Oddbean

These are valid arguments, but we can get a bit more creative here, and put our minds in problem-solving mode. There's multiple arguments presented here, so I'll just tackle one. And that is, "its too expensive". It's only expensive if you lose the base layer. It would cost you about 300 sats to timestamp your profile. That might be a worthwhile cost for those that value their profile. Furthermore on higher layers, you can get the cost down to something nearer to 10 sats, or even 1 sat. All be it with some security trade-offs, which the vast majority of end users wont notice, especially if the UX hides it. Even better, using testnet, you can get the cost down to zero, and the system would still work. Bear in mind, im just using the blockchain argument here, and not even talking about other options. Consistent data is useful for all sorts of things, and we can get the cost down to very low levels. Inconsistent data is problematic, and I'm not sure how easy it is to argue that it's a feature not a bug. It might be in some cases, but in another, or lot of cases, consistent data is desirable.

Indeed, timestamps are really useful in nostr notes. I agree that inconsistency is often undesirable, though I'm hesitant about concluding that relays should coordinate. It's my nostr client's job to ensure that the content is consistent across all the relays that I use. Perhaps I should be using fewer relays if I'm unable to maintain sufficient consistency across all my relays. It's enough to have just one relay with the outbox model 😀 I realize the UX suffers from having all the business logic in the client, but where do we draw the line between Twitter and nostr if our relays keep getting smarter?

Short answer: relays dont get smarter, clients do. You provide the proof in a tag, and it can be verified by the cllient. The client knows when a profile is the latest, and when it is stale, that's about it. This also opens the door for other use cases, other than profiles.