Over at the BBC, Bill Thompson attempts to burst the "social software" bubble before it even expands. Thompson is spot on with his criticism, especially the point about ignoring previous HCI, psychology, and computer mediated communication research. I would also add all of the effort by the CSCW community.
He paints with a little too broad a brush, leaving out such folks as the Social Dynamics Lab at HP, who really do solid principled work. But maybe that's his point. Those talking a lot about social software aren't doing much and don't fundamentally know what they're talking about. However, there are folks out there getting stuff done and investigating the frontier. They're just not overselling the short term gains and self-promoting their ass off.