I’ve been reading Jon Gertner’s The Idea Factory lately; it documents the history of Bell Labs, an R&D facility shared by AT&T and Western Electric that was directly or indirectly responsible for a number of innovations in telecom. Bell Labs researchers created transistors, signal extenders, and Unix, among other things. They developed the foundations of the Internet, statistical process control, and solar power. It’s shocking that one firm’s R&D efforts could affect the future of the Western World so profoundly, because there are few equivalents in modern life.
A few quotes from Gertner’s book help explain why Bell Labs was so singular:
“The techniques forged at Bell Labs—that knack for apprehending a vexing problem, gathering ideas that might lead to a solution, and then pushing toward the development of a product that could be deployed on a massive scale—are still worth considering today, where we confront a host of challenges that seem very nearly intractable.”
“Of [Bell Labs’] output, inventions are a valuable part, but invention is not to be scheduled or coerced. The point of this kind of experimentation was to provide a free environment for ‘the operation of genius.'”
“Davisson … had an almost unwavering commitment to what Kelly would later define as basic research—that is, research that generally had no immediate application to a product or company effort but sought fundamental knowledge regarding the deeper nature of things.”
From these snippets, we can see some of the traits that set Bell Labs apart. First, the firm cared deeply about finding and solving the right problems. They held a “notion that there were plenty of good ideas out there, almost too many … they were looking for good problems.” We live in a world where people have lots of incremental ideas (i.e. those that make a tiny improvement on existing products). Just glance through this “Uber for X” collection on Product Hunt to see how many derivative ideas are out there. I’m not saying that derivative ideas are bad, but am noting that they create a sort of convergence around some fundamental insight, product feature, or business model. Instead, intellectual exploration should focus on divergence and solving lots of different problems. By solving problems that are not correlated, researchers can create a much larger net impact.
Second, Bell Labs had a greater motive than the mere profit motive of its parent firms. Its goal was to use the genius of its researchers to explore the nature of reality, whether they were bound to find answers or not. Instead, Bell could defray risk across its massive project portfolio. By deploying so many resources across so many different initiatives, the Bell Labs management could be relatively assured that at least a few projects would hit the ball out of the park (similar to the VC model).
Third, unlike most firms’ focus on applied research, which was conducted with a specific goal or product in mind, Bell also devoted resources to basic research in order to explore the nature of reality. This type of research was even less certain of a profitable outcome, but in the cases in which it worked out, basic research could create staggering breakthroughs in a variety of disciplines.
This type and scale of exploration is practically lost in the modern world. The closest comparison might come from X, the R&D arm of Alphabet. However, X only has a few hundred employees, compared to the thousands that Bell Labs employed. And X mainly does applied, rather than basic, research.
In my view, the Bell Labs model would actually be most effective for a new type of university, in which students would be trained in a variety of disciplines through hands-on projects for the first two years.
Then, they might choose one or two relatively technical fields to specialize in and conduct mandatory theses in those fields, for the third year. The catch, however, is that those theses must aim to find novel knowledge, rather than focusing on expanding a granular detail of some old research. Consider Claude Shannon, a Bell Labs researcher, whose MIT master’s thesis explained how electrical switches could perform binary logic; his work laid the foundations for the computer and eventually, information theory and artificial intelligence. We need more people to take intellectual leaps of faith.
Finally, in their fourth year, students will be placed on teams that align with their interests in order to conduct both basic and applied research, and ultimately launch a product. This manner of education, centered around intellectual exploration, will likely help drive real, rather than incremental, innovation.