How is digital technology changing how we innovate? We continue reading Brynjolfsson and McAfee’s The Second Machine Age (previous posts).
Chapter 5 shifts the book’s focus so far from technology as such to the problem of innovation. “Innovation: Declining or Recombining?” begins by tying economic productivity to innovation, then reprising arguments in favor of a current slowdown in new ideas. Robert Gordon and Tyler Cowen are touchstones for the latter.
SMA sees a way out of innovation doldrums by way of the general purpose technology (GPT) concept (David and Wright, 1999, among others). GPTs are, as you might expect, technologies that “have the potential for important impacts on many sectors of the economy” (76). Think of steam power or electricity. So instead of viewing information technology as narrowly applicable to economies and societies, Brynjolfsson and McAfee argue that “ICT is a GPT”*. The digital world has impact far beyond generating new ways of “entertaining ourselves inexpensively online” (77).
How can this work? This is where the book’s promised theme of combinatorial or recombinant innovation comes in. New ideas are not necessarily brand new concepts or inventions, but rearrangements of preexisting resources and ideas. SMA offers a series of examples of this, including the inventions of polymerase chain reaction (PCR), Waze (again), the World Wide Web, Facebook, Instagram, and more (78-81, 83-88).
Brynjolfsson and McAfee are bullish in combinatorial innovation, arguing that the sheer number of possible combinations is immense (hearkening back to their notes on exponential growth curves). Brian Arthur and Paul Romer are cited to back this up, claiming combinatorial innovation is a wide-open field, powered by ideas.
Chapter 6: “Artificial and Human Intelligence in the Second Machine Age”
This is a very short and, to my reading, light chapter. SMA returns to artificial intelligence, claiming that continued development of AI will add still more power to innovation. Examples embodying this include Orcam, Watson’s deployment in medical situations, and the C-Path semi-automated cancer detection effort.
AI isn’t alone here, as Brynjolfsson and McAfee add to the chapter the creative and combinatorial possibilities inherent in increasing humanity’s ability to connect with itself. Social media and mobile devices play the star role here, and the chapter is flagged by a de Chardin quote.
However, that’s about all for chapter 6. We don’t get to explore the limitations of AI, both practical and theoretical. The crowdsourcing bit is really a formality, almost a placeholder.
Thus ends the 1st part of The Second Machine Age. So far the text has addressed technological issues. With the next chapters our focus will shift to to Brynjolfsson and McAfee’s concepts of “bounty and spread”, the economic and social effects of these emerging technological forces.
*GPT recalls for me Cory Doctorow’s forecast of a war on general purpose computing.