Kurzweil comes closest to showing me in error by pointing out that the published article used the word “foolproof” inappropriately, but as a note from the magazine’s editors acknowledges, that was not my doing:
Editor’s comment, 30 December 2010: Mr. Kurzweil’s objection to John Rennie’s critique begins with, and makes much of, the use of the word “foolproof.” In fact, that word was never used by Mr. Rennie. In an editing error, the word “foolproof” was inserted into a sentence in Mr. Rennie’s article, after Mr. Rennie had reviewed the editing. IEEE Spectrum regrets this lapse.
My response to Kurzweil’s rebuttal is as follows :
Ray Kurzweil was kind to take the time to reply to my article, though his responses, I think, show the gulf between his perceptions of what he is saying and how the rest of the world understands him. No doubt he sees that as our failure, not his.
According to his newly released “How My Predictions are Faring” document, since Kurzweil wrote to Michael Anissimov last January, his count of the number of predictions he made has increased from 108 to 147, and his own scoring of the predictions as “correct” or “essentially correct” has dropped from 94.5 percent to 86 percent. Perhaps he and I are seeing more eye to eye all the time.
IEEE Spectrum’s editors have been kind enough to acknowledge that the unfortunate word “foolproof” found its way into my article only after I had signed off on it. Whatever other inaccuracies, misquotes, misunderstandings and made-up factoids Kurzweil thinks he sees in my article, he has not specified them. Meanwhile, I stand by my statements and invite readers to follow the links to video of his TED talk and to Anissimov’s Accelerating Futures web site—just some of the many sources I consulted—to judge for themselves.
Putting aside the word “foolproof,” I assume Kurzweil is comfortable with my description of automated speech translation as widespread and real-time? Here are Kurzweil’s relevant predictions for 2009 in The Age of Spiritual Machines: “Listening machines can also translate what is being said into another language in real time, so they are commonly used by hearing people as well” and “Translating Telephone technology (where you speak in English and your friend hears you in Japanese, and vice versa) is commonly used for many language pairs.” The technology exists, but do readers feel that in their experience, this description is sufficiently accurate to count as “essentially correct” in Kurzweil’s scoring?
Kurzweil is simply not a reliable judge of his own prophecies. For example, in his “How My Predictions Are Faring” document (which he posted to the KurzweilAI.net site after my article appeared), he hails as “correct” his prediction about “The Computer Itself” that three-dimensional chips will be common by 2009, and he cites as proof the popularity of vertical stacking architectures in ubiquitous MEMS and CMOS chips. But those are image sensors, not microprocessors, which seems like a rather generous interpretation of what he could have meant by “computer.” He also boasts that chips with more sophisticated 3-D designs, better suited for microprocessors and memory, are expected to be 6 percent of the semiconductor market… by 2015. This is his idea of a validation?
Or this: “There is a growing perception that the primary disabilities of blindness, deafness, and physical impairment do not impart handicaps. Disabled persons routinely describe their disabilities as mere inconveniences.” Kurzweil deems that prediction “correct,” too. Would disabled audiences who heard that statement in 1999 have imagined the current world as the fulfillment of that promise?
Kurzweil waves away my review by saying that I cherrypick predictions rather than evaluating all of them. But as I noted, the problem with weighing Kurzweil’s record is that so many of his predictions are so open to interpretation that any scoring is debatable. Does the existence of a phone app prove a technology is in wide use? Do iPhones count as jewelry? Do they count as clothing? If a service becomes available in late December 2010, is that a hit or a miss? How about if it comes out two years from now? How about five years from now—on a 10-year prediction?
The specific predictions in question aren’t obscure ones. Kurzweil’s prediction about the rise of networked computing in everyday life is one of the standard proofs of his prophetic abilities mentioned in his introductions and press materials. (It’s the first one he mentions in “How My Predictions Are Faring.”) My point was that his claim looks less impressive if one looks at all the others who were implicitly making the same extrapolation yet who never get the same credit. Kurzweil’s reliance on his law of accelerating returns instead of “intuition” in this regard is a distinction without a difference.
Similarly, I continue to be baffled by Kurzweil’s statement about disappearing computers. He is wrong: I am more than happy to recognize the microprocessors embedded in objects as computers, and I said as much in the article. So why did he say in 2005 that by 2010, computers would disappear? Microprocessors were already commonly embedded in other devices. Kurzweil’s statement meant only that computers of all types would continue to get more powerful rapidly—which I called insipid. I stand by that description, too.