I don't know. I think we want some people to take risk. Spending a lot of time building something that you don't have any guarantee of being paid from is a risk, but for every 100 which fail, one will succeed in providing jobs and foreign investment. That is, in fact, the concept behind ycombinator in a nutshell. Not startups for the sake of startups, but startups for the greater good. And that means individuals taking risks.

Saying that working for existing corporations is more valuable to society underestimates the frequent and obvious value to society of the one-in-100 startup which creates new avenues of wealth for whole sectors of society. A dynamic polity should emphasize that and create more incentives for it - not even for the good of the individuals who do it, but for its own survival.