I've been in the weeds there, it's pretty decentralised on a architectural level. I see ATProtocol as analogous to Nostr but with 99% of the users on Primal. Fixing a hypothetical Nostr like that doesn't mean fixing the Nostr protocol. There's not much really to "fix". What it means is some Damus emerging and taking market share from this Primal behemoth.

Not too much the Bluesky team can do but encourage the makers of these future Damuses to speed up. What else could they do? The protocol itself is already open source, credible exit already assured, etc.