For real, trust online is an open and hard problem. It's only going to get harder with ai bots running amok.

I have some serious designs for a federated reputation system that is, as far as I can tell, novel but I haven't had time to really refine it and develop a proof of concept. Just a pile of notes for now

	▲	verdverm 4 hours ago \| parent [-]
		Have a look at how labellers work in ATProto. It forms a good foundation imo, perhaps sufficient as they currently stand. Good prior art if we abstract beyond just atproto, not sure what W3C might have already in the works that is similar enough. https://bsky.social/about/blog/03-12-2024-stackable-moderati... https://roost.tools is another group you may look into. They are broader in scope for Trust & Safety across the internet at large. Their current focus is a couple of OSS tools for builders, but the ambition is big and something to appreciate.