This post also appears in my O’Reilly blog.
In my series on downloads I looked at a couple of scenarios, based on traditional CD sales and iTunes sales, that indicated that the people who make music may end up making less money from downloads than they used to from CDs, even though CDs were not a great business to start with. Digital efficiencies have the effect of driving down prices, and the end of the trend may be music that’s worth next to nothing on a per unit basis. One possible outcome: Aggregators of that nearly free content may very well be able to make money, but small labels, and artists, may not.
My scenarios represented very small data samples. But they may be examples of a much bigger trend: a productivity crisis, one caused not by too little productivity, but by too much.
In the past, increases in productivity tended to be locally disruptive but broadly beneficial, a trade-off described by economist Joseph Schumpeter as Creative Destruction. When more efficient work produces more output per worker, some workers lose their jobs. But the overall increase in wealth creates more prosperity for everyone (assuming a roughly free and fair market), and more jobs. Those new jobs often required higher skills, but if displaced workers got training in those skills, they could end up earning higher wages. These days, telecommunications, computerized logistics and efficient transporation allow jobs to be shipped to lower-wage countries. Further down the line, automation eventually moves those jobs from low-wage humans to robots. On balance, though, if workers are being retrained for higher wage jobs, everyone may benefit.
But there may be a new problem, arising from the fact that with technology, each round of improved productivity causes the next round to come sooner. Because most of human history has been a story of scarcity, it’s our habit to assume that more and cheaper is always better. But as we see in many systems, a phenomenon that is benign when it happens at one rate can become threatening at a faster rate.
This kind of worry about the anti-human threat of technology is not new–when trains first started running, some feared that humans might be killed simply by moving at that speed. And so far, human adaptability has always triumphed, with life generally getting better for most people. For all the stresses of modern life, the average resident of the developed world is the healthiest, longest-lived, most prosperous person in the history of humanity.
But it seems to me there’s a natural limit to how fast people can learn how to do new jobs. If productivity improvements keep eliminating more jobs faster, the curve of productivity improvement will cross the curve of educational improvement, and that intersection could very well be a crisis point. It may be one we’ve already reached.
Much social tension in the US over recent decades has been a result of the divergence between the minimum educational level required by the average job and the level actually attained by a particular group of workers. In the 60’s, many African Americans, equipped with grade school educations from second rate schools, were left behind as factories moved to cheaper labor markets or became automated. African American men’s lost access to relatively well paid manufacturing jobs contributed to the decline of the cities, the break-down of families and a rise in crime. Later, whites with high school diplomas started to find that they couldn’t get good jobs, and we saw the rise of white supremacists and militias. Now many college graduates are struggling to find good jobs.
Meanwhile, the US economy has grown smartly, fueled by these technology-driven productivity gains. But as the overall economy has grown, we’ve seen a widening gap between the rich and everyone else: It looks like the value of an elite education (or really good social connections) has skyrocketed, while people below that level are being commoditized. Unemployment is low, but the unemployment percentage doesn’t reflect the many people who have given up looking for a job, the many people who are employed in low wage jobs.
Where does this lead? As more and more work can be done by machines, what do we do with the idled humans who can’t adapt fast enough? Pay them for not working, ending up with welfare for just about everyone? Or will we discover dramatically improved ways to educate people? Even if we do make that discovery, there will still be IQ limits to contend with–do people deserve to be poor just because they’re not brilliant? Or will people evolve in the direction of the very technology that poses the challenge? Ray Kurzweil argues just that in his new book The Singularity Is Near: When Humans Transcend Biology: He believes that in the not too distant future we will merge with brilliantly intelligent computers and become, in effect, a new species.