Users sign up for email hosting. They generally know if they are using gmail or fastmail or proton mail. And they generally pay for it. In a healthy nostr ecosystem it would be the same. Users would know which relay service(s) they signed up for. So the question then is: how much do they know about other relays?

If someone says "come join my chatroom on relay X" then they will know about relay X. It wasn't so hard.

If their client says "this nasty message came from relay Y, would you like to block relay Y?" then they can decide. It wasn't so hard.

Of course few people are going to study all the relays and make judgements about them all outside of specific reasons to.

I'm of the opinion that for outbox model, and for some applications on top of it, users don't need to be concerned about which relays are involved. I like hiding that detail. The relays can just magically work. But when you care about censorship or moderation or bad behavior, generally at that point you'll want to find out if you should cut off an account or perhaps an entire relay. The context (kind-1/1111? relay chat? global?) matters a lot.