually you are likely to run the artisanal coders who tend to do most of the true innovation out of the room.
Because by and large, agentic coders don't contribute, they make their own fork which nobody else is interested in because it is personalized to them and the code quality is questionable at best.
Eventually, I'm sure LLM code quality will catch up, but the ease with which an existing codebase can be forked and slightly tuned, instead of contributing to the original, is a double edged sword.
"make their own fork which nobody else is interested in because it is personalized to them"
Isn't that literally how open-source works, and why there's so many Linux distros?
Code quality is a subjective term as well, I feel like everyone dunking on AI coding is a defensive reaction - over time this will become an entirely acceptable concept.
For a human to be able to do any customization, they have to dive into the code and work with it, understand it, gain intuition for it. Engage with the maintainers and community. In the process, there's a good chance that they'll be encouraged to contribute improvements upstream even if they have their own fork.
Vibe coders don't have to do any of this. They don't have to understand anything, they can just have their LLMs do some modifications that are completely opaque to the vibe coder.
Perhaps the long term steady state will be a goldilocks renaissance of open source where lots of new ideas and contributors spring up, made capable with AI assistance. But so far what I've seen is the opposite. These people just feed existing work into their LLMs, produce derivative works and never bother to engage with the original authors or community.
> just like almost all transportation is done today via cars instead of horses.
That sounds very Usanian. In the meantime transportation in around me is done on foot, bicycle, bus, tram, metro, train and cars. There are good use cases for each method including the car. If you really want to use an automotive analogy, then sure, LLMs can be like cars. I've seen cities made for cars instead of humans, and they are a horrible place to live.
Signed, a person who totally gets good results from coding with LLMs. Sometimes, maybe even often.
As someone who enjoys working with AI tools, I honestly think the best approach here might be bifurcation.
Start new projects using LLM tools, or maybe fork projects where that is acceptable. Don't force the volunteer maintainers of existing projects with existing workflows and cultures to review AI generated code. Create your own projects with workflows and cultures that are supportive of this, from the ground up.
I'm not suggesting this will come without downside, but it seems better to me than expecting maintainers to take on a new burden that they really didn't sign up for.
even if this was true or someday will be (big IF), is it worth looking for valid counter workflows? example: in many parts of the US and Canada the Mennonites are incredibly productive farmers and massive adopters of technology while also keeping very strict limits on where/how and when it is used. If we had the same motivations and discipline in software could we walk a line that both benefited from and controlled AI? I don't know the answer.
That would only be a world where the copyright and other IP uncertainties around the output (and training!) of LLMs were a solved and known question. So that's not the world we currently live in.
The ruling capital class has decided that it is in their best interest for copyright to not be an obstacle, so it will not be. It is delusional to pretend that there is even a legal question here, because America is no longer a country of laws, to the extent that it ever was. I would bet you at odds of 10,000 to 1 that there will never be any significant intellectual property obstacles to the progress of generative AI. They might need to pay some fines here and there, but never anything that actually threatens their businesses in the slightest.
There clearly should be, but that is not the world we live in.
The problem was already there with lazy bug reports and inflammatory feature requests. Now there is a lazy (or inflammatory) accompanying code.
But there were also well-written bug reports with no code attached due to lack of time/skills that now can potentially become useful PRs if handled with application and engineering knowledge and good faith and will.
I don't think it really is - drive-by changes have been a net burden on maintainers long before LLMs started writing code. Someone who wants to put in the work to become a repeat contributor to a project is a different story.
I've gotta disagree with you here - it's not uncommon for me to be diving into a library I'm using at work, find a small issue or something that could be improved (measurably, not stylistically), and open a PR to fix it. No big rewrites or anything crazy, but it would definitely fit the definition of "drive by change" that _thus far_ has been welcomed.
How to differentiate between a drive-by contribution and a first contribution from a potentially long-time contrubutor.
And I would say especially for operating systems if it gets any adoption irregular contributions are pretty legit. E.g. when someone wants just one specific piece of hardware supported that no one else has or needs without being employed by the vendor.
This sounds complicated in theory, but it's easier in practice.
Potential long time contributor is somebody who was already asking annoying questions in the irc channel for a few months and helped with other stuff before shooting off th e PR. If the PR is the first time you hear from a person -- that's pretty drive-by ish.
Sounds like a better way to make sure you have to be part of a clique to get your changed reviewed. I’ve been a long-time bug fixer in a few projects over the years without participating in IRC. I like the software and want it you work, but have no interest in conversing about it at that level, especially when I was conversing about software constantly at work.
I always provided well-documented PRs with a narrow scope and an obvious purpose.
Why would I ask annoying questions when I can identify, reproduce, pinpoint the bug, locate it in code, and fix it? Doing it alone should make it clear I don't need to ask to understand it. And why would I be interested in small talk? Doubt many people are when they patch up their work tools. It's a dispassionate kind of kindness.
Not to mention LLMs can be annoying, too. Demand this, and you'll only be inviting bots to pester devs on IRC.
> Why would I ask annoying questions when I can identify, reproduce, pinpoint the bug, locate it in code, and fix it?
Because if the bug is sufficiently simple that an outsider with zero context to fix, there's a non-zero chance that the maintainers know about it and have a reason why it hasn't been addressed yet
i.e. the bug fix may have backwards-compatibility implications for other users which you aren't aware of. Or the maintainers may be bandwidth-limited, and reviewing your PR is an additional drain on that bandwidth that takes away from fixing larger issues
Hard disagree. Drive by's were the easiest to deal with, and the most welcome. Especially when the community tilted more to the side of non-amateurs and passionate people.
> Even if we assume LLMs would consistently generate good enough quality code, code submitted by someone untrusted would still need detailed review for many reasons
Wait but under that assumption - LLMs being good enough - wouldn't the maintainer also be able to leverage LLMs to speed up the review?
Often feels to me like the current stance of arguments is missing something.
> Wait but under that assumption - LLMs being good enough - wouldn't the maintainer also be able to leverage LLMs to speed up the review?
This assumes that AI capable of writing passable code is also capable of a passable review. It also assumes that you save any time by trusting that review, if it missed something wrong then it's often actually more effort to go back and fix than it would've been to just read it yourself the first time.
A couple weeks ago someone on my team tried using the experimental "vibe-lint" that someone else had added to our CI system and the results were hilariously bad. It left 10 plausible sounding review comments, but was anywhere from subtly to hilariously wrong about what's going on in 9/10 of them. If a human were leaving comments of that quality consistently they certainly wouldn't receive maintainer privileges here until they improved _significantly_.
It was maybe not quite clear enough in my comment, but this is more of a hypothetical future scenario - not at all where I assess LLMs are today or will get to in the foreseable future.
So it becomes a bit theoretical, but I guess if we had a future where LLMs could consistently write perfect code, it would not be too far fetched to also think it could perfectly review code, true enough. But either way the maintainer would still spend some time ensuring a contribution aligns with their vision and so forth, and there would still be close to zero incentive to allow outside contributors in that scenario. No matter what, that scenario is a bit of a fairytale at this point.
You can not trust the code or reviews it generates. You still have to review it manually.
I use Claude Code a lot, I generate a ton of changes, and I have to review it all because it makes stupid mistakes. And during reviews it misses stupid things. This review part is now the biggest bottleneck that can't yet be skipped.
An in an open source project many people can generate a lot more code than a few people can review.
This is not even about capabilities but responsibility. In an open source context where the maintainers take no responsibility for the code, it's perhaps easier. In a professional context, ultimately it's the human who is responsible, and the human has to make the call whether they trust the LLM enough.
Imagine someone vibe codes the code for a radiotherapy machine and it fries a patient (humans have made these errors). The developer won't be able to point to OpenAI and blame them for this, the developer is personally responsible for this (well, their employer is most likely). Ergo, in any setting where there is significant monetary or health risk at stake, humans have to review the code at least to show that they've done their due diligence.
I'm sure we are going to have some epic cases around someone messing up this way.
> Even if we assume LLMs would consistently generate good enough quality code, code submitted by someone untrusted would still need detailed review for many reasons - so even in that case it would like be faster for the maintainers to just use the tools themselves, rather than reviewing someone else's use of the same tools.
Wouldn't an agent run by a maintainer require the same scrutiny? An agent is imo "someone else" and not a trusted maintainer.
Yes, I agree. It was just me playing with a hypothetical (but in my view not imminent) future where vibe-coding without review would somehow be good enough.
Project maintainers will always have the right to decide how to maintain their projects, and "owe" nothing to no one.
That being said, to outright ban a technology in 2026 on pure "vibes" is not something I'd say is reasonable. Others have already commented that it's likely unenforceable, but I'd also say it's unreasonable for the sake of utility. It leaves stuff on the table in a time where they really shouldn't. Things like documentation tracking, regression tracking, security, feature parity, etc. can all be enhanced with carefully orchestrated assistance. To simply ban this is ... a choice, I guess. But it's not reasonable, in my book. It's like saying we won't use ci/cd, because it's automated stuff, we're purely manual here.
I think a lot of projects will find ways to adapt. Create good guidelines, help the community to use the best tools for the best tasks, and use automation wherever it makes sense.
At the end of the day slop is slop. You can always refuse to even look at something if you don't like the presentation. Or if the code is a mess. Or if it doesn't follow conventions. Or if a PR is +203323 lines, and so on. But attaching "LLMs aka AI" to the reasoning only invites drama, if anything it makes the effort of distinguishing good content from good looking content even harder, and so on. In the long run it won't be viable. If there's a good way to optimise a piece of code, it won't matter where that optimisation came from, as long as it can be proved it's good.
tl;dr; focus on better verification instead of better identification; prove that a change is good instead of focusing where it came from; test, learn and adapt. Dogma was never good.
At the moment verification at scale is an unsolved problem, though. As mentioned, I think this will act as a rough filter for now, but probably not work forever - and denying contributions from non-vetted contributors will likely end up being the new default.
Once outside contributions are rejected by default, the maintainers can of course choose whether or not to use LLMs or not.
I do think that it is a misconception that OSS software needs to "viable". OSS maintainers can have many motivations to build something, and just shipping a product might not be at the top of that list at all, and they certainly don't have that obligation. Personally, I use OSS as a way to build and design software with a level of gold plating that is not possible in most work settings, for the feeling that _I_ built something, and the pure joy of coding - using LLMs to write code would work directly against those goals. Whether LLMs are essential in more competitive environments is also something that there are mixed opinions on, but in those cases being dogmatic is certainly more risky.
> Or if the code is a mess. Or if it doesn't follow conventions.
In my experience these things are very easily fixable by ai, I just ask it to follow the patterns found and conventions used in the code and it does that pretty well.
I've recently worked extensively with "prompt coding", and the model we're using is very good at following such instructions early on. However after deep reasoning around problems, it tends to focus more on solving the problem at hand than following established guidelines.
Still haven't found a good way to keep it on course other than "Hey, remember that thing that you're required to do? Still do that please."
Licensing is dependent on IPR, primarily copyright.
It is very unclear whether the output of an AI tool is subject to copyright.
So if someone uses AI to refactor some code, that refactored code isn't considered a derivative work which means that the refactored source is no longer covered by the copyright, or the license that depends on that.
> It is very unclear whether the output of an AI tool is subject to copyright.
At least for those here under the jurisdiction of the US Copyright Office, the answer is rather clear. Copyright only applies to the part of a work that was contributed by a human.
For example, on page 3 there (PDF page 11): "In February 2022, the Copyright Office’s Review Board issued a final decision affirming
the refusal to register a work claimed to be generated with no human involvement. [...] Since [a guidance on the matter] was issued, the Office has registered hundreds of works that incorporate AI-generated material, with the registration covering the human author’s contribution to the work."
(I'm not saying that to mean "therefore this is how it works everywhere". Indeed, I'm less familiar with my own country's jurisprudence here in Germany, but the US Copyright Office has been on my radar from reading tech news.)
Your analogy with CI/CD is flawed because while not all were convinced of the merits of CI/CD, it's also not technology built on vast energy use and copyright violation at a scale unseen in all of history, which has upended the hardware market, shaken the idea of job security for developers to its very foundation and done it while offering no really obvious benefits to groups wishing to produce really solid software. Maybe that comes eventually, but not at this level of maturity.
But you're right it's probably unenforceable. They will probably end up accepting PRs which were written with LLM assistance, but if they do it will be because it's well-written code that the contributor can explain in a way that doesn't sound to the maintainers like an LLM is answering their questions. And maybe at that point the community as a whole would have less to worry about - if we're still assuming that we're not setting ourselves up for horrible licence violation problems in the future when it turns out an LLM spat out something verbatim from a GPLed project.
> That being said, to outright ban a technology in 2026 on pure "vibes" is not something I'd say is reasonable.
To outright accept LLM contributions would be as much "pure vibes" as banning it.
The thing is, those that maintain open source projects have to make a decision where they want to spend their time. It's open source, they are not being paid for it, they should and will decide what it acceptable and what is not.
If you dislike it, you are free to fork it and make a "LLM's welcome" fork. If, as you imply, the LLM contributions are invaluable, your fork should eventually become the better choice.
Or you can complain to the void that open source maintainers don't want to deal with low effort vibe coded bullshit PRs.
>Or you can complain to the void that open source maintainers don't want to deal with low effort vibe coded bullshit PRs.
If you look back and think about what your saying for a minute, it's that low effort PRs are bad.
Using an LLM to assist in development does not instantly make the whole work 'low effort'.
It's also unenforceable and will create AI witch hunts. Someone used an em-dash in a 500 line PR? Oh the horror that's a reject and ban from the project.
2000 line PR where the user launched multiple agents going over the PR for 'AI patterns'? Perfectly acceptable, no AI here.
> Using an LLM to assist in development does not instantly make the whole work 'low effort'.
Instantly? No, of course not.
I do use LLMs for development, and I am very careful with how I use it. I throughly review the code it generated (unless I am asking for throwaway scripts, because then I only care about the immediate output).
But I am not naive. We both know that a lot of people just vibe code the way through, results be damned.
I am not going to fault people devoting their free time on Open Source for not wanting to deal with bullshit. A blanket ban is perfectly acceptable.
Your reply is based on a 100% bad-faith, intellectually dishonest interpretation of the comment to which you’re replying. You know that.
Nobody claimed that LLM code should be outright accepted.
Also, nobody claimed that open source maintainers have the right to accept or decline based on whichever criteria they choose. To always come back to this point is so…American. It’s a cop-out. It’s a thought-terminating cliche. If you aren’t interested in discussing the merits of the decision, don’t bother joining the conversation. The world doesn’t need you to explain what consent is.
Most of all, I’m sick of the patronising “don’t forget that you can fork the project!” What’s the point of saying this? We all know. Nobody needs to be reminded. Nobody isn’t aware. You aren’t being clever. You aren’t adding anything to the conversation. You’re being snarky.
> Nobody claimed that LLM code should be outright accepted
Not directly, but that's the implication.
I just did not pretend that was not the implication.
> always come back to this point is so…American
I am not American.
To be frank, this was the most insulting thing someone ever told me online. Congratulations. I feel insulted. You win this one.
> If you aren’t interested in discussing the merits of the decision, don’t bother joining the conversation.
I will join whatever conversation I want, and to my desires I adressed the merits of the discussion perfectly.
You are not the judge here, your opinion is as meaningless as mine.
> Most of all, I’m sick of the patronising “don’t forget that you can fork the project!” What’s the point of saying this?
That sounds like a "you" problem. You will be sick of it until the end of time, because that's the final right answer to any complaints of open source project governance.
> You aren’t adding anything to the conversation. You’re being snarky.
I disagree. In fact, I contributed more than you. I adressed arguments. You went on a whinging session about me.
> That being said, to outright ban a technology in 2026 on pure "vibes" is not something I'd say is reasonable.
The response to a large enough amount of data is always vibes. You cannot analyze it all so you offload it to your intuition.
> It leaves stuff on the table in a time where they really shouldn't. Things like documentation tracking, regression tracking, security, feature parity, etc. can all be enhanced with carefully orchestrated assistance.
What’s stopping the maintainers themselves from doing just that? Nothing.
Producing it through their own pipeline means they don’t have to guess at the intentions of someone else.
Maintainers just doing it themselves is just the logical conclusion. Why go through the process of vetting the contribution of some random person who says that they’ve used AI “a little” to check if it was maybe really 90%, whether they have ulterior motives... just do it yourself.
I find the fact that people can't even be bothered to put their own thoughts into text and communicate via an LLM to be the most grotesque and dystopian aspect of this new AI era.
It looks like we are going to have large numbers of people whose entire personality is projected via an AI rather than their own mind. Surely this will have an (likely deleterious) effect on people's emotional and social intelligence, no? People's language centers will atrophy because the AI does the heavy lifting of transforming their thoughts into text, and even worse, I'm not sure it'll be avoidable to have the AIs biases and start to leak into the text that people like this generate.
These aren't even their thoughts, it's just a bot let loose.
I remember the first time I suspected someone using an LLM to answer on HN shortly after chatgpt's first release. In a few short years the tables turned and it's increasingly more difficult to read actual people's thoughts (and this has been predicted, and the predictions for the next few years are far worse).
No it doesn't. That bot's comment and every comment under its profile 100% reads like an LLM to anybody that has seen enough of them. I already knew that one was a bot before even clicking the profile. See enough of them and the uncanny valley feeling immediately pops out. Even the ones that try to trick you by typing in all lowercase.
An em-dash might have been a good indicator when LLMs were first introduced, but that shouldn't be used as a reliable indicator now.
I'm more concerned that they keep fooling everybody on here to the point where people start questioning them and sticking up for them a lot of times.
I've seen skills on the various skillz marketplaces that specifically instruct the LLM-generated text to replace emdashes with hyphens (or double-hyphens), and never to use the "it's not just <thing>, it's <other thing>" phrasing.
Also to, intentionally introduce random innoccuous punctuation and speling errors.
I do wonder if the way people speak is starting to change because of LLMs. The “it’s not just” thing (I forgot the name for it) is something that used to be a giveaway, but I am now seeing more and more people use it IRL. Perhaps I am just more vigilant towards this specific sentence construction that I notice it more?
I feel like the pattern here is donate compute, not code. If agents are writing most of the software anyway, why deal with the overhead of reviewing other people's PRs? You're basically reviewing someone else's agent output when you could just run your own.
Maintainers could just accept feature requests, point their own agents at them using donated compute, and skip the whole review dance. You get code that actually matches the project's style and conventions, and nobody has to spend time cleaning up after a stranger's slightly-off take on how things should work.
Well, it's not quite that easy because someone still has to test the agent's output and make sure it works as expected, which it often doesn't. In many cases, they still need to read the code and make sure that it does what it's supposed to do. Or they may need to spend time coming up with an effective prompt, which can be harder than it sounds for complicated projects where models will fail if you ask them to implement a feature without giving them detailed guidance on how to do so.
Definitely, but that's kind of my point: the maintainers are still going to be way better at all of that than some random contributor who just wants a feature, vibe codes it, and barely tests it. The maintainers already know the codebase, they understand the implications of changes, and they can write much better plans for the agent to follow, which they can verify against. Having a great plan written down that you can verify against drastically lowers the risk of LLM-generated code
You can do all the steps I mentioned as a random contributor. I've done it before. But I agree that donations are better than just prompting claude "implement this feature, make no mistakes" and hoping it one-shots it. Honestly, even carefully thought-out feature requests are much more valuable than that. At least if the maintainer vibe-codes it they don't have to worry that you deliberately introduced a security vulnerability or back door.
Or even more efficient: the model we already have. Donate money and let the maintainer decide whether to convert it into tokens or mash the keys themself.
The point isn't that agent output is magically better; it's that reviewing your own agent's output is way cheaper (intellectually) than reviewing a stranger's, because you've written the plan by yourself. And 'slop' is mostly what you get when you don't have a clear plan to verify against. Maintainers writing detailed specs for their own agents is a very different thing from someone vibe coding a feature request
You’re assuming that maintainers have a desire to use agentic coding in the first place.
Secondly, it would seem that such contributions would contribute little value, if the maintainers have to write up the detailed plans by themselves, basically have to do all the work to implement the change by themselves.
If you're curious to see what everyone else is doing, I did a survey of over 100 major source available projects and four of them banned AI assisted commits (NetBSD, GIMP, Zig, and qemu).
On the other hand projects with AI assisted commits you can easily find include Linux, curl, io_uring, MariaDB, DuckDB, Elasticsearch, and so on. Of the 112 projects surveyed, 70 of them had AI assisted commits already.
Thanks. Now I know which software to avoid: the ones that ban legitimate tool use. I have no respect for this protectionist prohibition. These people would insist on driving horse carriages 125 years ago because people were still getting used to driving automobiles.
Everyone is starting to make AI a moral question one way or another. So your moral view is progress must march on unimpeded by private actors?
I find that pretty original. I think progress will march largely unimpeded. I would be wary of unhinged government intervention, but I wouldn’t begrudge private actors for not getting on with the ticket.
I think you'll find the luddites to be a more informative historical analogy. A new tool arrives in an industry staffed by craftsmen, providing capital a lever to raise profits at the expense of quality. Is it surprising that worker co-ops would choose not to pull that lever?
The mistake here with both the Luddites and this is to mistake the tool for the actual problem (depending on where you sit), which is mechanization and automation and ultimately capitalism itself.
Opposing the machine does/did nothing.
Political organizing around unions, state regulations of the labour market, agitational political parties did (and can again).
I think we will be getting into an interesting situation soon, where project maintainers use LLMs because they truly are useful in many cases, but will ban contributors for doing so, because they can't review how well did the user guide the LLM.
* modelling a solution that is consistent with the existing modelling/architecture of the software and moves modelling and architecture in the right direction
* verifying that the the implementation of the solution is not introducing accidental complexity
These are the things LLMs can't do well yet. That's where contributions will be most appreciated. Producing code won't be it, maintainers have their own LLM subscriptions.
I still think there is value in external contributors solving problems using LLMs, assuming they do the research and know what they are doing. Getting a well written and tested solution from LLM is not as easy as writing a good prompt, it's a much longer/iterative process.
Some months back I would have agreed with you without any "but", but it really does help even if it only takes over "typing code".
Once you do understand the problem deep enough to know exactly what to ask for without ambiguity, the AI will produce the code that exactly solves your problem a heck of a lot quicker than you. And the time you don't spend on figuring out language syntax, you can instead spend on tweaking the code on a higher architecture level. Spend time where you, as a human, are better than the AI.
I don't know, I've had good experiences getting LLMs to understand and follow architecture and style guidelines. It may depend on how modular your codebase already is, because that by itself would focus/minimize any changes.
If an author of a PR just generated code with an LLM, the GitHub PR becomes an incredibly inefficient interface between a repository owner and the LLM. A much better use of the owner time would be to interact with LLM directly instead of responding to LLM generated PR, waiting for updates, responding again, etc.
As a project maintainer, I don't want to interact with someone's LLM. If a person submits a PR, using LLM or not, the person is responsible for any problems with it. How they respond to review is a good indicator if they actually understand the code. And if they used a bot to submit the PR, I'd simply consider it a spam.
Yep, the indirection through the PR author is almost always inefficient and error-prone unless the author is really knowledgable about the code (many aren't).
And in general a lot more people want to use LLMs to generate things than want to consume the things LLMs generate. Some of the more bullish people should think harder about this pretty clear trend.
that's already happening tbh. the real issue isn't hypocrisy though, it's that maintainers reviewing their own LLM output have full context on what they asked for and can verify it against their mental model of the codebase. a random contributor's LLM output is basically unverifiable, you don't know what prompt produced it or whether the person even understood the code they're submitting.
How is that different than before LLMs? You have no idea how the person came up with it, or whether they really understood.
We are inventing problems here. Fact is, an LLM writes better code than 95% of developers out there today. Yes, yes this is Lake Wobegone, everyone here is in the 1%. But for the world at large, I bet code quality goes up.
Maybe a future direction will be the submission of detailed research, specifications and change plans for feature requests. Something that can be assessed by a human and turned into working code by both slides.
I wonder if that is an opportunity to build an Open-Source platform focused on this, replacing GitHub as the collaboration platform of a time where code was valuable.
Some sort of LLM audit trail is needed (containing prompts used, model identifier and marking all code written by LLM). It could be even signed by LLM providers (but that wouldn't work with local models). Append only standard format that is required to be included in PR. It wouldn't be perfect (e.g. deleting the log completely), but it might help with code reviews.
This would probably be more useful to help you see what (and how) was written by LLMs. Not really to catch bad actors trying to hide LLM use.
This would be a useful feature to bake into the commits generated by agents. Heck you don’t even need to wait — just change your prompt to tell it to include more context in its commit messages and to sign them as Claude rather than yourself…
The GPL talks about "the preferred form for modification of the software", and I'm starting to think that anything which involves any kind of LLM agent should be including all the text that the user gave to it as well. Prompts, etc.
Of course, even then it's not reproducible and requires proprietary software!
Except the GPL is dependent on the author having copyright over the original software but the output of an LLM may not be covered by copyright as a derivative work.
some projects (I came across Mastodon's policy[0] which they say was inspired by ghostty and others) which have a more relaxed policy of basically "LLMs are ok so long as you understand what they did, and you own it".
But I think different projects have different needs.
The "interesting situation" is that maintainers are unable to cheaply distinguish slop from good contributions so they will simply stop accepting outside contributions.
This will cut off one of the genuine entry points to the industry where all you really needed was raw talent.
> any content submitted that is clearly labelled as LLM-generated (including issues, merge requests, and merge request descriptions) will be immediately closed
Note the word "clearly". Weirdly, as a native English speaker this term makes the policy less strict. What about submarine LLM submissions?
I have no beef with Redox OS. I wish them well. This feels like the newest form of OSS virtue signaling.
That would constitute an attempt to circumvent their policy, with the consequence of being banned from the project. In other words, it makes not clearly labeling any LLM use a bannable offense.
Don't ask don't tell looks like a reasonable policy. If no one can tell that your code was written by an LLM and you claim authorship, then whether you have actually written it is a matter of your conscience.
As a native English speaker I read this as two parts. If it's obvious, the response is immediate and not up for debate. If it's not obvious then it falls in the second part - "any attempt to bypass this policy will result in a ban from the project".
A submarine submission, if discovered, will result in a ban.
Using the phrase "virtual signaling" long ago became a meaningless term other than to indicate one's views in a culture war. 10 years ago David Shariatmadari wrote "The very act of accusing someone of virtue signalling is an act of virtue signalling in itself", https://www.theguardian.com/commentisfree/2016/jan/20/virtue... .
Somewhat off topic, but I can’t believe someone got paid to write that article, what a load of crap. It’s like saying that fallacies don’t exist because sometimes people incorrectly claim the other side is arguing fallaciously.
If you go by the literal definition in the article, it’s very clear what OP meant when he said the AI policy is virtue-signaling, and it has absolutely nothing to do with the culture war.
It's not a useful phrase because a "we accept AI-generated contributions" is also virtue signalling.
You have no doubt heard claims that AI "democratizes" software development. This is an argument that AI use for that case is virtuous.
You have no doubt heard claims that AI "decreases cognition ability." This is an argument that not using AI for software development is virtuous.
Which is correct depends strongly on your cultural views. If both are correct then the term has little or no weight.
From what I've seen, the term "virtue signalling" is almost always used by someone in camp A to disparage the public views of someone in camp B as being dishonest and ulterior to the actual hidden reason, which is to improve in-group social standing.
I therefore regard it as conspiracy theory couched as a sociological observation, unless strong evidence is given to the contrary. As a strawman exaggeration meant only to clarify my point, "all right-thinking people use AI to write code, so these are really just gatekeepers fighting to see who has the longest neckbeard."
Further, I agree with the observation at https://en.wikipedia.org/wiki/Virtue_signalling that "The concept of virtue signalling is most often used by those on the political right to denigrate the behaviour of those on the political left". I see that term as part of "culture war" framing, which makes it hard to use that term in other frames without careful clarification.
The LLM ban is unenforceable, they must know this. Is it to scare off the most obvious stuff and have a way to kick people off easily in case of incomplete evidence?
It is enforceable, I think you mean to say that it cannot be prevented since people can attempt to hide their usage? Most rules and laws are like that, you proscribe some behavior but that doesn't prevent people from doing it. Therefore you typically need to also define punishments:
> This policy is not open to discussion, any content submitted that is clearly labelled as LLM-generated (including issues, merge requests, and merge request descriptions) will be immediately closed, and any attempt to bypass this policy will result in a ban from the project.
You can slap on any punishment clause you want but verifying LLM-origin content without some kind of confession is shaky at best outside obvious cases like ChatGPT meta-fingerprints or copy-paste gaffes. Realistically, it boils down to vibes and suspicion unless you force everyone to record their keystrokes while coding which only works if you want surveillance. If the project ever matters at scale people will start discussing how enforceability degrades as outputs get more human-like.
What happens when the PR is clear, reasonable, short, checked by a human, and clearly fixes, implements, or otherwise improves the code base and has no alternative implementation that is reasonably different from the initially presented version?
If you're going to set a firm "no AI" policy, then my inclination would be to treat that kind of PR in the same way the US legal system does evidence obtained illegally: you say "sorry, no, we told you the rules and so you've wasted effort -- we will not take this even if it is good and perhaps the only sensible implementation". Perhaps somebody else will eventually re-implement it later without looking at the AI PR.
How funny would it be if the path to actually implement that thing is then cut off because of a PR that was submitted with the exact same patch. I'm honestly sitting here grinning at the absurdity demonstrated here. Some things can only be done a certain way. Especially when you're working with 3rd party libraries and APIs. The name of the function is the name of the function. There's no walking around it.
It follows the same reasoning as when someone purposefully copies code from a codebase into another where the license doesn't allow.
Yes it might be the only viable solution, and most likely no one will ever know you copied it, but if you get found out most maintainers will not merge your PR.
That's why I said "somebody else, without looking at it". Clean-room reimplementation, if you like. The functionality is not forever unimplementable, it is only not implementable by merging this AI-generated PR.
It's similar to how I can't implement a feature by copying-and-pasting the obvious code from some commercially licensed project. But somebody else could write basically the same thing independently without knowing about the proprietary-license code, and that would be fine.
You not realizing how ridiculous this is, is exactly why half of all devs are about to get left behind.
Like, this should be enshrined as the quintessential “they simply, obstinately, perilously, refused to get it” moment.
Shortly, no one is going to care about anyone’s bespoke manual keyboard entry of code if it takes 10 times as long to produce the same functionality with imperceptibly less error.
> Shortly, no one is going to care about anyone’s bespoke manual keyboard entry of code if it takes 10 times as long to produce the same functionality with imperceptibly less error.
Well that day doesn't appear to be coming any time soon. Even after years of supposed improvements, LLMs make mistakes so frequently that you can't trust anything they put out, which completely negates any time savings from not writing the code.
> no one is going to care about anyone’s bespoke manual keyboard entry of code if it takes 10 times as long to produce the same functionality with imperceptibly less error.
No one is going to care about anyone’s painstaking avoidance of chlorofluorocarbons if it takes ten times as long to style your hair with imperceptibly less ozone hole damage.
The problem is that even if the code is clear and easy to understand AND it fixes a problem, it still might not be suitable as a pull request. Perhaps it changes the code in a way that would complicate other work in progress or planned and wouldn't just be a simple merge. Perhaps it creates a vulnerability somewhere else or additional cognitive load to understand the change. Perhaps it adds a feature the project maintainer specifically doesn't want to add. Perhaps it just simply takes up too much of their time to look at.
There are plenty of good reasons why somebody might not want your PR, independent of how good or useful to you your change is.
How would you tell that it's LLM-generated in that case?
If the submitter is prepared to explain the code and vouch for its quality then that might reasonably fall under "don't ask, don't tell".
However, if LLM output is either (a) uncopyrightable or (b) considered a derivative work of the source that was used to train the model, then you have a legal problem. And the legal system does care about invisible "bit colour".
Both sides written by an LLM. Both sides written based on my explicit prompts explaining exactly how I want it to behave, then testing, retesting, and generally doing all the normal software eng due diligence necessary for basic QA. Sometimes the prompts are explicitly "change this variable name" and it ends up changing 2 lines of code no different from a find/replace.
Also I'm watching it reason in real time by running terminal commands to probe runtime data and extrapolate the right code. I've already seen it fix basic bugs because an RFC wasn't adhered to perfectly. Even leaving a nice comment explaining why we're ignoring the RFC in that one spot.
Eventually these arguments are kinda exhausting. People will use it to build stuff and the stuff they build ends up retraining it so we're already hundreds of generations deep on the retraining already and talking about licenses at this point feels absurd to me.
You may as well be the MPAA right now throwing threats around sharing MP3s. We're past the point of caring and the laws will catch up with reality eventually. The US copyright office says things that get turned over in court all the time.
I think the bigger point about enforcement is not whether you're able to detect "content submitted that is clearly labelled as LLM-generated", but that banning presumes you can identify the origin. Ie.: any individual contributor must be known to have (at most) one identity.
Once identity is guaranteed, privileges basically come down to reputation — which in this case is a binary "you're okay until we detect content that is clearly labelled as LLM-generated".
[Added]
Note that identity (especially avoiding duplicate identity) is not easily solved.
I suspect this is for now just a rough filter to remove the lowest effort PRs. It likely will not be enough for long, though, so I suspect we will see default deny policies soon enough, and various different approaches to screening potential contributors.
Any sufficiently advanced LLM-slop will be indistinguishable from regular human-slop. But that’s what they are after.
This heuristic lets the project flag problematic slop with minimal investment avoiding the cost issues with reviewing low-quality low-effort high-volume contributions, which should be near ideal.
Much like banning pornography on an artistic photo site, the perfect application on the borderline of the rule is far less important than filtering power “I know it when I see it” provides to the standard case. Plus, smut peddlers aren’t likely to set an OpenClaw bot-agent swarm loose arguing the point with you for days then posting blogs and medium articles attacking you personally for “discrimination”.
A sign to point at when you get someone is posting "I asked AI to fix this and got this". You can stop reading and any arguments right there. Saving lot of time and effort.
Just require that the CLA/Certificate of Origin statement be printed out, signed, and mailed with an envelope and stamp, where besides attesting that they appropriately license their contributions ((A)GPL, BSD, MIT, or whatever) and have the authority to do so, that they also attest that they haven't used any LLMs for their contributions. This will strongly deter direct LLM usage. Indirect usage, where people whip up LLM-generated PoCs that they then rewrite, will still probably go on, and go on without detection, but that's less objectionable morally (and legally) than trying to directly commit LLM code.
As an aside, I've noticed a huge drop off in license literacy amongst developers, as well as respect for the license choices of other developers/projects. I can't tell if LLMs caused this, but there's a noticeable difference from the way things were 10 years ago.
> As an aside, I've noticed a huge drop off in license literacy amongst developers
What do you mean by this? I always assumed this was the case anyway; MIT is, if I'm not mistaken, one of the mostly used licenses. I typically had a "fuck it" attitude when it came to the license, and I assume quite a lot of other people shared that sentiment. The code is the fun bit.
> I always assumed this was the case anyway; MIT is, if I'm not mistaken, one of the mostly used licenses
No, it wasn't that way in the 2000s, e.g., on platforms like SourceForge, where OSS devs would go out of their way to learn the terms and conditions of the popular licenses and made sure to respect each other's license choices, and usually defaulted to GPL (or LGPL), unless there was a compelling reason not to: https://web.archive.org/web/20160326002305/https://redmonk.c...
Not being able to publish anything without sifting through all the libs licences? Remembering legalese, jurisprudence, edge cases, on top of everything else?
MIT became ubiquitous because it gives us peace of mind
Sarcasm? Nobody will be contributing with a complexe signing process like that, and it doesn't guarantee anything in the end, it's like a high tech pinky swear
Yep, that’s why my forks of all their libraries with bugs fixed such as https://github.com/pmarreck/zigimg/commit/52c4b9a557d38fe1e1... will never ever go back to upstream, just because an LLM did it. Lame, but oh well- their loss. Also, this is dumb because anyone who wants fixes like this will have to find a fork like mine with them, which is an increased maintenance burden.
The PR doesn't disclose that "an LLM did it", so maybe the project allowed a violation of their policy by mistake. I guess they could revert the commit if they happen to see the submitter's HN comment.
Dunno but a commenter already noted that some begins to say: "No LLM generated PR, but we'll accept your prompt" and another person answered he saw that too.
Hugely unpopular opinion on HN, but I'd rather use code that is flawed while written by a human, versus code that has been generated by a LLM, even if it fixes bugs.
I'd gladly take a bug report, sure, but then I'd fix the issues myself. I'd never allow LLM code to be merged.
Because human errors are, well, human. And producing code that contains those errors is a human endeavor. It bases on years, decades of learning. Mistakes were made, experience was gained, skills were improved. Reasoning by humans is relatable.
Generating slop using LLMs takes seconds, has no human element, no work goes into it. Mistakes made by an LLM are excused without sincerity, without real learning, without consequence. I hate everything about that.
For the parent there's immaterial value knowing that is written by a human. From what I read in your comment, you see code more as a means to an end. I think I understand where the parent is coming from. Writing code myself, and accomplishing what I set out to build sometimes feels like a form of art, and knowing that I build it, gives me a sense of accomplishment. And gives me energy. Writing code solely as a means to an end, or letting it be generated by some model, doesn't give that same energy.
This thinking has nothing to do with not caring about being a good teammate or the business. I've no idea why you put that on the same pile.
Sure, but back in reality no you’re not? No more than any other contributor?
If I want to use an auto-complete then I can, and I will? Restricting that is as regressive as a project trying to specify that I write code from a specific country or… standing on my head.
Sure, if they want me to add a “I’m writing this standing on my head” message in the PR then I will… but I’m not.
No, you can't. See, that's where you are just wrong: when you don't respect the boundaries an open source project sets that you want to contribute to then you are a net negative.
Restricting this is their right, and it is not for you to attempt to overrule that right. Besides the fact that you do not oversee the consequences it also makes you an asshole.
They're not asking for you to write standing on your head, they are asking for you to author your contributions yourself.
They are asking me to author my contributions in a way that they approve of. The essence of the request is the same as asking someone to author them whilst standing on their head.
Except they don’t, won’t and can’t control that: the very request is insulting.
I’ll make a change any way I choose, upright, sideways, using AI. My choice. Not theirs.
Their choice is to accept it or reject it based purely on the change itself, because that’s all there is.
Their boundaries. If they don’t want to accept the code, cool. Nobody is forcing them to, and I respect that.
But if they can’t enforce their boundaries, because they can’t tell the difference between AI code and non-AI code without being told, then their boundaries they made up are unenforceable nonsense.
About as nonsense and enforceable as asking me to code upside down.
I'll make this blunt: if you're a guy then half the population is not capable of 'enforcing their boundaries' against you, more so if you count children. The problem you seem to have is to think that if someone is not capable of enforcing their boundaries that they are not allowed to have those boundaries and that it is your god given right to do whatever the F* you want just because you can. That's not how the world works, nor is it how it should work.
Boundaries - of all kinds - are not unenforceable nonsense, they are rights that you willingly and knowingly violate.
I’ll make this blunt: I expect you to write in all caps as per my policy below. You’re really just going to do whatever the F* you want in lowercase just because you can? That’s not how the world works, nor how it should work.
Markdown files - of all kinds - are totally not unenforceable nonsense, they are rights of a real legal entity (the git repository) that you willingly and knowing violate every time you don’t comment in all caps.
And yes, before you ask, this discussion is definitely one in which it is appropriate to bring up rape and pedophilia.
I am authoring my contributions, using Clause Code as a tool. It doesn't make me an asshole.
If the maintainers don't want to accept it, fine. Someone will eventually fork and advance and we move on. The Uncles can continue to play in their no AI playground, and show each other how nice their code is.
The world is moving on from the "AI is bad" crowd.
Forking the code can be perfectly reasonable, with this or any other disagreement about policy. The main point of contention in this thread is whether you ought to lie about having used an LLM. I agree with Jacques: doing something like that would make you an asshole.
Even before AI, getting a fix into an open source project required a certain level of time and effort. If you prefer to spend your time on other things, and you assume it will eventually get fixed by someone else, using an LLM to fix it just for yourself makes sense.
If you rely on llms, you're simply not going to make it. The person who showed their work on the math test is 9/10 times is doing better in life than the person that only knew how to use a calculator. Now how do we think things are going to turn out for the person that doesn't even think they need to learn how to use a calculator.
Just like when people started losing their ability to navigate without a GPS/Maps app, you will lose your ability to write solid code, solve problems, hell maybe even read well.
I want my brain to be strong in old age, and I actually love to write code unlike 99% in software apparently (like why did you people even start doing this career.. makes no sense to me).
I'm going to keep writing the code myself! Stop paying Billionaires for their thinking machines, its not going to work out well for you.
I went into software because I like building things and coming up with solid solutions to business problems that are of use to society. I would not describe myself with "love to code". It's a means to an end to pay bills and have a meaningful career. I think of myself more like a carpenter or craftsman.
I used a coding agent for the majority of my current project and I still got the "build stuff" itch scratched because Engineers are still responsible for the output and they are needed to interface between technical teams, UX, business people etc
> I think of myself more like a carpenter or craftsman.
> I used a coding agent for the majority of my current project and I still got the "build stuff" itch scratched because Engineers are still responsible for the output and they are needed to interface between technical teams, UX, business people etc
Then you are the opposite of a carpenter or a craftsman, no matter what you think about it yourself.
And yet, I find a coding agent makes it even more fun. I spend less time working on the boilerplate crap that I hate, and a lot less time searching Google and trying to make sense of a dozen half-arsed StackOverflow posts that don't quite answer my question.
I just went through that yesterday with Unity. I did all the leg work to figure out why something didn't work like I expected. Even Google's search engine agent wasn't answering the question. It was a terrible, energy-draining experience that I don't miss at all. I did figure it out in the end, though.
Prior to yesterday, I was thinking that using AIs to do that was making it harder for me to learn things because it was so easy. But comparing what I remember from yesterday to other things I did with the AI, I don't really think that. The AI lets me do it repeatedly, quickly, and I learn by the repetition, and a lot of it. The slow method has just 1 instance, and it takes forever.
This is certainly an exciting time for coders, no matter why they're in the game.
> The person who showed their work on the math test is 9/10 times is doing better in life than the person that only knew how to use a calculator
Sure but once you learn long multiplication/division algorithms by hand there's not much point in using them. By high school everyone is using a calculator.
> Just like when people started losing their ability to navigate without a GPS/Maps app
Are you suggesting people shouldn't use Google Maps? Seems kind of nuts. Similar to calculators, the lesson here is that progress works by obviating the need to think about some thing. Paper maps and compasses work the same way, they render some older skill obsolete. The written word made memorization infinitely less valuable (and writing had its critics).
I don't think "LLMs making us dumber" is a real concern. Yes, people will lose some skills. Before calculators, adults were probably way better at doing arithmetic. But this isn't something worth prioritizing.
However, it is worth teaching people to code by hand, just like we still teach arithmetic and times tables. But ultimately, once we've learned these things, we're going to use tools that supercede them. There's nothig new or scary about this, and it will be a significant net win.
>I want my brain to be strong in old age, and I actually love to write code unlike 99% in software apparently (like why did you people even start doing this career.. makes no sense to me).
I am old now, and the unfortunate truth is that my brain isn't working as fast or as precise as when I was young. LLMs help me maintain some of my coding abilities.
It's like having a non-judgemental co-coder sitting at your side, you can discuss about the code you wrote and it will point out things you didn't think of.
Or I can tap into the immense knowledge about APIs LLMs have to keep up with change. I wouldn't be able to still read that much documentation and keep all of this.
I agree but only in the very long term. I think short-medium term, it's not going to matter as the MBA types get so caught up in the mania that results matter even less than they normally do.
One doesn't exclude the other. I still program myself; I actually have more time to do so because the LLM I pay some billionaire for is taking care of the mundane stuff. Before I had to do the mundane stuff myself. What I pay the billionaire is a laughable fraction compared to the time and energy I now have extra to spend on meaningful innovation.
Dangerous that all these projects keep going MIT. We wouldn't have an open source community if it weren't for protections against modification without sharing. Almost all software today would be proprietary, as it was before.
No. People shared code because they wanted to. Open standards are great tools against emerging monopolies. So the losing side used that. IBM lost OS/2 vs NT war. They propped up Linux. Intel wanted to have a second option to Microsoft in server space. AMD wants to gain some developers against Nvidia Cuda monopoly. That's the reason they contribute. Even Linux's own leadership decided against extra freedoms for users; they rejected GPLv3 to keep company contributions going. That's why LLVM gets the first implementations of certain optimizations and architectures, yet being permissive licensed.
Quite a bit of the Linux userspace is already permissively licensed. Nobody has built a full-fledged open source alternative yet. Because it is hard to build an ecosystem, it is hard to test thousands of different pieces of hardware. None of that would happen without well-paid engineers contributing.
You pay taxes to a government using it to wage wars bombing children schools.
Will you now live in hut a on the forest because you don't consent to it?
Not necessarily a bad idea, but I think the bigger issue here and now is the increasing assymmetry in effort between code submitter and reviewer, and the unsustainable review burden on the maintainers if nothing is done.
Honestly, given that that GPL model would be far below SOTA in capabilities, what exactly would be its use-case? Why would anyone try to use an inferior LLM if they can get away with using a superior one?
It doesn't make sense, because GPL means only GPL comes out, not only GPL can go in:
>Many of the most common free-software licenses, especially the permissive licenses, such as the original MIT/X license, BSD licenses (in the three-clause and two-clause forms, though not the original four-clause form), MPL 2.0, and LGPL, are GPL-compatible. That is, their code can be combined with a program under the GPL without conflict, and the new combination would have the GPL applied to the whole (but the other license would not so apply).
https://en.wikipedia.org/wiki/License_compatibility#GPL_comp...
A model that contains no GPL code makes sense so that people using non-GPL licenses don't violate it.
Are you and Redox just going to fall behind? Projects that used to take months take days or hours.
It seems well intentioned, but lots of bad ideas are like this.
I was told by my customer they didn't need my help because Claude Code did the program they wanted me to quote. I sheepishly said, 'I can send an intern to work in-house if you don't want to spend internal resources on it.'
I can't really imagine what kind of code will be done by hand anymore... Even military level stuff can run large local models.
Projects that used to take months still take months. LLM’s are only useful for throwaway low-quality slop. Perhaps some times the sloperator will get lucky and the end result isn’t something that will bite them in the ass. But the rest of us foresee a mountain of tech debt that will come knocking one day.
I am wondering why people spam OSS with AI slop pull requests in the first place?
Are they really that delusional to think that their AI slop has any value to the project?
Do they think acting like a complete prick and increasing the burden for the maintainers will get them a job offer?
I guess interacting with a sycophantic LLM for hours truly rots the brain.
To spell it out: No, your AI generated code has zero value. Actually less than that because generating it helped destroy the environment.
If the problem could be solved by using an LLM and the maintainers wanted to, they could prompt it themselves and get much better results than you do because they actually know the code. And no AI will not help you "get into open source". You don't learn shit from spamming open source projects.
That is one reason I believe. Being told that you need open source presence to be employed.
I think some people also like the feeling of being helpful. And they do not understand reality of LLM outputs. See comments posting AI generated summaries or answers to question. With no verification or critical checking themselves.
I genuinely feel bad for you. At least it isnt difficult to make the transition to AI Agent coding. Even untrained people are able to do it.
At some point your manager is going to force you to AI code. At best you can try to find some healthcare or finance company that is too cheap to buy a machine that can locally run 400B models.
My manager can try. I’ll say no, and he’ll be welcome to start the slow process of laying me off, which will end in my employer having to pay me 10 months worth of wages. By the time I’ll be faced with having to find a new job, I’ll be retiring anyway. Using LLMs is not in my job description, and I will not do it.
Untrained people are producing utter garbage. I am 100% in on AI coding and haven’t typed into the IDE in months but the craft remains pretty much as challenging as ever.
I understand the knee-jerk reaction to restrict LLM's, but that feels like a failing prospect. They're going to be doing an incredible amount of heavy lifting on code generation, so why would you intentionally cut out what will likely be 90% or more of potential contributions? Wouldn't it be better to come up with a system that tags the type of contributor, ie. human vs. AI?
What about building an Agentic architecture that reduces your review burden? Just a thought.
That's a shame. We used claude to migrate redox very successfully to pure rust cranlift on mac: https://github.com/pannous/redox (but then got stuck on multi-cpu assembly):
One thing that is missing is a standard no-LLM policy, like the "Contributor Covenant Code of Conduct". On PostgREST we recently added a strict no-LLM policy [1], basically linking Gentoo's AI policy, which we found the most apt in lack of a standard.
The interesting tension here is that "no LLM-generated code" is easy to state but hard to enforce - a developer who uses an LLM to understand a concept and then writes the code themselves is indistinguishable from one who didn't. The policy probably works as a cultural signal more than a technical guarantee, which might be exactly what they want.
the skills that protect against displacement long-term are exactly what vibe coding erodes. an engineer who built with AI but never developed the instincts to spot its mistakes has a gap they don't know they have. this maintainer problem is a preview: when you can't tell the difference between a PR from someone who understood the code and one from someone who just prompted into it, the verification burden doesn't disappear. it shifts to whoever has enough skill to catch the errors.
That’s a shallow way to get people to own up to their code. I haven’t read the original link but going by the title, a blanket no-LLM is bad. Driven by human and supported by AI should be the way.
Hiring managers could help here: the only thing that should count as a positive when - if - you feel like someone's open source contributions are important for your hiring decision is to make it plain that you only accept this if someone is a core contributor. Drive-by contributions should not count for anything, even if accepted.
> This policy is not open to discussion, any content submitted that is clearly labelled as LLM-generated (including issues, merge requests, and merge request descriptions) will be immediately closed, and any attempt to bypass this policy will result in a ban from the project.
It sounds serious and strict, but it applies to content that's 'clearly labelled as LLM-generated'. So what about content that isn't as clear? I don't know what to make of it.
My guess is that the serious tone is to avoid any possible legal issues that may arise from the inadvertent inclusion of AI-generated code. But the general motivation might be to avoid wasting the maintainers' time on reviewing confusing and sloppy submissions that are made using the lazy use of AI (as opposed finely guided and well reviewed AI code).
Very next sentence lists the penalty for lying. So you can defraud the project, but only if you can walk the walk and talk the talk well enough for them to never notice you’re using an LLM. At that point it’s more effort than just complying with the policy.
So... my prediction is that they will either have to close off their dev process or start using LLMs to filter contributions in the attempt to detect submissions from LLMs.
While I am more on the AI-hater side, I don't consider this to be a good idea:
"any content submitted that is clearly labelled as LLM-generated (including issues, merge requests, and merge request descriptions) will be immediately closed"
For example:
- What if a non-native English speaker uses the help of an AI model in the formulation of some issue/task?
- What about having a plugin in your IDE that rather gives syntax and small code fragment suggestions ("autocomplete on steroids")? Does this policy mean that the programmers are also restricted on the IDE and plugins that they are allowed to have installed if they want to contribute?
> What if a non-native English speaker uses the help of an AI model in the formulation of some issue/task?
Unfortunately, when I have seen this in the context of the Rust project, the result has still been the typical verbose word salad that is typical of chat style LLMs. It is better to use a dedicated translation tool, and post the original along with the translation.
> What about having a plugin in your IDE that rather gives syntax and small code fragment suggestions ("autocomplete on steroids")?
Very good question, I myself consider this sort of AI usage benign (unlike agent style usage), and is the only style of AI I use myself (since I have RSI it helps having to type less). You could turn the feature off for just this project though.
> Does this policy mean that the programmers are also restricted on the IDE and plugins that they are allowed to have installed if they want to contribute?
I don't think that follows, but what features you have active in the current project would definitely be affected. From what I have seen all IDEs allow turning AI features on and off as needed.
> and post the original along with the translation
this so many times - it's so incredibly handy to have the original message from the author, for one I may speak or understand parts of that language and so have an easier time understanding the intent of the translated text. For another I can cut and translate specific parts using whatever tools I want, again giving me more context about what is trying to be communicated.
> What if a non-native English speaker uses the help of an AI model in the formulation of some issue/task?
How can you be sure the AI translation is accurately convening what was written by the speaker? The reality is you can't accommodate every hypothetical scenario.
> What about having a plugin in your IDE that rather gives syntax and small code fragment suggestions ("autocomplete on steroids")? Does this policy mean that the programmers are also restricted on the IDE and plugins that they are allowed to have installed if they want to contribute?
Nobody is talking about advanced autocomplete when they want to ban AI code. It's prompt generated code.
>What if a non-native English speaker uses the help of an AI model in the formulation of some issue/task?
Firefox has direct translation built in. One can self-host libretranslate. There are many free sites to paste in language input and get a direct translation sans filler and AI "interpretation". Just write in your native language or your imperfect English.
Translation software does not solve the problem that the tone that you have to use in English is often very different from the tone in your native language. What I would write in German would sometimes not be socially acceptable for English speakers.
If the native language is very different from English, this problem gets much worse.
This is a problem that LLM claim to partially mitigate (and is one reason why non-native speakers could be tempted to use them), but hardly any classical translation tool can.
I'd be interested to see examples of this where the translated text is not editorializing and behaves like other machine translators, just better. I expect there to be some missed nuance when translating languages, but I also expect LLMs to clobber it.
> What if a non-native English speaker uses the help of an AI model in the formulation of some issue
I've seen this excuse before but in practice the output they copy/paste is extremely verbose and long winded (with the bullet point and heading soup etc.)
Surely non-native speakers can see that structure and tell the LLM to match their natural style instead? No one wants to read a massive wall of text.
if (foo == true) { // checking foo is true (rocketship emoji)
20 lines of code;
} else {
the same 20 lines of code with one boolean changed in the middle;
}
Description:
(markdown header) Summary (nerd emoji):
This PR fixes a non-existent issue by adding an *if statement** that checks if a variable is true. This has the following benefits:
This PR does not just improve performance, it fundamentally reshapes how we approach performance considerations. This is not just design --- it's architecture. Simple, succinct, yet powerful.
They're too methodical, and duplicate code when they're longer than a single line fix. I've never received a pull request formatted like that from a human.
The problem is that the well you are drinking from has in fact been poisoned. Maybe you think you can tolerate it but some projects are taking a policy decision that any exposure is too dangerous and that is IMO perfectly reasonable.
I see a lot of oss forks in the future where people just fork to fix their issues with LLMs without going through maintainers. Or even doing full LLM rewrites of smaller stuff.
Probably, but on the other hand, this is almost literally the definition of technical debt -- it's great to get fixes uptreamed precisely so that you don't have to maintain your own fork, keep it in sync, etc. an LLM can likely lower the burden of that but the burden still exists.
I assume that most of these purely llm generated unwanted contributions will just end up in dead end forks, because my impression is that a lot of them are just being generated as GitHub activity fodder. But the stuff that really solves a problem for a person - eh, good. Problem solved is problem solved. (Unless it creates new problems)
The purpose of a LLM ban is to encourage use of LLMs to submit PRs, not discourage. The longer term effect is to eliminate FOSS competency from the hiring process.
It takes some human effort to set up a slop generator. Have the slop generator make 100 buckets of slop, humans will work hard accepting or rejecting the buckets, somewhat less than 100 buckets will be approved, the payoff for the owner of the slop generator is now they have "verified FOSS developer contribution" on their resume which translates directly into job offers and salary. Its a profitable grift, profitable enough that the remaining humans are being flooded out. The ban makes successful submission to Redox even MORE valuable than before. They can expect infinite floods of PRs now that a successful PR "proves" that Redox thinks the human owner of the slop generator did the work and should therefore be offered more jobs, paid more, etc. Technically, they're hiring and paying based on ability to set up a slop generator which is not zero value, but not as valuable as being an Official Redox Contributor.
In the long run, this eliminates FOSS competency from the hiring process. Currently FOSS competency and coding experience indicates a certain amount, however minimal, of human skill and ability to work with others. Soon, it'll mean the person claiming to be a contributor has no problem violating orders and rules, such as the ones forbidding AI submissions, and it'll be a strong signal they actively work to subvert teams for their own financial reward and benefit. Which might actually be a hiring bullet point for corporate management in more dysfunctional orgs, but probably not help individual contributors get hired.
They're certainly welcome to do whatever they're like, and for a microkernel based OS it might make sense--I think there's probably pretty "Meh" output from a lot of LLMs.
I think part of the battle is actually just getting people to identify which LLM made it to understand if someones contribution is good or not. A javascript project with contributions from Opus 4.6 will probably be pretty good, but if someone is using Mistral small via the chat app, it's probably just a waste of time.
AI has the potential to level the playing field somewhat between open source and commercial software and SaaS that can afford armies of expensive paid developers.
Time consuming work can be done quickly at a fraction of the cost or even almost free with open weights LLMs.
Let someone from the Redox team go read [1], [2], and [3]. If they still insist on keeping their position then ... well. The industry is being redefined as we speak and everyone doing the push-back are pushing against themselves really.
“Our approach is harness-first engineering: instead of reading every line of agent-generated code, invest in automated checks that can tell us with high confidence, in seconds, whether the code is correct. “
that’s literally what The whole industry has been doing for decades, and spoiler: you still need to review code! it just gives you confidence that you didn’t miss anything.
Also, without understanding the code, it’s difficult to see its failure modes, and how it should be tested accordingly.
So you read the three-part series of blogs that are packed in details in 3 minutes after I shared the link and put yourself into a position of entitled opinion and calling my position a silly take? Sure thing.
Their profile generally comes up here on HN very often with Dunning-Kruger effect like comments so it makes me believe it is no AI. AI would do a better analysis, for the better or worse.
Implementing a Redis and Kafka rewrite (in Rust) but with workload-aware and self-balancing JIT-like engine deployed at Datadog-scale is no fluff. You obviously have no idea what you're talking about.
> The industry is being redefined as we speak and everyone doing the push-back are pushing against themselves really.
No, they’re pushing back against a world full of even more mass surveillance, corporate oligarchy, mass unemployment, wanton spam, and global warming. It is absolutely in your personal best interest to hate AI.
While I appreciate the morality and ethics of this choice, the current trend means projects going in this direction are making themselves irrelevant (don't bother quipping at how relevant redox is today, thanks). E.g. top security researches are now using LLMs to find new RCEs and local privilege escalations; no reason why the models couldn't fix these, too - and it's only the security surface.
IOW I think this stance is ethically good, but technically irresponsible.
Even if we assume that LLMs become good enough for this to be true (some might feel that is the case already - I disagree, but that is beside the point), there is no reason why OSS maintainers should accept such outside contributions that they would need to carefully review, as it comes from an untrusted source, when they could just use the tools themselves directly. Low effort drive-by PRs is a burden with no upside.
I think clients who care about getting good software will eventually require that LLMs are not directly used during the development.
I think one way to compare the use of LLMs is that it is like comparing a dynamically typed language with a functional/statically typed one. Functional programming languages with static typing makes it harder to implement the solution without understanding and developing an intuition of the problem.
But programming languages with dynamic typing will let you create a (partial) solutions with a lesser understanding the problem.
LLMs takes it even more easy to implement an even more partial solutions, without actually understanding even less of the problem (actually zero understanding is required)..
If I am a client who wants reliable software, then I want an competent programmer to
1. actually understand the problem,
2. and then come up with a solution.
The first part will be really important for me. Using LLM means that I cannot count on 1 being done, so I would not want the contractor to use LLMs.
In my experience with the right set of guardrails LLMs can deliver high quality code. One interesting aspect is doing security reviews and formal verification with agents that is proven to be very useful in practice.
Not sure how they can expect to make a viable full OS without massive use of LLMs, so this makes no sense.
What makes sense if that of course any LLM-generated code must be reviewed by a good programmer and must be correct and well written, and the AI usage must be precisely disclosed.
What they should ban is people posting AI-generated code without mentioning it or replying "I don't know, the AI did it like that" to questions.
The problem is the increasing review burden - with LLMs it is possible to create superficially valid looking (but potentially incorrect) code without much effort, which will still take a lot of effort to review. So outright rejecting code that can identified as LLM-generated at a glance, is a rough filter to remove the lowest effort PRs.
Over time this might not be enough, though, so I suspect we will see default deny policies popping up soon enough.
Because it takes a massive amount of developer work (perhaps more than anything else), and it's very unlikely they either have the ability to attract enough human developers to be able to do it without LLM assistance.
Not to mention that even finding good developers willing to develop without AI (a significant handicap, even more so for coding things like an OS that are well represented in LLM training) seems difficult nowadays, especially if they aren't paying them.
>Not sure how they can expect to make a viable full OS without massive use of LLMs, so this makes no sense.
Humans have been doing this for the better parts of 5 decades now. Don't assume others rely on LLMs as much as you do.
>Not to mention that even finding good developers willing to develop without AI (a significant handicap, even more so for coding things like an OS that are well represented in LLM training) seems difficult nowadays, especially if they aren't paying them.
I highly doubt that. In fact, I'd take a significant pay cut to move to a company that doesn't use LLMs, if I were forced to use them in my current job.
> Because it takes a massive amount of developer work
You know what else takes "a massive amount of developer work"?
"any LLM-generated code must be reviewed by a good programmer"
And this is the crux of the matter with using LLMs to generate code for everything but really simple greenfield projects: They don't really speed things up, because everything they produce HAS TO be verified by someone, and that someone HAS TO have the necessary skill to write such code themselves.
LLMs save time on the typing part of programming. Incidentially that part is the least time consuming.
The submitter is supposed to be the good programmer; if not, then maintainers may or may not review it themselves depending on the importance of the feature.
And yes of course they need to be able to write the code themselves, but that's the easy part: any good developer could write a full production OS by themselves given access to documentation and literature and an enormous amount of time. The problem is the time.
> The submitter is supposed to be the good programmer;
And how will that be assured? Everyone can open a PR or submit a bug.
> The problem is the time.
But not the time spent TYPING.
The problem is the time spent THINKING. And that's a task that LLMs, which are nothing other than statistical models trying to guess the next token, really aren't good at.
There are only 4 successful general purpose production OSes (GNU/Linux, Android/Linux, Windows, OS X/iOS) and only one of those made by the open source community (GNU/Linux).
And a new OS needs to be significantly better than those to overcome the switching costs.
> There are only 4 successful general purpose production OSes
Feel like you are using a very narrow definition of "success" here. Is BSD not successful? It is deployed on 10s of millions of routers/firewalls/etc in addition to being the ancestor of both modern MacOS and PlaystationOS...
> And a new OS needs to be significantly better than those to overcome the switching costs.
Who cares if nobody switches to it as their daily driver? The goal you proposed was "viable", not "widely used". The former is perfectly possible without LLMs (as history has proved), and the latter is unrelated to how you choose to make the OS.
Just because they have been made before LLMs doesn't mean it can be done again, since there was just one success (GNU/Linux) and that success makes it much harder for new OSes since they need to better then it
