My advice: Take notes. Don't work on it.

Sure, there are a dozen things I would do differently if starting over including using a different elliptic curve, pushing for binary packets, giving relays identities, negotiating new features in-protocol, event kinds being bitflags, having slightly more composibility in some places, etc. But nothing about the way it is now breaks things signficantly.

If you think there is going to be a successor protocol, would you pour your heart and soul into the next nostr project? Probably not. I don't think nostr is broken, so it doesn't require fixing.  I'm all-in on nostr, including starting a new project soon.