Ashish describes his early work on the division of labor in innovation, which grew out of his advisor Nathan Rosenberg's argument that specialized capital-goods firms (think James Watt, or the textile-machinery makers who carried that technology to Japan) drove both technical progress and its spread across countries. That led to his work on markets for technology, which he now describes as an early attempt to understand startup business-model choice: when an inventor should commercialize a technology themselves, and when they should license it or sell components to others. UOP, the refining-technology firm, is his example. He credits Bresnahan and Gambardella's insight that fragmenting the value chain makes sense when users are diverse ("broad markets, not deep markets"), and he points to modular, well-defined interfaces as a condition for these handoffs to work.

The second half covers why Silicon Valley has been so hard to replicate, and where his current thinking has gone. Arora says specialization doesn't require geographic clustering (the chemical industry shows this), and that trying to recreate Silicon Valley is quixotic. Better to build on a region's own strengths. His forthcoming book, Inventing Prosperity, demonstrates that the US system of specialized universities, startups and big firms works well in digital and life sciences, but poorly in sectors like new materials and energy, where integrating research inside large corporate labs (as in roughly 1930s–1980s America) worked better. He argues that losing that integration is hurting growth, and that universities' role needs to be rethought to fill the gaps. He closes by saying he keeps working because he enjoys learning, being humbled by what he doesn't know, and being surrounded by smart colleagues and students.