Ask HN: Is MVP overly pessimistic?

The idea of MVP seems very in vogue right now, which I think is appropriate in the context of lean times and lean startups.

However, it's occurred to me that MVP may be overly pessimistic because it ignores potential network effects.

For example, a signup flow that ends in a "not available now/just kidding" kind of screen is NOT going to a result in a potential new user tweeting to their followers about your service. It will probably end in a modest amount of frustration. In light of this, perhaps MVP cannot be applied at overly broad levels - would it represent too much of an underestimate?

Maybe the real question is -- can you MVP your entire business? Or only features, after you've launched?