Interestingly, most bubbles end in a scenario where the decline of a popular asset causes a domino effect through out the system. Silicon valley might by in a 'facebook bubble' - we anchor all of our prices on facebook, and then hope like hell that it isn't over valued.

Leaving out the fact that Alex's data is wildly inaccurate that is probably the worst possible way to try to value a company. The most obvious issue is that Pinterest has only been successful for a few months. There's a long list of Facebook apps that were hot for a while then dropped out of existence. Pinterest may still be added to that list.

You have precisely identified the bubble, inadvertently:

Why are we determining what these companies are worth based on things like number of uniques and users, rather than the actual amount of revenue they bring in? Users, in-and-of themselves, are useless. Revenue is always and everywhere the end goal. Somehow we've begun to take at face value that more users=more money, and I don't think that's the case.

What if you don't agree that Facebook is worth $100B?

Alexa traffic measurements are too unreliable to base any kind value estimation on.