The hypothesis for ImageNet—that computers could learn to "see" from vast visual data—was sparked by Dr. Li's reading of psychology research on how children learn. This demonstrates that radical innovation often emerges from the cross-pollination of ideas from seemingly unrelated fields.

David Epstein's book *Range* shows that breakthrough innovators often switch disciplines. By entering a new field "through the side door," they bring different mental models and "far analogies" that allow them to see solutions incumbents cannot.

The most powerful creations are a synthesis of different, seemingly unrelated concepts. Like Velcro emerging from a prickly burr and a sock, you can create a unique perspective by bridging two different areas of your experience, such as an NFL locker room and a predominantly white private school.

Claude Shannon's indecisiveness about majoring in math or engineering inadvertently became his greatest strength. This dual training gave him a unique ability to blend abstract theory with practical application, a combination essential for his breakthroughs in information theory.

Shannon's work reinforces that major breakthroughs are not always the result of a direct, goal-oriented pursuit. Instead, he believed that following one's natural curiosity on seemingly simple or useless topics often leads to the most valuable and impactful discoveries.

Breakthrough creativity, like that behind Disney's *Frozen* or behavioral economics, is often "innovation brokerage." It doesn't come from a blank slate but from combining established concepts from disparate fields—like mixing psychology with economics—to create something new and powerful.

Shannon's mentor, Vannevar Bush, held a deep conviction that 'specialization is the death of genius.' He tested this by directing Shannon, an expert in engineering, to solve problems in genetics, a field he was completely unfamiliar with, with groundbreaking results.

The idea for a living computer came not from biologists, but from engineers with backgrounds in signal processing. This highlights how breakthrough innovations often occur at the intersection of disciplines, where outsiders can reframe a problem from a fresh perspective.

The development of neural networks wasn't a linear path. It involved a cycle where computer scientists and psychologists alternately abandoned and revived the concept. When one discipline hit a wall or lost interest, researchers in the other field would pick it up, solve a key problem, and reignite progress.