What DARPA has is a lot of money without a lot of oversight. Don't underestimate the power of that model. If you look at the organizations that have been extremely effective at innovation, it's organizations where research isn't subject to market pressures. Bell Labs was bankrolled by AT&T's telephone monopoly. Xerox PARC was bankrolled by Xerox's extremely lucrative copier business. Intel, IBM, Google, etc. Innovation is associated with companies that have such tremendous market power that they can be benevolently forward-looking, instead of having to pinch every penny to survive in the dog fight that is a competitive market place.
Tim Wu's book "The Master Switch" gives great insight into this aspect of the AT&T monopoly.
Thanks for pointing me to Tim Wu - very interesting analysis.
But does anyone have hard facts of how effective they are at applied problem solving? Or do they just get lucky every once in a while and have good PR? In particular have they developed some better way to run R&D programs?
I don't know but my limited interaction with their program managers did not impress me but this is purely anecdotal.