I read this book after seeing a review in the Lexington column of The Economist magazine. The Economist review is good, but I do have a few quibbles with the book. First, let me say that the general theme, “average is over,” is I believe, right. I agree with Cowan that those who are able to work effectively with electronic devices and tools will be those who are able to hold down the higher paying jobs of the 21st century. Effective education is a part of the equation, as are motivation, perseverance, and native intelligence. Those who lack one or more of these qualities will have a hard time being in the “top 15%” of society, as Cowan postulates.
Now, on to my issues with the book. Cowan’s focus on chess, while understandably important to him, is but one case study, and I think that he uses empirical observations about chess and chess computer programs to project a narrow domain on both human endeavor and computational theory. Not only that, but he lets this example run on far too long. More important, though, is Cowan’s focus on “genius machines” and their ability to develop theoretical frameworks beyond the ken of humans. In chess or other computational problems, there are data and algorithms. Lots of data and sophisticated algorithms can produced results that are beyond human abilities due to the immense amount of computation and number crunching involved. By mining data, computers can find correlations that we didn’t know existed. But, can they determine if a relationship is causality or coincidence? To the extent that the answer is “computable” a computer is a wonderful tool.
I’m much less sanguine about a set of silicon having a spark (perhaps at the Kurzweil singularity) that allows it to develop theoretical frameworks (without significant fundamental changes in the science of computers). More data and faster execution won’t make the qualitative changes necessary for those results. Cowan is right that we are getting further removed from understanding the technologies behind our tablet computers and smartphones. However, as “magical” as the iPad may be, as Arthur Clarke said, “Any sufficiently advanced technology is indistinguishable from magic.” At its heart, the iPad, your desktop computer, or a supercomputer is still a von Neumann computer with one instruction from a program being executed after another. The processor (running its own sophisticated microcode) may support parallelism, predictive execution of code, modification of the code it’s executing, and a myriad of other performance enhancing features. However, it still at heart is executing one instruction after another and processing data with those instructions…albeit very quickly with a precision that a human can never match. Cowan’s economic analysis may be good, and his theory on the impact of technology on the workplace generally accurate, but he’s not a computer scientist.
All in all, it’s a good book. While my left-leaning democratic tendencies make me believe that we do need to provide a higher level of social services for those who are digital roadkill, Cowan’s thesis that the economic impact of technology is just beginning to be felt is something that I agree with wholeheartedly. Cowan highlights problems that our politicians need to address and that the electorate needs to understand.