Tim posted below on this blog about the new generation of social-networking sites and whether they can attract another round of users. (What are we on now, the third generation since Friendster launched in 2002? These things have almost as short generational cycles as fruit flies!)
As he rightly points out, there are many incentives but also challenges to bringing out a social network at this point in time. It's similar in some ways to the challenges faced by serialized TV shows that came in the wave after 24 and Lost.
Serialized shows where a story arc crosses over an entire season (or even across seasons) require an investment in time, memorization of facts, and emotional connection with the characters. The fast-follower round of shows had a harder time because people have a maximum capacity for these things--they can't keep investing in numerous shows. So the first ones out of the gate tend to stick around and the followers drop away.
Social networks are the same way, people can only invest so much each time a new one appears, and the more of them there are, the more niche they are likely to be, simply because of the math. If future social-networking sites are going to succeed they will probably have to be niche-oriented rather than mass-market like Friendster, MySpace and how Facebook has evolved. Either that, or we'll see an emergence of "meta" networks that pull together friends and content from across them all. But that requires opening of standards that so far only Facebook has done.
We may also see social networks capture a certain generation of users (within a tiny two- to three-year timeframe) who stay with the network and "age" with it, and successive generations will adopt new networks. You network just with your cohorts, in other words. This happens in many other industries (in cars and music people grow up with certain brands and bands, which evolve with their audience), but seems untenable in the long run with social networks because we rarely just want to network within our cohort group.
That's not to say that someone can't come along and figure out a quantum leap improvement in social networking by addressing an area of user experience no one else has tackled. Apple was not the first to make mp3 players, and Google was not the first search engine. Each succeeded by being an order of magnitude better in user experience than the existing offering. But neither one was in an area where switching costs were high (mp3s could easily be moved from an old player to a new iPod, search requires no time or emotional investment from the user to get payback). This is why both Apple and Google have built stickier networks of services around their core offerings--to increasing the switching costs. At this point it is going to be difficult for anyone to seriously displace them in the mainstream.
The best thing for users, and probably for the social-network providers themselves in the long run, is to open up standards and API's in the same way Facebook has done, and in the same way that happened with IM platforms. A rising tide will lift all boats.