How are new technological innovations changing the economy and society? We continue reading of Brynjolfsson and McAfee’s recent book, The Second Machine Age (previous posts):
From this week’s SMA-related news: I’ve been playing with the Expresso web application. It’s a simple tool. Simply paste in a chunk of text, and Expresso analyzes it for a bunch of factors: passive voice, complex nouns, rare words, extra long and short sentences, frequent words, and more.
It automates the function of an editor or writing instructor with incredible speed and significant scale, apparently.
And it was built by this grad student. Another sign of innovative digital technology hitting the worlds of work and consumption.
More ambitiously, the Los Angeles Times now publishes earthquake stories written by a program called Quakebot.
On to the book!
Chapter 9: The Spread
Once upon a time economic productivity and median income advanced together. Then the two decoupled in the 1970s, and their separation has widened since. Consider, for example, this graph: 
That widening split is a good image for this chapter’s focus on growing inequality. Here Brynjolfsson and McAfee contrast the many positive results of technological progress, which they’ve dubbed bounty, with socioeconomic division, which they call spread. “[T]here are large and growing differences among people in income, wealth, and other important circumstances in life.” (127) Spread and bounty share a cause: “[t]he main driver is exponential, digital, and combinatorial change in the technology that undergirds our economic system.” (133)
How does this spread play out? Brynjolfsson and McAfee identify several forms, starting with unemployment. Employers can substitute technology (and capital) for workers – a very old story, but one accelerating in the present and into the near future. Technology also empowers certain professions and people with advanced degrees to become more effective and well-compensated. The authors cite Autor, Acemoglu, and others, who dub this “skill-biased technical change” (136).
Add business organization to the mix and the result is a demand for “more educated workers and reduce[d] demand for less-skilled” ones. (137) “[S]ome workers (usually the less skilled ones) are… eliminated from the production process and others are augmented” (138). In between the two, middle class jobs “collapse in demand” (139).
That spread is uneven across the population. If we look at median income and the proportion of people employed, we see “fewer people are working, and wages for those who are working are lower than before.” (145) At the top capital’s owners are doing just fine, although chapter 9 concludes by considering the possibility that competition will whittle down their bounty.
Chapter 10: The Biggest Winners: Stars and Superstars
Here the authors zoom in on the dynamics by which some have a little, while most have a lot. To begin with, they add another spread-driving category, building on skill-biased technical change to discern talent-based technical change (148). This is the story of the superstar economy, wherein the leading contributors to fields capture a larger than ever share of their market’s rewards. Now the story goes beyond superstars:
[T]he evidence suggests that the spread of incomes continues at high levels of income with a fractal-like quality, with each subset of superstars watching an even smaller group of super-duper-stars pulling away. (149)
“Much of this growth is linked to the greater use of information technology,” as you might expect. (151; 152-7) Brynjolfsson and McAfee also make room for non-technological drivers, including social acceptance of higher compensation, changes in tax policy, and trade developments (157-159), not to mention the effects of power law curves (159-162).
That’s a lot to chew on from these two chapters. Second Machine Age now integrates many of its previous themes, and pushes them into society: aspects of innovation; dynamics of technological change. The next chapter, 11, will continue along these lines. Then chapter 12 will pivot towards recommendations for readers.
Here’s an SMA-themed cartoon to end on:
(with help from Naked Capitalism)
Related to automating the writing of stories, I’ve seen demos of Narrative Science’s technology, which has been used to write sports stories as well as investment analyses. Hoping the final chapter doesn’t invoke magical reskilling in its list of recommendations.
Tangent: The cynic in me notes the uniform skin pallor of the human employees of the month in the cartoon and wonders whether or not it was intentional.
Steve, what did you make of those Narrative Science demos?
For now, it relies on structured data to make inferences from, and templates for writing styles and what to emphasize or call out. An Atlantic piece notes the possibility of applying it to standardized testing: “Instead of simply tallying wrong answers, your kid’s standardized test results make highly specific study suggestions—in language that would do an English teacher proud.”