Using an LLM to filter content and create topic-focused relays is a great idea. I’ve thought about having admins review posts before, but that’s obviously far too inefficient.
Discussion
you mean creating opaque twitter algorithm on a server to censor content