Rate limit against spam. Don’t rate limit against normal user activity. It’s not that difficult to understand what I’m saying. You are simply wanting to be argumentative. Please kindly fuck off.
Detecting spam is the trick. New npubs could have a probationary period until "vetted". WoT-based lifting. In general, the incentive ought to be to create content with which other trustable npubs want to interact. Bots will always create multiple npubs, so a simple blacklist for spam isn't enough.
That pushes the filtering burden onto the best/heaviest users, tho, which will push them off the relay. Only useful for an onboarding relay, or something.
We could set up a charity for an onboarding relay.
That isn't a crazy idea, right? We just need one client-independent relay that every client may offer as a default, for 30 days per npub or so, so that new users have someplace to start off from. nostr:nevent1qvzqqqqqqypzphtxf40yq9jr82xdd8cqtts5szqyx5tcndvaukhsvfmduetr85ceqydhwumn8ghj7argv4nx7un9wd6zumn0wd68yvfwvdhk6tcpr3mhxue69uhhg6r9vd5hgctyv4kzumn0wd68yvfwvdhk6tcqyqseu50rhpcps56a4x09yp8ztfqnru3fhwuwx56r8w2qsd0msakjudvnlgq
This seems like general infrastructure for public use, that benefits all users by making the protocol more attractive to newbies, rather than something we can expect individual client developers or relay operators to finance and run indefinitely. Otherwise, we risk skewing viable business models and getting newbies trapped in a client. This would also even the playing field a bit between clients, as currently only the VC-funded stuff can afford "free to use" relays. That might spur innovation.