Post #4823618 · Elektrine

2026-09-22 15:51 UTC

@ska@social.treehouse.systems @ell1e@hachyderm.io @aelspire@aelspire.info @ariadne@social.treehouse.systems There will be backported LLM-assisted commits, and I also don’t consider every LLM-assisted commit to be slop. To me, slop is LLM-generated code of low quality. Code that is of high quality and been thoroughly reviewed by humans is not slop. I have Assisted-by tags in some of my contributions. LLMs have found bugs that I missed. It has found vulnerabilities in third-party code I didn’t spot. I will never submit code I cannot personally vouch for. The only possible exception would be machine-checked formal proofs of a theorem whose claim I have reviewed. in that case I trust the proof checker more than I trust myself.

Replies (1)

@dalias@hachyderm.io 2026-09-22 16:58

@alwayscurious@infosec.exchange @ska@social.treehouse.systems @ell1e@hachyderm.io @aelspire@aelspire.info @ariadne@social.treehouse.systems To me slop is a matter of provenance not quality. If there could legitimately be unknown copyright claims on the code because it's a derivative of some unknowable part of the corpus in a meaningful way (i.e. large enough to be subject to copyright), then it's slop. It's also slop even if it's smaller than that, if the submitter does not understand it. If it's small and the submitter understands it, I still don't want someone working that way contributing, because they're degrading their own cognitive capability and likely to push for pro-AI policies that drive away good contributors. But at least it's not tainting slop.

Open ##4823616