Post #4009248
2026-07-22 10:01 UTC
Replies (2)
-
@tomgag@infosec.exchange 2026-07-22 12:45
And, just to be clear, since what I wrote seems prone to misinterpretation: If there were a magic spell able to flag with confidence code generated with AI by more than X% amount, whatever X or surrounding policy is, then it would all make sense. But I don't believe in magic, and you shouldn't either. Here are a few examples of what I'm afraid it's going to happen: FLOSS project "OpenHemorrhoid" splits over an intestine dev dispute on tabs-VS-spaces. Now we have "OpenHemorrhoid" and "LibreHemorrhoid" accusing each other of AI-generated code at every PR. Indie dating sim game "Kawaii Shinobi Kanojo" is targeted and harassed with unproven AI-gen reporting because the lack of nonbinary characters is interpreted as a queer-phobic political stance of the dev. Libre "ICE tracker and reporting" app is taken down not by US authority overreach, but by mass allegations of AI-generated code. Well, I hope I'm wrong, the community is very mature after all, I'm sure none of this will happen. #sarcasm
-
@emilyskidsister@hachyderm.io 2026-07-22 10:57
@tomgag@infosec.exchange I'm going to move off of #codeberg since it's not helpful to argue over what "mostly" and "written" means, but I'm glad there is a place for this kind of thing. I'm still glad I donated, and will continue. They've help ensure there are options that aren't large US companies. It took a coffee break with glm to deploy Forgejo. I'm hoping there will be a good way to accept drive-by PRs for my small projects. imo, it's okay that it isn't Codeberg.