There's a blindingly obvious scalability problem with NIP-05: https://github.com/nostr-protocol/nips/blob/master/05.md

Homework problem: what is it? How could it be done better?

Reminds me of how old versions of git didn't use to validate hashes by default. A terrible idea that they were forced to fix.

Particularly when nostr security practices are still awful, eg how often people end up posting private keys into random websites.

note1a05e89lfjqprptagynvv6jhcmqhjfl7mp940jhq65yeu8pmfrlxsg65gt0

On chain payments in the last three months haven't been much more expensive than Lightning.

If inscriptions continue to keep blockspace in high demand, that graph will be much more interesting over the next three months.

Does nostr not support direct image relay because it doesn't actually scale? 🤔

It's a revealing design decision. Fast computers and short messages let many sins slide.

You should define what your threat model actually is, and what specifically is the use case. Eg delegated decryption for email is very different than group chat.