The most important seem to be that it assumes people are risk-neutral and optimizing their expected wealth, instead of trying to reach goals like maximizing the probability you'll be able to buy a house in SF. If that were the goal, I think early stage startups would fare incredibly worse in this analysis.
The next most important seems to be that there's no way for a company to fail in this model. The companies that become worth less can't become worthless. Since these early private companies' stock aren't liquid, you can't cash out before they fail. You get zero dollars from those, which is a huge difference from the model.
And the whole crazy tax thing. If you quit as this model suggests, you almost always have to exercise within 90 days (spending a bunch of your money for an uncertain future). Then the IRS sends you a huge bill if you left a successful company where you might get more than zero dollars. But you still haven't earned any cash from the stock.