The mitigation is of course more companies and open source projects getting involved. As far as I understand it, the protocol is pretty well designed, and there is a bunch of open source out there already. It just needs more people to get involved.
But the article pointing out that you can migrate your Bluesky account to a thing called Eurosky was actually news to me. As is the existence of non Bluesky owned applications on top of atproto. I might give those a try actually. It seems a lot has happened since I last looked at this stuff a few years ago.
That still leaves governance of the protocol as a problem. That would ideally need to be untangled from Bluesky as well. But that seems doable as well. The more independent atproto software projects are out there, the harder it is for Bluesky to make breaking changes. That naturally pushes for some independent governance. I wouldn't be surprised to see that happen over time. But worst case, there could be some kind of fork/split of the network.
It would be interesting to see some attempt to bridge with other networks. A mastodon / bsky hybrid. Why make people choose? Despite X's demise, it's still X versus everything else. There's also Meta's Threads and a few others. The whole "not X" space fragments users over multiple networks that sort of could federate if they would but they really aren't. I think Threads is nominally mastodon compatible at least but actually federating is a bit frowned upon by Mastodon users.
And while we are "at" it, why not support email as well? Which is the og. federated communication network with essentially all internet users as active users. It's not perfect, but nothing better ever managed to replace it.
Separating identity from protocol is a good design choice. Cross protocol federation is a logical next step. It's all about receiving content from people, not about staying in walled gardens that belong to share holders.