Tim Lee offers two views of the economic case for innovation. I am certainly more Hayekian — I think expertise and information is inherently dispersed and the Hayekian view is more granular, if you will. It’s more likely to discover and serve niches and inefficiencies.

I think Tim hits the nail on the head that: Building an iPad didn’t just require millions of dollars, it required that those millions be spent by the right company. And the only way to figure out which company is the right one is to let a bunch of them try and fail.

I’ll go further. The Hayekian view is still not “rational” in an economic sense. Creating a startup, or a radically new product, is usually a recipe for economic failure.

What the Hayekian view allows, and why I love it, is that it makes room for “madness”. It allows economic irrationality on a granular scale.

Startups are optimized for the personal ambitions of their founders, not for economic success. The creation of Google, Apple and others were not rational or irrational, per se — they were driven by the desires of their founders. Billions of dollars later, we craft their histories into a narrative that sounds like economics.

The process of going from nothing to something, however, is inherently irrational — or more precisely, impossible to reason about statistically. Peter Thiel calls this the 0 to 1 problem. To embark on it is to leave the world of economic modeling.

The Hayekian view doesn’t require large, supposedly-rational, profit-maximizing institutions for progress. Rather, progress requires madness, that is, the freedom to pursue choices whose rationality can’t be measured.