On some level I'm surprised this approach isn't more common. With diminishing marginal utility of wealth, a small chance of making millions has a much higher expected value than an almost zero chance of making billions. Plus, even if the billions were worth proportionally more, we actually tend to underweight high risk/high reward scenarios compared to a perfect expected value maximizer. E.g. you wouldn't pay much to get a 1/10,000 shot at unimaginable bliss, even if the expected value math works out.

I'm guessing part of it is that at some point, say under a 5% chance, we stop calculating altogether and it just becomes "so you're saying there's a chance!" If you have a 1% chance of making something as big as Basecamp and a 0.001% chance of starting Facebook, they both feel about the same, so might as well shoot for the moon.