This notion that people can't pick winners is, in my opinion quite inaccurate. The need for light weight electronics for aerospace systems, and the view that semiconductors were 'the answer' to this problem led to a massive amount of government funding (50% of all R&D funding) and support (the government was far and away the largest customer for semiconductor devices) almost directly led to the creation of the Silicon Valley we are all so proud of [1].

Second, the notion that prizes can solve this problem is simply unreasonable. If a major goal (ie a cure for cancer) takes 20 years to develop, no organization can pursue the goal without intermediate funding. In order to provide intermediate funding with some semblance of accountability, you essentially create the existing R&D structure that exists in America today.

The other advantage to government funding is that with proper open access protocols, innovative ideas aren't trapped in the organizations that created them. There's no doubt that companies like SpaceX could have achieved what they have without the backbone of research done by NASA and published in NASA's amazing technical report series.

1. A nice article on the history of government support for the transistor industry. http://coursesa.matrix.msu.edu/~business/bhcweb/publications...