Bo Peabody recently had an Op-Ed piece in the Washington Times (via PEHub), titled "Twitter.Org?", where he suggested:

I launched the social networking site Tripod in 1995. By 1998, it was
the eighth-largest site on the Web. But Tripod was never a successful
business. Social networks aren't great places to advertise. You can't
charge users for their services. And they never gain enough momentum to
survive in the stock market. Indeed, no social network has ever made it
as a public company.
…
Instead of expecting profits that won't materialize, the
entrepreneurial community should instead operate social networks as
not-for-profit organizations. Wikipedia has grown phenomenally with a
not-for-profit business model, and while Wikipedia has its problems,
its fate is in the collective hands of its users rather than in the
hands of media companies or the stock market. Facebook and Twitter
should enjoy the same comfort.

Bo Peabody is a social media veteran and the article is a personal one.  It's explicitly rooted in his experience.  His point can be valid if you take social networks and user-generated content from a media perspective.  So I would agree with him in the case of the likes of MySpace or Bebo.

However, he points out Twitter and Facebook specifically, and there, I disagree with him.  This blog has always seen Twitter and Facebook as attempts at owning the identity layer of the internet.  Until Facebook, there was no meta data on the identites of internet users.  That's been changing rapidly over the last couple of years, and especially last year, Twitter has emerged as a contender in the identity layer – one that perhaps has a thinner layer of data on the users, but can grow even faster due to its simplicity.

I also find the notion of a not-for-profit, or utility-like, Facebook or Twitter very interesting and worthwhile. However, my reasoning would not be due to their limited profit generating ability, as proposed by Peabody, but the idea that something as powerful as the identity layer of the internet perhaps ought to be a public-domain asset, just like the internet itself.