isn’t that just more computationally demanding for everyone. now i’m just spamming relays with many very-specific queries
You don't need to constantly spam relays if you keep a good local performant cache index which is what is missing in most clients.
From what I understand of Nostr and apps in general (which is not a lot) and from what I see connecting through Tor, I think this makes sense: nostr:nevent1qqst5qe2ff477l9lu4ms6xkyxurg7qkye5m7wdudnddhl8jhl3pd5tgpzpmhxue69uhkummnw3ezuamfdejsygxsm6lelvfda7qlg0tud9pfhduysy4vrexj65azqtdk4tr75j6xdspsgqqqqqqslsu9zc Maybe we just aren't there yet in Nostr development, the point at which you can and should develop a good local cache