If it's true that for a social network to be successful it needs some quick growth at the beginning to have a critical mass, then Diaspora is in trouble. They're in trouble because they release non-usable alpha versions of their code, loose the buzz ("Diaspora released another version today, X works now, but it's still not good enough to replace Facebook."), and by the time they have a 1.0 version nobody's going to care.

I'm saying it might not make sense to develop software in an open-source manner whose existense relies of viral growth of users dependant on a one-off buzz generated when the 1.0 is released.