"Crackpots"? Is this a serious, honest comment? Or is a crackpot anyone who doesn't choose and create philosophically and technically inferior products for the sake of "Just one more social network that will surely be worth millions if I can make it centralized and proprietary and non-inter-operable and get just enough users." SOUNDS GREAT.

"P2P is harder to build" is an equally loaded statement, though you've caught my intrigue. I wrote a WebRTC signaling server in a weekend that is capable of letting peers establish N:N connections with each other for video conferencing.

Outside of cases where TURN would have been necessary, the clients were all connected, P2P with encrypted connections. (And even with TURN, I believe end-to-end encryption could be ensured, but the traffic would go through a third party)

Sorry, I didn't take the time to sugar-coat my comment. One lesson I've taken from Web 2.0 over the last decade is that users do not care. So why should I care about them more than they care about themselves? We'll see if PRISM changes that, but I suspect it won't.

WebRTC makes P2P a lot easier, especially since it doesn't require installing software. I think getting peer connectivity is 10% of the problem. Video conferencing is pretty much a best case for P2P; general apps are much harder. Things like data sync and handing churn are very complex and not needed in client-server architectures.