Most of my opinions about nostr's architecture come from a conversation I had with a friend of mine who used to work at Twitter. He explained that Twitter is only able to deliver relevant content quickly by pushing it to a huge network of special purpose caches. This is easily done with nostr relays in theory, the hard part is coordinating how caches (relays) get primed. I think some combination of push and pull will be necessary - if indeed the problem can be solved without central coordination. 

 So like having a Smart Relay? 

 I definitely think that's part of it 

 What else do you think we could need?  (Nymble is looking to commission a smart relay in the near future and would like to get some perspective on what else should be built into it)
 

 Well-established strategies that people can expect smart relays to implement. DVMs that can perform aggregation/analysis/discovery on relays, including AUTH-protected ones. Clients that can use smart relays. Conventions for using different relays differently. Lots of things, mostly based on establishing good conventions and a shared mental model for this stuff. 

 DVM is definitely something we want to build.

We were also thinking of setting up a page (at a .bit domain controlled by the client) that would act as the canonical source of information for the .bit name holder so that clients can be guarantied to receive every note the people they follow have ever posted (or at least the ones they want them to see).  An added benefit is that it seems that clients would be able to pull information faster with such an architecture.