We could also implliment TrustNet which was developed for Cabal & Secure Scuttlebutt

https://github.com/cblgh/trustnet

That would let me block some content that you're happy to see. Relays could look at the trust scores of its own hosted content to figure out blocks. Or rather than blocks, ways of separating groups of users…. you don't delete all spam, you just put it off in it's own corner of the world talking to itself.