I do agree experimentation is useful, but some things will never produce useful results. Sending data over Nostr is known to be inefficient: it has a lot of overhead for everyone involved. While nostr identities as websites may be useful, this can be achieved in simpler ways like a replaceable event that has hostnames and have actual utility for things like relays with public key identities. (DNS/IP/I2P/Tor/etc addresses)