Home Creators Posts Import Register Favorites Logout

Downloads

Content

Hey everyone!

Today, we have:

00:05 KDE's proposed AI policy leads to massive backlash

https://www.gamingonlinux.com/2026/09/kde-developers-attempts-at-creating-llm-guidelines-are-not-going-well/

05:54 GNOME dev offers a "no AI at all" policy

https://blogs.gnome.org/alatiera/2026/09/23/the-gnome-llm-policy-that-i-want/

Have a great day!

Files

Comments

Quidam

I have now switched to the new RUST desktop Roundel, actaully Roundel OS, which is the Roundel Desktop on Debian. Clearly, it's built with AI assistance but it's the fastest desktop I've ever used and has so many unique features . Whilst KDE & Gnome is spending all it's energy arguing, other devs are making a desktop that makes them suck. I have it running on my old laptop, one step up from a potato, and it is as quick as my framework running Gnome.

Quidam

It's funny, many of these people have been using YouTube, Google, Facebook, Microsoft, etc for decades and never complained about their massive, energy hungry data servers, stealing their data and posts. Now it's AI, that can do their jobs, they are suddenly finding every reason to block it. If they hate AI, they should stop watching YouTube and using Google Products, as it's probably the biggest AI company and certainly the biggest data thief.

NotMyName

There is escaping work that uses AI: version freezing. Whether you use the release of GPT-3, Microsoft Copilot, Claude, or explicit known commits, you can know at what point 100% of the code was written by humans, and when that started to decline. Fully forking the whole stack isn't necessary, you can just fix known security bugs and preserve what was made. Also, love ya Nick, but genie-out-of-the-bottle is copy-paste propaganda from AI companies designed to make people give up and subscribe instead of make things better. They were saying it before it was any good, because it's marketing. An AI-focused future isn't a forgone conclusion based on its characteristics, but by people's attitude.

AArexx AAron Ruscetta

There is a perfect A\bducted I\ntellect policy -- NONE AT ALL -- there is not a project on the planet that needs FAKE developers who rely on environment destroying, Hannibal Lechter Stochastic Parrot Parasites to eat up their user's brains while sucking out the neurons of their own. There is NOTHING that is inevitable about the use of this cancerous DATA RAPE tech. ...THat being Ranted... I just saw a great talk by a KDE UX designer who had some solid thoughts on ethical LLM. From a share with some UX developer friends... worth checking out the link I think! .... The video is a tech talk from a FLOSS designer / developer working on KDE, I'm sharing because it seems to me he is presenting some novel and unique ways of approaching the problems of computer interface design, exploring how we might integrate some of the valuable extensions of tools like obsidian into the iconic icon and screen desktop presentations of your data. < https://www.youtube.com/watch?v=V7AfAcQwLW0 > Full disclosure is to note my confirmation bias for his deeply critical views of A\bducted I\ntellect, though he does give fair consideration the potential for ethical application of properly scaled and private LLM NeuruoNet tech. [ interesting thoughts on this point at about 22 to 25 min., with mention of a Swiss LLM model called APERTUS trained entirely on ethically sourced data ]

Marco Schmidlin

I agree with you, Nick, that the genie is out of the bottle, like it was when cars were invented compared to horses. But I don't see it as that negative for the open source space. Of course, there are all those negative aspects that you all know about (ethically, environmentally, etc.), but there is also a big chance for open source, especially bigger projects like KDE, to finally lift off because they were always short on development resources. Now, a few genius developers can achieve so much more that I think that open source will thrive. Existing proprietary software solutions have lost a lot of value, since it's so easily replaceable with something new, and that leads us to more choice. Therefore, you can pick solutions that are open to development, which is power to people and companies to even more evolve the software freely. You can't do that with proprietary software, and therefore they lose their value. But I also agree that the aspect of having a people community will be lost this way, since the focus will mainly be on outcome and results. This is sad, but also, there could be a middle ground where community and outcome could both be achieved.

narF

"You can't put the genie back into the bottle" I disagree. Of course we could! Bigger society changes happened when there was motivation. When Covid happened, we all stopped going outside. In another era, we stopped believing in gods and kicked out monarchy. These are much bigger changes that affected the majority of population. Unlike AI which is currently used only by a minority of people (despite what AI companies want us to believe) Let's look at the current situation. New AI models are trained by massive data centers and only like 5 companies in the world have capacity to do it. We could pass a regulations about how AI need to be trained in ethical and environmental safe ways. Enforcing those rules wouldn't be too hard, since we only have about 5-10 companies to watch, those with actual capacity to train new AI models. Yes, it would require international coordination, and it wouldn't be easy, but it's not impossible. We've done it with Covid, we can do it again.

Nick - The Linux Experiment

I tried it in a VM, and it was unusable there, all kinds of visually buggy, but I didn't have time to try it out on a real PC

Nick - The Linux Experiment

This s not the same scale at all :) Just for AI, it's 3.5 times toe existing power consumption that took decades to mount up. It' snot that AI can do people's jobs (it clearly can't yet), it's because the massive scale is insane in a world where we're already exhausting existing resources. We don't need MORE energy usage today. The timing is the worst one possible, when considering the state of the world.

Nick - The Linux Experiment

I do think it's too late. No government will back down on AI, not when every other one is pushing it that hard. Without government regulation, there is no stopping this thing; It's not propaganda, it's just realism: even if everyone banded together to not use AI as customers, Companies will use it, governments will use it.

Nick - The Linux Experiment

I still think it's a net negative overall: even if this helps project develop faster, do we need that? When every other non FOSS project gains the same advantage, with more devs ready to use AI to grow even faster? As a whole, AI isn't making the world a better place right now, and I don't see how it could in the future either :/

Nick - The Linux Experiment

No country is going to want to take that step if it means other countries can ignore these rules and keep going. The entire world would need to band together at the exact same moment to pressure every major government to back down and regulate. This never happened for anything crucial in the world, not with climate change, not with Covid (each country did their own thing at their own timing and pace, or with nuclear weapons, or with international law, or with international sanctions. I don't see it happening for a tech that props up capitalism so hard when every government wants more capitalism right now. in a ideal world, sure, we could do it, but no country is going to go at it on their own, and I don't see how the current political landscape would favor international cooperation on such an issue. I hope I'm proven wrong.

Quidam

Yes. I had the same thing. Probably the lack of graphic acceleration in VM, usual thing. Works great on hardware, but wait for its first release though

Quidam

I agree with the power consumption issue. Hopefully, they will create new tech to resolve this that will benefit the planet. My comment still stands about people using Facebook, Google/YouTube, X, etc. It's rather hypocritical that people make contents about the negative effects of AI and then share this content on platforms that are responsible for creating these problems. I saw one YouTuber, after a rant about the effect of AI on the environment, went on to complain AI was scanning his Gmail. I shouted at my screen and unsubscribed.

Marco Schmidlin

"do we need that?": yes please 😁 I'm just not such a patient person. I had to wait for multi monitor high DPI support for years, which was absolutely painfull to me. I know, now we have all these things. Now it seems everything is there, but the next thing is surely over the horizon. Regarding the speed of proprietary software projects: I think this is a big win for smaller highly skilled teams, since you can't just throw an indefinite amount of developers at a problem/product and have a faster output. Else, the swarm of developers could be replaced with AI itself. So far, I don't see any proprietary software grow much quicker than before (yes, Windows now has the taskbar on the side, yeah 😂). But I agree with you that the net overall is negative. Now it's just more the question of what we as the FOSS community want to do with the situation: fight a losing game (IMHO it's lost) or make the best out of the situation (make FOSS projects better than proprietary)

Marco Schmidlin

Now another aspect came to my mind. I actually always wanted to contribute to an open-source project, but I was always lacking time and energy to do that properly because I just have limited resources (and yes, a family 😉) . If the hurdle for contributing for a skilled developer is lowered by just orchestrating agents, I'm much more willing/able to contribute code. But as you already said, it only works if you're a highly skilled developer, you won't learn much with it and you never get a community feeling, that is a big part of FOSS.