Interesting. Everyone seems to be experimenting with better ways of deciding what from Nostr should be shown to us. I'm increasingly curious about a different problem: what if the client tried to identify conversations worth entering rather than content worth reading?