PlayStation 3 emulator RPCS3 developers battling "AI slop code pull requests"

http://www.gamingonlinux.com/2026/05/playstation-3-emulator-rpcs3-devs-battling-ai-slop-code-pull-requests/

194 points · 15 comments · view on lemmy.world

15 Comments

DonutsRMeh@lemmy.world · 17 pts · 102d (3 replies)

Why do people do that? Is it fun or something? I've been using AI a lot lately, but ONLY on my personal projects. I've never submitted a single PR to any project with AI. It just feels wrong to be honest.

thingsiplay@lemmy.ml · 20 pts · 101d (1 reply)

Probably different reasons to do so. One reason I can imagine is, some users of Ai submitting slop genuinely think they are helping and don't know how bad the situation is.

mnemonicmonkeys@sh.itjust.works · 18 pts · 101d

Or just wanting to have a bunch of PR's in their portfolio

Anon518@sh.itjust.works · 2 pts · 101d

They can't code but want a feature/fix implemented and the dev team is limited?

thingsiplay@lemmy.ml · 10 pts · 102d (3 replies)

Unpopular opinion: RPCS3 is the best Playstation 3 emulator.

Squizzy@lemmy.world · 8 pts · 101d

Isnt that the very popular opinion?

deadcade@lemmy.deadca.de · 6 pts · 101d (1 reply)

... Are there others?

thingsiplay@lemmy.ml · 3 pts · 101d
artyom@piefed.social · 1 pts · 102d (6 replies)

Not to be dramatic but, is this the end of open source? Or at least the open contribution systems? I feel like I see this story every day now...

MrMcGasion@lemmy.world · 23 pts · 102d (2 replies)

Almost certainly not. All you would really need to do to stop this is limit who can submit pull requests to verified devs. Things go back to the way they were during the clunkier, before times when everyone used Subversion, when there were more hoops to jump through to contribute code. Make it so you need an existing contributor to "mentor" first-time contributors (making sure they aren't an AI, are writing competent code that follows project standards, etc.) before giving access to submit dozens of pull requests.

fluffykittycat@slrpnk.net · 4 pts · 102d

We really need a stronger mentorship culture in general

artyom@piefed.social · 4 pts · 102d

All you would really need to do to stop this is limit who can submit pull requests to verified devs.

That seems like it would greatly limit the number of contributors, and especially over time, since you'd never have any new ones.

chromodynamic@piefed.social · 3 pts · 102d (2 replies)

Need some way to sort contributors by "trust", e.g. if their PRs have been accepted before, if their account has existed before GenAI was invented, etc

Corngood@lemmy.ml · 6 pts · 102d

It should be possible to use a distributed web of trust for this.

https://en.wikipedia.org/wiki/Web_of_trust

artyom@piefed.social · 1 pts · 102d

So no one new is allowed to contribute?