I think your AlgoRelay can be used to protect against spam, but allowing the user to configure the algo and share the score with other AlgoRelays. If everyone run their own AlgoRelay and shared the score with one-another, that which everyone agrees on is spam and/or badly tagged, etc., would be naturally and organically filtered out of existence even though there is no centralized censorship/moderation. One could even imagine running _different_ such relays, thus creating differently curated feeds… heck someone who loves spam, could have a special feed for that too. :)