krisHow to deal with LMM slop in FOSS projects, regarding discussions on labeling/information
The way i see it there are 2 clear positions (non exclusive)
A. People who don't want LLM output in their lives, like, at all.
B. People who don't see a fundamental problem with LMM output, as long as X Y Z
Getting people from group B to attribute their commits to AI, or setting up rules around it immediatly goes into hair splitting about what constitutes LLM use. Hard or hopeless case. Easy concern-trolling target.
Getting people from group A to label their contributions #noai, everyone in A would be more than happy to have a clear way of communicating this.
In a way, we're retreating to a "slop until proven innocent", but it can also be a way to turn this into something positive.
Abstinence in itself can be a positive thing, and we should try to communicate this.
What i'm still not 100% about is what positive term we could use, #noai is fine, but it's inherently negative. I'm personally a bit allergic to generalist humanism, and reluctant to use "made by humans" etc. Organic software comes to mind, CREATED BY MOIST ENTITIES OF FLESH is another option. If anyone has suggestions here please do!
Feel free to boost for reach
DecksDark