Remember that in issues of optimization, the question is always, "for what am I optimizing?" Startups have to scale as quickly as possible, or they'll run out of money or be supplanted by a competitor. A startup team pumps all their effort into a 2 to 5 year period, after which point they'll likely have achieved dominance in their market, been bought by a larger company, or failed. The game is short, and the payoff is bimodal (either very high or close-to nothing). Workign one's self into the ground for a few years to come out a millionaire maybe worth it.

In contrast, the employee at a typical large corporation wants to have his job (or a very similar one) and his salary (or one somewhat higher) for the next ten years and his employer similarly wants to have the continuous use of his labor.

A good rule of thumb is that if somthing seems insane or just wrong, consider the incentives involved. Many times it is actually insane, but just writing it off as so, doesn't help you understand it.