You heavily under estimate the network effect.
If that was true, terrible sites like eBay and Craigslist would have been replaced a long, long time ago.
specially ebay which is hated among even it's most loyal sellers refuse to leave ebay, because they have buyers. The network effect is even stronger with Facebook. I highly doubt that we are going to see another social network get 900 million users anytime soon.
Facebook's advantage over any competitor is that everyone is already on Facebook. I could see early adopters moving to what you're describing, but it would take ages to move the other 90% of the users over. So at least for the next few years, it's unlikely that people will transition en masse to a completely new service.
They're on FB. They're also on Google. A hell of a lot of them are still on Yahoo mail. They're using Craigslist. And eBay. And Amazon. And The New York Times and/or Fox. Hulu. Whatever Apple's calling its offerings these days. Even Microsoft. To say nothing of a bunch of B2B service sites that have large user lists (ADT, Delta Dental, Beico, etc.).
There are a bunch of sites that have a very large userbase. Large enough at least to get past the network effects problem sufficiently to start boostrapping social onto things if they want to do so.
Which is one of the reasons that network effects, in and of themselves, offer such transient advantages. You may be able to corner the network effect in one area for a time, but you're not going to conrner all network effects in all areas, leaving you vulnerable to intrusion and erosion.
Several of the players listed above have a vastly better advertising or direct revenue capacity (Google, Apple/iTunes, eBay, Craigslist, and most of all in my mind, Amazon).
Craigslist is a wonderful example of a site that leave huge amounts of potential revenue behind (charging for only a small fraction of jobs and housing listings) to create a hugely compelling advantage in the classified advertising market, and supporting some "social" features (there is a pretty active, if not grossly overwhelming, forums section). Amazon also has discussion. Google has entered directly into the frey. The direct problem of social networking is one of coupling a modicum of engineering talent to a financing capacity and and existing network.
I've said it before and I'll say it again: social is mostly a way of tying notices and feeds to specific users, and there's absolutely no inherent reason why it needs to be based on a centralized application model. Diaspora is already a community-based project, and may be heading toward a decentralized hosting model. Once that happens, IMO, the jig is up. Not because Diaspora will instantly subvert Facebook, but because all the 2nd-tier social networks will have the option of joining FB (or whomever leads the siloed SN space at the time), ceding significant power and control to them, or of playing in the open space.
It's the same strategy that's lead Linux to subvert all other operating systems on the server, embedded, and handheld/mobile markets (less successfully still in desktop/laptop). Sun, IBM, HP, LG, Samsung, HTC, Motorola, Google, Amazon, B&N, etc., couldn't take on Microsoft directly with their own OS offerings, but they could join forces by agreeing on an open standard which offered none of them a proprietary advantage of itself, but allowed them to exercise other business strengths to benefit (and yes, Apple's been doing quite well with an alternative strategy, though also based on an open UNIX base).
My point is that the underlying functionalities of Facebook are features of the web in general, not of Facebook specifically.
Facebook made them simple to use and put them all in one place and now has grabbed an impressive amount of share as a result, but they are not on the most solid of ground, as their social functionality could easily be swallowed by the browser itself, or by a clever lump of javascript.
And remember, the inline advertising is only required by their business model, not by the functionality of the service itself.
[edit] That said, I sort of agree with you that they probably have a few years before the mass exodus, I wouldn't give them a lot more than that however.
