Slop-free FOSS ๐Ÿ–‘๏ธ


FOSS apps, projects, and services with accounts on the Fediverse that do NOT use AI, LLMs, Claude, vibe coding, etc.

fedidevs.com/s/MTA2OA/

#NeverSlop


I should make a list of FOSS apps, projects, and services that have Fediverse accounts and don't code with AI, then strike them from the list with a big ELIMINATED notice if they do start drinking the slop kool-aid.

EDIT: Made one fedidevs.com/s/MTA2OA/

Honorary mentions
infosec.exchange/@catsalad/117341552145650963


reshared this

in reply to Cat ๐Ÿˆ๐Ÿฅ— (D.Burch)
Don't forget to always have one finger on the block button, and one eye on the retractions list. Fortunately that list above added the latter for this kind of cases, I'd recommend adding a notification whenever the page changes so you're on the safe side.
in reply to Cat ๐Ÿˆ๐Ÿฅ— (D.Burch)

Bevy only has one co-authored by Claude commit, however their AI Policy does indeed allow for LLMs, so....

โš ๏ธ ELIMINATED


@bevy :aneobot_explode:

We are wary of the use of AI in the development of Bevy, and are particularly troubled by its impact on artists, but tolerate careful human-driven AI use in some code-related areas consistent with the above values.
in reply to Cat ๐Ÿˆ๐Ÿฅ— (D.Burch)

github.com/bevyengine/bevy/issues/23418#issuecomment-4088119141
Their response to this being discovered was effectively 'Ew, fuck, our bad...' and they'll purge it eventually.

Losing your shit at such a large project letting one commit slip through, a project lead-maintained by a visibly trans dev for that matter, is a bit dickish, tbh.

in reply to Raeve
@raeve Oh the one Claude commit is not why I removed them from the list, it is that they "tolerate careful human-driven AI use in some code-related areas" (as per their AI contribution policy)
in reply to Cat ๐Ÿˆ๐Ÿฅ— (D.Burch)

You're doing the impossible dichotomy thing where a glorified autocomplete must be argued entirely out of every usage in a way that fully exhausts their maintenance efforts until they give up, because somehow that shitty autocomplete is so pervasive and obvious that it will overtake regular human effort in the repo with lazy hallucinatory garbage.

Or maybe they'll just do what you don't seem to have continued reading about, this crazy thing called having standards that shitty autocomplete will either fail to meet or be indistinguishable from human-written code and otherwise evasive of enforcement without disclosure.

in reply to Raeve

Yeah, in fact, I fully fucking agree with their assertion: Zero-tolerance doesn't work. Slopcoders will just try to hide it better. When they make it clear that they will grill a submitter on their rationale and understanding and that they won't accept discussion or documentation written by LLMs (the hard part, mind), I feel like this presents a better trap than trying to argue every little code completion.

This isn't a permissive policy, it's a "Go on, try it, we'll rip your balls off during review" policy.

in reply to orfanik

@orfanik @kde TBD...

circumstances.run/@davidgerard/117311106218081864


lol, the attempt to ram through a pro-slop KDE policy just got closed yet again because people found out about it

invent.kde.org/plasma/plasma-workspace/-/work_items/187#note_1648917


in reply to Cat ๐Ÿˆ๐Ÿฅ— (D.Burch)

@orfanik

floss.social/@kde/117325537492794402


We apologise if any of our messaging over the past days has created confusion. For the sake of clarity:

1. KDE is open, diverse and accepting and we have the track record to prove it. Please come join us and see for yourself.
2. KDE does not harbor "fascists" and we have the track record to prove it.
3. We are not writing a set of "pro-LLM" policies. Quite the contrary. The current draft actually restricts its use much more than currently codified.

[Continued >]


in reply to Cat ๐Ÿˆ๐Ÿฅ— (D.Burch)
not sure how strict your requirements are for inclusion here but Inkscape has a few (openly) claude AI 'assisted' commits, i.e gitlab.com/inkscape/inkscape/-/commit/eb8e143091e1
in reply to voxel

@voxel The full LLM policy is being actively developed but appears to lean towards banning generative AI

gitlab.com/inkscape/inkscape-board/-/work_items/5#note_3140424828

Oh, and there's also this:
social.treehouse.systems/@dysfun/116182391411729473