The future of online identity is indeed decentralized and not distributed, meaning that users will always have some super nodes to handle their identity on behalf of them. In my opinion Facebook/Twitter/etc are not identity providers, they are silos. Sure they are very successful ones and can even used as identity providers at some places, but as long as they don't open up they can easily die anytime.

The author suggests that services built on top of these Silos that provide proofs of connection between all the identities. I welcome such initiatives and but I doubt they will lead anywhere, cause they are built on top of silos. And a silo, as soon as it figures out it loses money, it will cut down that connection.

What won't die is decentralized published standards and protocols that handle the Identity management through the internet. Starting from plain DNS, we can get AoR for SMTP, SIP, XMPP and on top of that we have frameworks that facilitate the identity management like Oauth2, OpenID etc. All open and standardized. We are getting there, we just need some more time I guess.

That's why I always thought that, Google, who owns emails has much more value than Facebook, that asks for your email. If facebook dies, you lose one aspect of your digital social part. If you lose your email though, you almost lose your online identity. I really can't get how Zuckerberg has missed that.