Currently, we have all [our social] groups siloed. Because we have many different contexts and levels of intimacy with people in these groups, we're inclined to use different systems to interact with them. Facebook for gaming, friends and family. LinkedIn for customers, recruiters, sales prospects. Twitter for friends and celebrities. And so on into specialist communities: Instagram and Flickr, Yammer or Salesforce Chatter for co-workers.
The situation is reminiscent of electronic mail before it became standardized. Differing semi-interoperable systems, many as walled gardens. Business plans predicated on somehow "owning" the social graph. The social software scene is filled with systems that assume a closed world, making them more easily managed as businesses, but ultimately providing for an uncomfortable interface with the reality of user need.
An interoperable email system created widespread benefit, and permitted many ecosystems to emerge on top of it, both formal and ad-hoc. Email reduced distance and time between people, enabling rapid iteration of ideas, collaboration and community formation. For example, it's hard to imagine the open source revolution without email.
Dumbill seems not to have noticed that the various services he mentions, from Facebook to Twitter to Instagram, are already built around an "interoperable system": it's called the World Wide Web. Those aren't incompatible platforms, they are merely services you have to sign up for — just like Google.
Ah, but, "Though Google+ is the work of one company, there are good reasons to herald it as the start of a commodity social layer for the Internet. Google decided to make Google+ be part of the web and not a walled garden." Well, yes and no. You can see Google+ posts online, if the poster chooses to make them public, but you can't participate in the conversation without signing up for the service. In other words: just like Facebook, Twitter, Flickr, and so on.
In the end, it seems to me that Dumbill is merely saying that if all of us decide to share all our information with just one service, we'll have a fantastic "social backbone" for our online lives. And that may be true. Now, can we stop to ask whether there may be any costs to that decision?