It never hurts to repeat a good point: The silos we're all moving into now can all be leapfrogged. The second the silos aren't simplifying, but instead hampering, innovation and expression, they'll start to loose. Maybe they are beginning to loose right now. Google seems intent - in everything but search - to make sure the silo is just a convenience, not in-circumventable fact.
In other silos-are-breaking news: The T-mobile/Sidekick catastrophe, and we could end up on a modern, open, federated communications platform instead of a silo, and a silo so stressed by growth, that it's really hard to get data out of it that is more than 2-3 days old.
Googler's current and ex- are also still working on moving people to open social networks. Open social maybe wasn't really the sweet spot that one hoped for, but new projects are trying to get there. One wonders along the way if Twitter is moving to a Mozilla business model? Free to use, we just sell the behaviour of our users.

Facebook is going to have move as well - just because of the things you can do with data you can get at. What I'm trying to say, really, is that if you plan to be doing something interesting in two years, I would personally bet on desilofication over "better, but closed" data, just on general principle of cycles.