The assumption is (I believe) that innovation is better when it's not goal-oriented. Currently much of the funding goes to projects that have a clear path to results, but finding that path is a prerequisite, and is not funded.