It seems we're at a bit of a crossroads. It seems like the world both needs:

- Permissionless email (i.e. for agents, empowered users who can program now)

- Pervasive email allow listing

Wonder if these can both exist at the same time, i.e. having a "public" email that is read first by AI (let's imagine we're in a world where prompt injections weren't so possible) and heavily filtered, along with one that is private and allow-list gated (via some easier-than-gpg-to-use identity marker).