I'm a "pasty scientist" paid by DARPA, in biotech. I work for someone who previously founded a brain-machine-interface company that had a small IPO (cyberkinetics). I would love nothing more than to see things I work on get turned into viable products. This seems to require two things: a) public investment in vague-but-cool-sounding-what-if projects, and b) easy ways to translate them into usable products.
For (a), someone should be funding things that might be valuable "platforms" for society: the human genome project, ARPANet, basic mathematics research (Fourier analysis, instrumental in many, many commercial products, was one such basic mathematical result for a long time). The value of these platforms may take a long time to pan out (if ever), so it makes sense to fund them as a society. Perhaps government per se isn't required - it could be done by some collection of nonprofits - but the funding won't come from many private firms. Further, if we want these platforms to be of use to many companies, perhaps they shouldn't come from private firms.
I think about it like I think about LAPACK. LAPACK is used by almost every scientific computing library for matrix multiplication. Now if Intel and AMD want to come along and make proprietary versions of it (e.g. Intel's Math Kernel Library), or if Matlab or Mathematica want to wrap them into an easy-to-use product, or someone needs it for a robot controller for a factory, that's awesome! The point is that there is some value in public research the market can draw on to make nice things (the ROI on LAPACK must be HUGE for whatever nominal cost + non-nominal man-hours it takes).
(b) is difficult, but there have been some efforts. One such is INVO[0], an effort at Northwestern that helps researchers develop their ideas without losing too much control in the process. They just founded cd2[1], the center for device development, that helps clinicians and engineers navigate the bureaucratic and regulatory hoops, and ultimately pitch. The key is that like YC, they have a strong network of people who understand this process, who to talk to, what you're really signing when you sign a piece of paper, etc. They pay you (for a year!) to work on your project.
My question (updated for clarity): What do potential founders of not-widget-startups want to be able to pursue the kind of innovation mentioned in the article? In particular, what would help translate academic proof-of-concept ideas into viable, fundable, products?
It would seem that asking VC's to fund cooler projects at the expense of profits is not the right way to go. Maybe a better strategy is to ease the link between cutting-edge research and consumer-facing products. What would help with that?
[0](http://invo.northwestern.edu/about) [1](http://cd2.northwestern.edu/)