I don't see why a blockchain or "currency" would be needed. It adds a large administrative overhead (the blockchain, a diverse pool of peers, etc.), is exclusio

I don't see why a blockchain or "currency" would be needed. It adds a large administrative overhead (the blockchain, a diverse pool of peers, etc.), is exclusionary as you say, and causes (a weak form of) centralisation. The benefit of a blockchain is having arbitrary adversarial agents reach a global consensus (such as who owns what); but that's much stronger than what's required for blocking spam. You can make spam expensive by using a proof-of-work system like HashCash ( https://en.wikipedia.org/wiki/Hashcash https://en.wikipedia.org/wiki/Hashcash ), which is the idea that inspired mining. HashCash adds proof-of-work to a message, which is specific to that message. The receiver verifies the proof then discards it. There's no datastructure to maintain, no consensus required, no need to involve anyone other than the sender and receiver, etc. HashCash didn't catch on for email, since its 'fire and forget' nature prevented senders and receivers from negotiating a non-zero amount of work. Such problems can be avoided in this kind of real-time protocol: just add an extra step to ask the receiver how much work it requires, then do the work, then send the message. With such a negotiation layer, there's no need for every message to require the same amount of work. You could receive the message without hashcash, and reply with a cost based on how spammy it looks. Everyone would be free to implement this however they like: give a high cost to everyone; check senders against public blacklists; give a whitelist of 'friends' a lower cost; reduce the cost if we've previously 'liked'/'followed'/etc. the sender; look up the sender in a pagerank/web-of-trust/friend-of-a-friend/etc. database; use a spam filter to score the message content; etc.