Its not foolproof but I swear I saw somone maintaining a list, possibly hosted on github.
It was bringing awareness to (fully) open source apps/devs that had started using vibe code or partial AI in the app/software code in some way.
I can't even remember some of the more surprising offenders but wanted to check over my installs.
It might have been android based but possibly not.
7 Comments
Ediacarium@feddit.org · 6 pts · 9h
List of FOSS that contains vibecode: https://codeberg.org/ethical-foss/open-slopware
List of FOSS that doesn't contain vibecode: https://codeberg.org/brib/slopfree-software-index
The lists may not be complete, e.g. OrcaSlicer is missing from the first list (has both
AGENTS.mdandCLAUDE.md)lavember@programming.dev · 3 pts · 5h
very weird they (1st linked repo) include Immich as having "AI functionality" and said functionality is face recognition and other features unrelated to LLMs
I wonder how much this sort of effort against AI is harming the contrarian vision of stopping it then helping, because we look stupid blindly hating anything to do with a neural network
the problem I have is the reductionist take that any project should be deemed 'tainted' if it has a permissive AI policy such as "use whatever you want but you'll be banned if its slop"
there are great arguments against LLMs, as they are today, to not exist, which I strongly agree.
it doesn't mean the best course of action is to target FOSS projects for tolerating its use, deeming them tainted for it, given they are not contributing to LLM's existence. thinking WE, as users or developers of said projects, are perpetuating the cycle of abuse is shifting blame from the massive corporations actually pushing all of it.
it feels the same like with "carbon offset" ideas, where the concept of global warming is framed as "if only we do our part by buying from this carbon-neutral brand we will save everyone!" whereas the absolute majority of worldwide carbon emissions are from corporations and not individuals, and said framing only serves to mask that fact
AI should be regulated 'til oblivion, but can we not point fingers at each other and actually focus the discussion on the real pushers?? because saying that, for instance, Immich, an actually great project built by people for people, is "tainted" because LLM contribs are "ok" and has facial recognition is absolutely stupid because it deflects so hard from the actual people fucking us over
Dave@lemmy.nz · 1 pts · 16m
Also interesting that it (Immich) is labelled as AI permissive with a link to a section saying no AI generated code in PRs is allowed.
ghodawalaaman@programming.dev · 1 pts · 5h
that list of open-slopeware made me depressed since almost all of the software I use are in there :(
Zetta@mander.xyz · 2 pts · 7h
Imo, there is a material difference between vibe code and LLM assisted code. Vibe code would be somebody like me with pretty much no coding experience making an end product. And LLM assisted coding would be a real coder who knows what they're doing who uses LLMs for assistance and fully reviews the code.
Whether it's disclosed or not, I am of the belief that pretty much every single piece of open source (and closed) software that is even a little bit popular most likely has LLM assisted code. So theres no avoiding that.
jaredwhite@humansare.social · 3 pts · 5h
If someone types into a prompt and they get a bunch of code spat out, I don't care if they reviewed it or not. It's slop.
The only LLM-assisted code I can think of which I personally would not classify as vibe coding is the old-school copilot autocomplete. Unfortunately, that's not what anyone really means nowadays—so-called agenetic coding has become the norm.
Also, regarding your last assertion, see: https://handcraftedcode.org/
atrielienz@lemmy.world · 2 pts · 5h
https://lemmy.world/post/41818548
This?