most interesting for us i think is the idea of long search paths, which they describe as an innovation that looks like a very large leap from previous ideas, but is actually composed of a lot of small(er) improvements that combine over time, and are known only to the inventor