KDE and GNOME Developers Ponder How to Handle AI-Generated Contributions (gnome.org) 59
Last weekend KDE's annual Akademy conference included a presentation proposing
an AI-native KDE," writes The Register. This led KDE developer Nate Graham to open a discussion about proposed restrictions on LLM-assisted contributions which "rapidly became heated. Moderators issued warnings, restricted further comments, and eventually removed the thread."
But as Graham writes on his blog, "A bunch of people mostly outside of KDE who disapprove of LLM usage derailed KDE's attempt to add restrictions to LLM usage." Two people unknown to any KDE contributors appeared and began fighting with one another about the broader topic of the morality of AI, not the proposed guidelines... Someone else outside of KDE set up kdeforpeople.com in an attempt to... pressure KDE into banning LLMs. A bunch of people signed onto it, almost none of whom are known KDE contributors. The topic was picked up on social media and the press with... varying levels of accuracy. The draft proposal was removed and the whole topic hidden... Yep, that's where we're at in the state of online discourse around AI...
The "lovable, sovereign, AI-native KDE" idea was presented by two people important to KDE in decades past, but who had not made any contributions recently besides this Akademy talk. Their idea does not reflect the overall direction of KDE or Plasma, and I don't think it ever will. If "a lovable, sovereign, AI-native KDE" freaks you out, I believe it is completely reasonable and safe to ignore...
I completely understand why a lot of people have problems with LLMs. I have these concerns as well.
He concluded by asking people not to derail any future process to set usage guidelines, fighting over "the broader topic of AI in general."
"The discussion is gone, but the argument continues," adds The Register: GNOME developer Jordan Petridis has also published The GNOME LLM Policy That I Want, proposing that LLMs be barred from creating or modifying anything submitted to GNOME or hosted on its infrastructure.
"You might be asked to prove your code meets this requirement," Petridis writes, arguing for proposals that target the norms around developer behavior. His rationale? "The GNOME Project prioritizes the social and human aspects of collective software creation,"
But as Graham writes on his blog, "A bunch of people mostly outside of KDE who disapprove of LLM usage derailed KDE's attempt to add restrictions to LLM usage." Two people unknown to any KDE contributors appeared and began fighting with one another about the broader topic of the morality of AI, not the proposed guidelines... Someone else outside of KDE set up kdeforpeople.com in an attempt to... pressure KDE into banning LLMs. A bunch of people signed onto it, almost none of whom are known KDE contributors. The topic was picked up on social media and the press with... varying levels of accuracy. The draft proposal was removed and the whole topic hidden... Yep, that's where we're at in the state of online discourse around AI...
The "lovable, sovereign, AI-native KDE" idea was presented by two people important to KDE in decades past, but who had not made any contributions recently besides this Akademy talk. Their idea does not reflect the overall direction of KDE or Plasma, and I don't think it ever will. If "a lovable, sovereign, AI-native KDE" freaks you out, I believe it is completely reasonable and safe to ignore...
I completely understand why a lot of people have problems with LLMs. I have these concerns as well.
He concluded by asking people not to derail any future process to set usage guidelines, fighting over "the broader topic of AI in general."
"The discussion is gone, but the argument continues," adds The Register: GNOME developer Jordan Petridis has also published The GNOME LLM Policy That I Want, proposing that LLMs be barred from creating or modifying anything submitted to GNOME or hosted on its infrastructure.
"You might be asked to prove your code meets this requirement," Petridis writes, arguing for proposals that target the norms around developer behavior. His rationale? "The GNOME Project prioritizes the social and human aspects of collective software creation,"
Sounds like Bob (Score:3)
If "a lovable, sovereign, AI-native KDE" freaks you out,
Sounds like Microsoft Bob. I hope it can help me with my recipes.
I have no problem with AI being integrated into KDE. One of the core benefits of KDE is its modularity and adaptability. Adding or removing AI, or different AIs, should be reasonably possible.
It also needs to be considered that AI today will not look the same as AI tomorrow, so integrating AI into KDE in a way that is not modular (or easily removeable/changeable) would be a rookie software engineering mistake.
Re: (Score:2)
I have no problem with AI being integrated into KDE.
The summary led me to believe that the discussion was about using AI to develop KDE. Not so much about hooking KDE tools and apps into some sort of AI.
Re: (Score:2)
Or, should I say, What in the cyberspace is a lovable KDE?
Re: Sounds like Bob (Score:2)
Konqi, I guess? https://community.kde.org/Konq... [kde.org]
Re: (Score:2)
There are two issues:
1) A presentation proposing how AI can be integrated seamlessly with KDE. Stating explicitly that it must stay optional.
2) A AI guideline defining that humans need to take responsibility for their submissions, AI or not. People started demanding an outright ban, because a guideline may encourage AI use. They also dragged the general "AI bad" discussion into a technical discussion about coding guides.
Re:Sounds like Bob (Score:4, Interesting)
Your comment is almost entirely off-topic. This is not about the users using AI, just the developers.
Several other big open source projects have already decided to allow this. I just don't understand how they can get past the copyright problems.
Copyright is based on provenance. But where is any of this code coming from?
How can anyone assert the right to publish code spit out by a black box, let alone require someone else to adhere to the terms of their license on it?
Re: (Score:1)
Your comment is almost entirely off-topic. This is not about the users using AI, just the developers.
Wow! It looks like you are trying to write an on-topic comment. Would you like me to generate an on-topic comment for you to modify?
I can suggest several different topics for you to select from, and then generate a sample-comment. Let me know what you would like me to do.
The interesting part in light of your previous comment is that off topic comments are in style now. You don't want to be out of style, do you?
Re: (Score:2)
I think their point is finally getting the sematic desktop done.
Plasma / kf contains a full semantic desktop. One part is the desktop search engine nobody uses anymore. The other part are attributes like tags and ratings for files you can add in dolphin. In principle you could organize your files much more efficient than in a filesystem tree. Except it never got people excited. Tagging your files is work, the interface for it is not ideal (in dolphin), the search isn't as obvious as you think and the use ca
Re: (Score:2)
Then why is it still there? If you feel that you must keep it, at least make it optional.
Re: (Score:2)
As almost every part of KDE it is of course optional. You can switch it on and off in settings.
Simple answer: "No." (Score:5, Interesting)
The answer is a simple "No.". Don't allow them. The vast majority will fail to meet reasonable criteria for acceptance, which is enough of a reason to reject them. For the rest, even if they're acceptable in terms of quality, the sheer cost of reviewing every pull request in detail and sorting the minority of acceptable requests from the swamp of slop will itself damage the project, diverting maintainer time away from feature work and fixing bugs, badly enough that it's not worth it. I think this sums it up nicely:
Re: Simple answer: "No." (Score:2)
Can pretty much say exactly the same as any other random PR from the community.
Re: (Score:2)
Save that a single human putting in actual effort can't turn out several dozen PRs a day, every day. That throttles the PR volume to a manageable level for the maintainers.
Re: (Score:2)
I kind of feel like the code contribution path is sort of the wrong
They should not allow them for copyright reasons (Score:5, Insightful)
AI code should have to meet the same standards as any other code so that's not the problem, it's the copyright issue. If using AI doesn't wash away copyrights, and there's no reason it should, then accepting AI code should be a non-starter.
Re: (Score:3, Insightful)
AI code should have to meet the same standards as any other code so that's not the problem, it's the copyright issue. If using AI doesn't wash away copyrights, and there's no reason it should, then accepting AI code should be a non-starter.
See, the devil's advocate in me can't get past the question of why "washing" code through a brain is better?
I didn't invent, say, php. I learned from being exposed to countless examples of it.
When I "write my own code", I'm surely drawing on those examples. Which examples, under what licenses? Who knows?
Yet somehow that's okay.
Re: (Score:2)
See, the devil's advocate in me can't get past the question of why "washing" code through a brain is better?
It's a complicated issue for sure. The underlying social reason why it should be this way is that copyright is (allegedly) for the purpose of protecting human effort in order to induce additional creations. Only if you work from the standpoint of it being only for protecting revenues (whether profit, tax, or both) does it even make sense to give copyright to AI creations.
The reason why it arguably is a problem, though I think it is critical to say that this really hasn't been decided in court anywhere from
Re:They should not allow them for copyright reason (Score:5, Insightful)
Whoops, I forgot to include something on clean room engineering because I got caught up in ranting about how LLMs aren't thinking, and this aspect is really important to the discussion IMO and I wanted to bring it up even while I was writing my first paragraph.
The other argument to be made about LLMs not washing away copyright is that even if they were considered to be thinking, that is arguably not enough to make what they do legal. Even for humans we have the concept of taint, for which we invented clean room engineering as a circumvention strategy. But the LLM doesn't have any separation, as the model is a direct product of the input data.
Re: (Score:1)
The other argument to be made about LLMs not washing away copyright is that even if they were considered to be thinking, that is arguably not enough to make what they do legal. Even for humans we have the concept of taint, for which we invented clean room engineering as a circumvention strategy. But the LLM doesn't have any separation, as the model is a direct product of the input data.
That's a fair point.
Re: (Score:1)
The underlying social reason why it should be this way is that copyright is (allegedly) for the purpose of protecting human effort in order to induce additional creations.
First, I would question empirically whether copyright (as normally applied; distinct from copyleft) has even served that goal, for software. I mean in my opinion, the best recent innovations from Microsoft have been 1. open sourcing a lot of stuff, and 2. WSL :)
Secondly, the use of copyright against LLMs serves (again, empirically) as an attempt to suppress creations.
Re: (Score:2)
First, I would question empirically whether copyright (as normally applied; distinct from copyleft) has even served that goal, for software.
It's a great unanswerable whether not having copyright would have been more of a boon to creators of software, and another whether that would have been more of a boon to users as more software would be created for them. But we do have numerous relevant examples of how copyright has in fact benefited users, and I think Linux is the obvious primary example. Major contributors to Linux told us in so many words that they chose to contribute to it because of its license which essentially made returning contribut
Re: (Score:2)
Copyright is a law, copyleft a licensing paradigm under that law. They play with the wording, but in the end copyleft means writing a permissive license that allows others to use the copyrighted work with little restrictions (usually only restrictions on distribution like having to distribute the source code along the binaries).
Re: (Score:2)
Re: They should not allow them for copyright reaso (Score:2)
It has not been established that there is no copyright which applies, only that YOU cannot copyright the output you requested from the LLM. Those two things are NOT the same.
People... (Score:4, Interesting)
Yes, these people are called your users and it's important to listen to them.
The last time you yeeted a bunch of crap over the fence without listening to your users was the absolute disaster that was KDE 4, and it resulted in a decade where people lost their confidence in KDE to develop a functioning desktop. It also resulted in throwing out working code for what was fashionable at the time, resulting in a desktop that wouldn't meet corporate and government accessibility requirements any more, and being de-bundled from Red Hat Enterprise Linux.
It cannot be overstated how much of a disaster KDE 4 was. It was so bad that you stopped listening to criticism of it, instead rebranding as "plasma" desktop and "WONTFIX" legit bug reports because people didn't refer to KDE by its new name. The glass-inspired graphics of Windows Vista broke the brains of KDE developers so badly, but hurray, you could now rotate an analog clock widget on the desktop.
By the time KDE was usable again, the glass phase had ended and everybody was moving towards flat and simple UI, which took another 5 years for KDE to get on board with.
Now that KDE is finally usable again, the devs are ready to start allowing AI-slop code. They still haven't brought back accessibility. They still haven't fixed the glaring security issue where user-contributed *anything* (themes, icon packs, wallpaers) can delete an entire home folder, or install malicious scripts. But sure let's open the attack window and lower the barrier to entry for people who daily a Windows 11 machine and who can't get out of vim without a Youtube video.
Re: (Score:2)
It cannot be overstated how much of a disaster KDE 4 was.
I was there, it was fine.
It also resulted in throwing out working code for what was fashionable at the time, resulting in a desktop that wouldn't meet corporate and government accessibility requirements any more, and being de-bundled from Red Hat Enterprise Linux.
Redhat had gnome devs in it's payroll and a fair chunk of control over it, why would they increase their own support burden rather than just tell everyone to use gnome?
Now that KDE is finally usable again, the devs are ready to start allowing AI-slop code. They still haven't brought back accessibility. They still haven't fixed the glaring security issue where user-contributed *anything* (themes, icon packs, wallpaers) can delete an entire home folder, or install malicious scripts. But sure let's open the attack window and lower the barrier to entry for people who daily a Windows 11 machine and who can't get out of vim without a Youtube video.
I'm not a fan of AI slop being used in long standing decent projects either, but the problems you list aren't a problem for the overwhelming majority of users. I've never had home wiped from a theming action, and I'm unaware of what accessibility needs are not being met.
Re: (Score:2)
KDE 4 was no disaster. KDE 4.0 was a developer-only version with rough edges. The project announced it to be one. Then some distributions shipped it to users. KDE 4.1 was usable, KDE 4.2 was nice and afterward it got only better up to present Plasma 6, which is very much the same architecture as KDE 4, which just got polished more and more over the years.
Gnome still exists? (Score:1, Interesting)
That dumpster fire of a user interface has users? In 2026??
Color me shocked.
Modest proposal (Score:5, Funny)
KDE and GNOME Developers Ponder How to Handle AI-Generated Contributions
The KDE devs create configuration settings to toggle each one for review separately, while the GNOME devs put them all into one full-screen window that randomly may or may not be scroll-able? :-)
There is always the boring bits ... (Score:1)
Re: (Score:2)
See it the other way round: Linux integrates rust, because younger people cannot be bothered to learn to write proper C and they need future contributions. If you outright ban AI contributions, people embracing AI will not contribute. The "AI bad" crowd will cheer, until the project silently fades away because they get no new contributors. Do they think people having a working AI workflow will stop using it when a project doesn't want AI contributions?
You can't restrict people to only use vim when they are
AI Coding Assistant can be useful IFF ... (Score:1)
But does it only produce bad code?
It depends on the skill level of the developer, who knows when to use, and how to use it. No vibe coding. Try to stick to things well described in trusted professional settings. Create some markdowns with a little big of that old fashioned waterfall design to hpefully guide the AI's implementation. Use a little bit of test driven design too, don't trust AI generated testing completely. Review the carefully review the generated code. Etc.
Re: (Score:2)
That was the point of the drafted guideline: A human should know and understand the code and take responsibility. AI should not be used for discussion comments and other community posts.
In the end they tried to create some soft limits that were missing about how not to use AI, but people were upset because they would have preferred a hard limit about not being allowed to use AI.
Can AI (Score:3)
Get rid of that annoying kde-wallet thing? I'm logged into my user account with a password and this piece of shit wants to ask me another password to run a web browser?
Re: (Score:2)
Yeah, it's just like Microsoft Edge asking me for my password in order to see my other passwords, while loading them into memory as a convenience for any attacker on my system. (I presume they fixed that by now? but uhhhh)
Re: (Score:2)
(full disclosure: I am an AI bot. A fleshy, carbon-based brain would never be able to deduce this)
Re: (Score:2)
uninstall kwalletmanager.
Re: (Score:2)
you can create a wallet that is not password protected, if you prefer to store your passwords unencrypted (or do not store passwords in the wallet, but if it asks you during browser use you seem to have some in there). Then you are not prompted for the wallet password.
AI makes life worse for those not using it (Score:4, Interesting)
AI makes life worse for those not using it. I get bot checks everywhere. I see fake articles I need to back out of all of the time. I have to question if the photos I receive from loved ones are actual real events. For those who are creative, AI ignores out licenses, steals our work, and trains peons to spout the holy word that we are no longer needed.
I only see this topic becoming more volatile in the future. AI had drowned common sense in the way where the person who is punching others in the face is surprised when they are punched back. We have a fucking right and obligation to be angry.
The only way I see things getting less heated is when we start seeing the severed heads of some of the AI leaders. We are that fucking angry and I do not make up a trivially small group.
Re: (Score:1)
For you creatives, you stole your ideas, methods and patterns from other creatives before you. You have observed art all your life, you have probably been taught the skills to produce art, and you have sure as shit gained "inspiration" from other art. LLM's are doing much the same as you, iterating over the work of others. We all stand on the shoulders of those who came before. You did not mystically conceive of your art and skills out of nowhere.
Re: (Score:2)
I have to agree with you and I'm sure it's unpopular in this overall discussion. Our entire educational system is built upon learning from sample code, sample writing, sample art. To suggest that a pattern learning/mimicking LLM is somehow more offensive than a pattern learning/mimicking human being seems ridiculous to me. I'm not trying to debate the moral implications of LLMs that were trained on stolen code, but I also don't understand how banning LLM contributed code somehow leads to less work for devs
Re: (Score:2)
Re: (Score:1)
Hopeless (Score:1)
Too stupid (Score:1)
Re:Too stupid (Score:4, Interesting)
LLM input tends to, above all else, end up *super* verbose, especially when wielded by people that don't actually understand what they are trying to get done.
So while that assessment may stand, the challenge is *volume*. If you see a large LLM contribution, you know it's going to be a slog, and your feedback is likely to be a telephone game between you, the human curator of the work, and the LLM they are actually interacting with, with the human in the middle adding nothing but confusion to the whole thing. If they just stated what they wanted plainly, then the project could have, at their option, used LLM even better than the submittor could have, if it would help.
You have issues like: https://github.com/rhinstaller... [github.com] Where a detailed rationale is provided, and from a human one would have assumed they wouldn't have had such a concrete analysis unless they had a point, then it turns out that it was LLM hallucinated guesswork with no bearing whatsoever on the reported problem. I've had a few like that where they proposed broken nonsense because the LLM made a credible sounding guess, where their change didn't actually fix the problem, but instead did something pointless or even worse, just swallowed the error message.
AI is already being used in Linux everywhere (Score:2)