from TMP_NKcYUEoM7kXg4qYe@lemmy.world to selfhosted@lemmy.world on 15 Mar 09:31
https://lemmy.world/post/44284092
The list is (in my opinion) just human written slop.
The requirements for some software to make it in the list are so lax, that the list loses its meaning. It really isn’t a list of “slop” but just list of software which is somehow connected to AI. In my opinion it’s similar to the anti systemd hate. People think they understand some software better than the maintainer who wrote it.
Examples of bs criteria:
- Having AI features - This one is a conundrum to me. Why do you need a list for it? If you are using the app, you know if it has AI features or not and if you have not noticed, why are you bothered?
- Having
agents.mdin the repository - This one does not mean anything on it self. Could just be there for others to use, not necessary for the maintainers. It’s Open Source, remember? Someone else might read it and might want to use different tools than you. - KeepassXC - This one is bizzare. They link KeepassXC’s blog post, which explains why KeepassXC is very much not “slop”. Every tiny change has to be reviewed by a maintainer regardless if it was human or AI written. Because they allow LLM usage, they also encourage people to disclose their usage. This is what I think should secure codebases aim for.
Ironically I will link to the list: codeberg.org/small-hack/open-slopware
Also the KeepassXC blog post: keepassxc.org/…/2025-11-09-about-keepassxcs-code-…
Yes, I am aware that I have no authority over which links you can post and which ones not but hopefully this post will convince you.
Perhaps someone will create a better list of actual slop? I think the idea of it is pretty good, this particular list is crap though.
#selfhosted
threaded - newest
Counterpoint: I think it’s reasonable to inform people of what a dev team allows into their project, or how they’re directing the development of the product.
For some people, it will be a hard no. Others maybe don’t mind AI interoperability if it’s not in the codebase itself. Others won’t mind if the app is critical to their workflow, but may reject apps that are less critical. Others don’t care at all.
I look at it like the Denuvo labeling curator on Steam. Nothing wrong in informing people.
Having agents.md is a silly criteria… One project I’m with is doing GSoC this year and another org recommended using agents.md to tell the AI that will inevitably be used that it needs to tell the person that AI copy paste will not be accepted and please don’t bother unless you can demonstrate understanding of what you are doing.
I’m not sure if this also applied to the (now deleted?) original repository since this is a fork, but I agree that the name of the project is misleading.
The readme file mentions the following:
So it is supposed to be a list of software that uses AI to any degree and leaves it up to the reader what to do with said information.
And to counter your argument:
It isn’t always clear whether features use AI unless you actively look for it. Some people prefer avoiding software that uses it to a certain or any degree.
If you think this is a bad list, then I believe that’s because you misunderstood it’s purpose, though I don’t necessarily blame you.
You could open an issue to suggest changing the name.
I’d never heard of this list, so thanks for sharing. I have to say while some of the projects seem to have been included due to minor offences, I’m really disappointed in some of my favourite FOSS projects.
This uses some pretty terrible criteria. Being critical is imporant, and slopware should be looked upon very spektically. But please do this properly and dont slop together questionable lists.
Downvote, hide, and move on. If it isn’t against the rules, it’s allowed.