I don’t understand the bottleneck. Where is it slow and where is it compute intensive? On the individual relays? On the clients.?

This is why many ppl are not using Nostr as much as before. I personally can't afford replacing my devices every 6 months. So if you guys can look at it, it would be great.

It should be a bug.

See it on Primal where it causes duplicates. See it on Amethyst where it eats battery and heats phones. I worry about relays becoming smart in order to address these issues. Gotta keep the pipe dumb, focus on problems at the edges, and make protocols smarter.