OpenID might actually work
It seems like one of the big problems the majority of new startups face is the user-base leap. A good social site will have a much harder time attracting users unless the site already has a pretty vibrant community; of course, it’s hard to attract the first batch of users because your site doesn’t have that vibrant community, and to get that vibrant community…
The solution is, of course, to have a site that does something useful when the social aspects are taken away. Youtube was a pretty decent video hosting site - the community interaction came as a nice bonus over time. Delicious just happened to be the best online bookmarking service out there, and the community features - while central to the site - can easily be ignored.

OpenID is a service that faces these issues as well. Developers always have pages of feature requests, and adding support to a service with a handful of compatible websites is bound to take a lower priority then nearly anything else on their plate.
Luckily for OpenID, this hasn’t really been an problem. Sure, the service may have only been available to a handful of sites for quite some time, but that didn’t matter. People got their toes wet with the site and warmed up to the idea, and more sites followed suit, and during that stage of the service’s life there was still a great product to be found.
The best thing about OpenID, in my opinion, is that it’s a genius solution to a problem that’s starting to plague more and more of us. As more desktop-quality web apps take center stage, a service that keeps track of all your information and then makes registration and logins a breeze will become and even more enticing proposition. Until then, however, people are content to use it for just a handful of sites; of course, from the looks of it, they won’t have to anymore.