Short version: Founders of companies often have some typical biases that help them create companies, but also can cause risk later.

Seems to me that some of those biases are useful per se (believing in one's success), and some of them increase the variance of outcome, thus increasing probability of extreme success, but also of extreme failure (black and white thinking, willingness to destroy already existing solutions, hasty extrapolation).

The former describe a valley of bad rationality from inside: if you already have a bias, another bias may help you by reducing the first bias. For example, if humans have an instinct to "not act above your social status", and our social status calibration mechanisms are ruined by moving from small tribes to global economy, then believing in your personal excellence may give you enough perception of status that you finally dare to do things which actually didn't require that much status in the first place. If people get too easily discouraged by a probability of failure, then not believing in failure may give them enough motivation to work.

The latter seem like a variations of "wanna be a millionaire? buy a lottery ticket!". Buying a lottery ticket may be a loss in average, but at the end many millionaires may have buying a lottery ticket as an important step in their history. Actually, maybe the lottery tickets in average are not as bad as we would believe, if there are cognitive biases preventing people from buying them (in the metaphor of "buying a ticket" = founding a company; of course in real life with real lottery our cognitive biases make us buy more lottery tickets than would be optimal). If you choose a random idea and follow it blindly, there is a small chance of huge success (and a large chance of just being stupid). But if you overanalyze things and spend your life procrastinating, you will get nowhere.

My guess is that with human brain the best you can do is to separate the analytical aspect from the motivational aspect. (If you try to do them both as the same time, you will get horrible outcomes at both: your thinking will be "motivated", and your motivation will be ruined by keeping the possibility of failure too salient.) One option is to make a rational choice, and then close your eyes and start running forward. This would not work well for problems that only appear later. Other option is to split the task among multiple people: someone creating an optimistic vision (e.g. Steve Jobs), someone taking care about the actual risks (managers). In our world, the optimistic person would be the founder, but they need to hire a competent manager soon. Yet another option would be to maintain proper compartmentalization; for example to have special regular meetings where you consider things as rationally as possible, but outside of the meeting room it is forbidded to express any doubt in success.