Please link to your other feedback, I didnt see it.
At scale, Nostr ultimately needs multiple relays running at enormous scale to have any hope of a large "marketplace" (twitter, telegram, whatever). As the network grows, amateur relays will fall away, and you will be where we started with a handful of major providers.
The idea that there would be reliable relays with consistent data for many people is fantasy. There will be what there is today, businesses offering hosting/servers for users/businesses, and a few people hosting things themselves.
So, Pubky is designed around giving users a way to easily switch their provider if needed.
There is no question where to find me, or which sessions are authorized. There is no putting private keys directly into applications.
If I want to have a Pubky business profile/account, i don't have to share privkeys or make retarded future-proof msig schemes.
I can keep my keys cold, and hot swap or modify my endpoints in a verifiable and secure way while taking advantage of any (compatible) centralized provider I prefer.
To be fair, many Nostr devs realise that spraying and praying data on relays isn't sustainable for most use cases, and the Outbox model seems to be a step towards people entering a relationship with a specific home relay.
But if we take that to its extreme, then if the few big relays are only serving to facilitate discovery, then these Outbox relays could be better as HTTP api servers not websocket. Blossom seems to have opted for that which is good.
Nostr does not need multiple relays at enormous scale, that premise is completely incorrect. Users don't need view of the entire network at all times, they tend to group into circles, that's why Americans bitcoiners are seldom served up Indonesian politics tweets etc. And those circles can easily run on a handful of relays.