The architectural design and the caching system will change so much that the more you delay the more you will have to rewrite later on.

Why should we use outbox for this use case? Can you explain it like I'm a noob?

Because: 1. centralizing is bad. The main relay set can moderate everything. 2. the fixed set of relays will know the location of each one of your users. 3. You want people to choose relays that are physically closer to them for performance. 4. If I follow people on Nostr and they are using bitchat, I don't want to have to move to their location to talk to them, I just send to their relays. 5. Relays can become topical communities within locations.