Flathub has replaced its blanket ban on AI-generated applications with a disclosure requirement. Reviewers can still reject submissions, and manifests remain off-limits for AI-generated content.
I think they should just create a policy where unmaintained software gets removed? Or flagged where it’s easy to filter out.
Most AI slop just gets abandoned, but a real developer who wants to maintain FOSS software long term with some AI assistance is quite a different thing.
It’s especially frustrating because I don’t think I’ve seen a single app store on a distro prominently show when the last update was for a flatpak app. Even flatpaks own website requires scrolling all the way down to the bottom of an app to see that info.
Plus the app store interfaces/apps are terribly slow and bloated, gnome software will sit in the background doing nothing useful and consume 700MB+ of RAM, searching for an app takes forever, and it will lock up sometimes while installing so you can’t continue browsing/searching for more apps. I’ve tried other stores and they all share the same issues for the most part.
Overall the whole situation is just pretty bad and frustrating. I use the terminal for flatpak updates and such, but it’s not any good for finding new apps.
An unforeseen consequence from AI Slop is that normal people can now create quite handy software for niche use cases, but the barrier to entry is so low that those people often abandon it quickly too. Rather than security risks, abandonedware is starting to become a bigger problem.
I don’t know about other normal people, but I’m certainly actively maintaining my own vibecoded apps as long as it’s affordable to do so. By that I mean they give me the niche functionality I want & new features are not added anymore, but if a pillow or pyqt update breaks my app, I may figure out how to fix it even on my own.
i feel like the coupons aren’t working
the cat eating eagles are being printed faster
there should be a place for ai apps, that is separate from human built apps
I’d just put them in a different repo that can be toggled on and off by the user. If you don’t want slop you toggle off the slop repo and go about your day.
Bit then what gets put in there? If developer used AI at all? If thrybise it to fix bugs.
We could call it SlopHub
sounds nice
That seems like a lot of work for something that can be done with a search filter.
This suggests that it has to be disclosed to the reviewers, which is a good thing, but how will end users be made aware?
I only use 6 flatpaks (due them not being available in the arch official repo’s) and I get them via a gui package manager. I’m assuming I am going to have to modify how I do this for any future installs, and check my existing ones, but then what about future releases/updates?
A ‘no ai’ swtich/toggle would be swell.
You are aware, that if you are using Linux you are using AI generated code?
Yes, I’m also aware that is redundant to this conversation. There is no reason for decent people to just give up on decency just because one layer (even the top layer) of something is f*cked up.
If people gave up that easily, we would all be working in coal mines 14 hours a day, living in trash societies, with slave labour, and trash governments… oh
Hopefully someone forks the kernel, and there is an increase in anti ai distro’s, de’s etc.
Ya the only thing that keeps me from being omega worried about it is that I use Debian and by the time code gets to me that shit has been tested to hell and back. So less likely some AI nonsense is going to break something.
Your ai nonsense distro is protecting you from the flatpak ai nonsense?
idk why you think Debian is AI nonsense but my point was that AI code tends to be shipped without actually being looked at or tested by a human being. Debian is known for being extremely slow to make changes. So by the time stuff gets to stable it has been around long enough that it’s been tested a lot.
you might want search ‘debian ai vote’
This resolution therefore affirms that generative AI is neither exempt from nor subject to special rules beyond the standards already expected of Debian contributors. The responsibility for every contribution rests with the contributor who submits it, who remains accountable for its technical quality, legal acceptability, and suitability for inclusion in Debian.
My desire is to have my system function. I like stability. If a project has tons of AI slop being thrown at it it takes time to sift through that and find code that’s not garbage. Debian takes the time to do that. Other distros are much faster paced and simply don’t have the time to do that properly before code shows up on your machine. That’s what I was talking about. The Debian Stable pipeline. I’m downstream so the AI slop should hopefully get noticed and patched out before it reaches me. So I don’t see Debian Stable as AI slop. Maybe I’m wrong. Maybe AI will ruin the Debian Project. But I’m hopeful the project’s policies that have kept it going for decades will continue to keep things stable going forward.
Probably the right decision given there’s no way to truly tell what level of AI generation or effort within the project itself with AI has taken place.
I understand the willingness to ban low effort low quality one shot projects, but given most software nowadays incorporating some level of LLMs, I can’t ever see a blanket ban working out anywhere.
that’s like saying that because we can’t all always 100% tell if someone is cheating at an online game we should forget about banning cheaters - you can prohibit a behavior you aren’t always able to sus out and use a combination of detection heuristics and making that behavior unwelcome (i will say i believe the correct policy is to label projects which discernibly use llms rather than banning them)
“It’s too op, I can’t compete, nerf it”
do you understand what an analogy is?
i don’t think llm content is cheating; i think it’s usually garbage and that slop projects are very rarely maintained for more than a few monthsif people want to risk it they can risk it if they want i just want it labeled or in a separate category so i don’t have to waste my own time
i just find the argument that it is impossible to tell something 100% means you shouldn’t bother deeply suspect since we don’t apply that to other things we care about
Cheaters are annoying because they win.
If the new projects were all as useless as you claim, you wouldn’t compare them to cheating.
This is a direct contradiction. My conclusion: You’re salty that they are more useful than your projects.
By and large flathubs metadata has allowed for pretty transparent overviews. Stick the same filtering that safe and unsafe containers have and everyone retains a clear choice.
🤦♂️
I don’t have an issue with this. If it’s AI you can usually tell, if you can’t tell then even if it’s AI it’s not slop and it’s fine to use.
And human code is so good it doesn’t have any vulnerabilities? Well I have a bridge to sell you.
That goes for AI and human made software. You should only use security critical software that is trusted by the community.
sure, that’s true, but if we are being serious do you think software made nearly exclusively by a human (or humans) who care or software made mostly by a non-deterministic token generator working on behalf of a human who couldn’t bother to write the code themself is more likely to be maintained long enough to gain community trust
Using AI isn’t just a sign of laziness. I don’t see an issue with using the tools available to do something in less time than usual. If a program is vibe coded then of course it’s usually going to be of bad quality, but if it’s a mix of human and human checked AI then it can still be a high quality project.
my concern isn’t “laziness” - it’s that i want software that is maintainable over something flashy that will be rotting in a year
someone who knows what they are doing using autocomplete is fine in my opinion if that is truly how it is being used. the problem is that once you accept an llm writing the rest of a line of boiler plate for you it is a short step into it writing a function body for you which i trust far less.
it is easy to just accept llm content as ‘good enough’ and become lulled into not thinking critically about the security and performance implications of the code it is popping out. /if/ humans who know what they are doing are /truly/ checking the output the same way they would scrutinize an intern they don’t trust’s code then… sure. we should be serious though about the fact that that is neither what the ai companies are incentivizing nor what they are advertising.
all of that aside, my most pragmatic concern is that projects slapped together with meaningful llm aid usually are not - in practice - maintained for long unless they were already a well maintained project pre-llm mass adoption. checking a repo for something shiny and new and finding that claude is a contributor is, in my experienced, /very/ tightly correlated with finding that /every single/ top level file/directory was last changed as part of the same commit n number of months ago (n > 2).
part of why i pick and use the tools i use is because i have /some/ confidence that i won’t have to drop them and deal with identifying (and then working through the hassle of sidegrading to) a new tool for that task every ~6 months. i would hate to see long maintained projects i use, or even cool new projects i follow, hamstring their maintainability in the medium and long term in exchange for a burst of speed in the short term.
I’m weary of the slippery slope fallacy embedded in your reply, I’m not sure if using some AI often results quality to fall out the window.
Also I don’t really trust any new projects to be maintained for a long time, unless I trust the developer(s) have a clear commitment to supporting it in the long term.
this project (configarr) is one I’ve been using in myself-hosted setup clearly uses AI to help with development, although I’m not sure to what extent. It’s been maintained consistently for 2 years.
Avoiding using brand new programs has worked quite well for me not being left with unmaintained software so far.
Humans are lazy as long as AI is affordable and that is a big if right now more and more people well use it for more and more things. How many people remwbmwr phone numbers or directions anymore?
It means you contain the generated code.
