If relays become private-read only, which I think will be a common (but not dominant) scenario, they'll have to self-index.
For as much as I like https://nostr.band/ I think it will be sad if we end up with a handful of search engines indexing the entirety of the Nostr public network. If Nostr is successful it will be hard for hobby search engines to exist, right? The amount of stuff will be too huge.
Does it make sense or does it scale to have search engines that prune content and only keep references? Or search engines that only index content from the last year from a set of relays? Or only from a set of public keys according to some algorithm? I don't know.
Discussion
maybe mix in a choose your own read more nip with private-read only so like sub stack, you still index the top of the articles publicly