Very true. It also makes a difference as to which resource is being pulled, whether it is cached, what transport is being requested (SSL, compression, etc).
I really suspect the website would fall long before it hits anything close to 4.2 million requests (which the author also seems to except).
Visitors don't come neatly one after the other. You might only have 1M requests a day but get random spikes with 100 requests at the same time.