Here's the tricky thing about privacy and social networks: Facebook's privacy support is actually pretty good. Whilst people might tell you in the abstract that they want more privacy from Facebook, figuring out what they would change in concrete terms is very hard. For example, they might say "I don't want to see ads" - but given the choice, they don't want to pay for anything either. So this feedback ends up being pretty useless, equivalent to hearing "I want everything and a pony". It's not a basis for a product.
Google learned this one the hard way with Google+. The original way Google+ tried to differentiate itself from Facebook was with circles. The idea is, Facebooks relatively singular notion of "friend" doesn't reflect the way real people work, this means it doesn't respect people's privacy and so people use the product less .... therefore by giving them better tools, they'd win a lot of users. Facebook responded that they'd tried the same thing, it turns out people don't like making lists of friends and controlling their sharing at a fine grained level, so it wouldn't work. And guess what? Facebook were right. Sure, you interview people in focus groups and they say one thing. In reality they might do something else.
So - decentralised open source social networks. Not gonna work. People might sound enthusiastic when you pitch it to them in the abstract, but actually Facebook works fine for them, and the kind of privacy that matters to them (can people see who views their profile?! Can my parents see my drunken party pics?) is already well supported and tuned.
Ultimately what will do off Facebook, eventually, is a change in how people use social networking that for whatever reason they cannot replicate in their main product.