Many people (including me) have wondered why large tech companies don't do this, and many people (including me) have quit and founded startups instead. I think it must be intrinsically hard to do. Look at it this way: there already is a mechanism for spawning startups that distinguishes winners and lets founders reap rewards. It's called the open market. So essentially, for an organization to succeed at this would mean beating the market. That's hard.

I'm unconvinced by the analogy between hedge fund managers and startups. The variation in what the former are doing is miniscule compared to the variation in the latter. That doesn't make it easy to be a winning trader, but it does make it far easier to identify winning traders and build an organization around them.