The community around NOSTR are basically building a kind of semantic web, where users identities are verified via their public key, data is routed through content agnostic relays, and trustworthiness is verified by peer recommendation.

They are currently experimenting with replicating many types of services which are currently websites as protocols with data types, with the goal being that all of these services can share available data with eachother openly.

It's definitely more of a "bazaar" model over a "catherdral" model, with many open questions and it's also tough to get a good overview of what is really going on there. But at least it's an attempt.