[yason]

"Imagine if the guys at Google just sat around waiting for their server to kick the bucket before they fixed it; they'd be out of business."

Funny that you said that. Somebody correct me if I'm mistaken but I've understood that's exactly what they do at Google. Due to their scaling, self-replicating, and dynamic architecture, they figured out that it's cheaper to just let the boxes fry and replace them rather than trying to maintain and service their gazillion servers. I don't know how many boxes they replace each day but I'd guess it's routine.