More Kurzweil December 17, 2007January 6, 2017 JUSTICE IS BLIND (BUT REPUBLICAN) From the McClatchy newspaper chain: WASHINGTON – The Justice Department’s voting rights chief stepped down Friday amid allegations that he’d used the position to aid a Republican strategy to suppress African-American votes. . . . SEX AMONG OCTUPI Nick Altenbernd: ‘In between quadrupling and 16-tupling (sesidecitupling, I think) is octupling, not sextupling [as per Friday’s column]. You’re obsessing about that orgasm button again.’ ☞ Agh! You are so right! KURZWEIL FOR FREE AND IN COLOR Bruce Stephenson: ‘The Kurzweil C-SPAN interview video is actually available on-line for free (though it requires RealPlayer), here.’ ☞ The full-length video. Although there was something pretty wonderful about just listening to it, freeing your imagination to soar into . . . TECHNORAPTURE Ken Doran: ‘Reading Friday’s column, the word that occurred to me was technorapture. I was hoping that I had invented it, but Brother Google informs me otherwise, at least as to first invention. Those who toss it around seem to have something with more religious overtones in mind, but that doesn’t mean we can’t appropriate it with our own spin.’ AN OPPOSING VIEW Peter Kaczowka: ‘I am an admirer of Ray Kurzweil, and I realize he’s smart. In fact, I worked five years for a ‘Peter S,’ who had previously worked closely with Kurzweil on his reading machine for the blind. Peter S told me I was the best software engineer he had ever known, which presumably included himself and Kurzweil. I may not be as smart as Kurzweil, and I only have a lowly Dartmouth Math BA (1972) compared to a MIT PhD which both Kurzweil and Peter S have. But I am considered a better engineer, so I feel qualified to judge Ray’s statements. ‘I spent five years writing medical imaging software. I attended the Radiology Society of North America (RSNA) convention every year, spoke to radiologists and engineers. My software is used in medical image viewing systems from GE, Fuji, all the major medical equipment vendors. More than half of all medical image viewing stations now sold worldwide run my software. I thus feel qualified to say that modern medical technology is overrated, particularly diagnostic imaging. Radiologists and technologists tend to agree, derisively calling them ‘scams’ instead of ‘scans’. It is Sales and Marketing people that generate the false hype, also true for pharmaceuticals. ‘I am prouder of my DVR software work from 2000-2004 at a company bought by Motorola. If you now buy a DirecTV HD DVR it is running my software. (The DVR itself runs Linux. Imagine that, your DVR is more sophisticated than your Windows PC!) I authored two US patents (Google ‘Peter Kaczowka’), which relate to ‘whole home DVR,’ the ability to watch your single DVR from any TV set in the house, connected by the cable. According to this, that feature will be available from DirecTV in the future. ‘So I am (or was) a technologist. I worked with technologists for 30 years, and I got tired of their self-aggrandizing hyperbole. Psychologists call it ‘cognitive dissonance’: the tendency to see one’s own actions in the best light. I call it hubris. ‘Predictions of the wonderful things that technology will do have been consistently overly optimistic. That is particularly for Artificial Intelligence (AI), Kurzweil’s own field. Here are some wrong predictions by experts: – in the late 60s Marvin Minsky of MIT predicted that within 20 years we would have computers that out-performed humans. Minsky was an adviser on Kubrick’s movie ‘2001: A Space Odyssey’. It is now after 2001, and are we on Jupiter’s moons? Can you talk to your computer? Can your computer read lips? When 2001 came out in 1969, suppose you had predicted that in 2001 ‘people will mainly communicate with computers by rolling a ball on a table, thus moving a graphic cursor on the screen, and clicking a button after said cursor is over the chosen task’? – my ex-wife’s uncle led all speech recognition work at the NSA, from 1950 until the 90s, when he told me ‘in 1950 we were sure that within 20 years, we would have full vocabulary, speaker-independent voice recognition. We now think that may never happen.’ (This means the NSA may be listening to our phone calls but their computers can’t decipher them. I find that reassuring.) – regarding the promise of nuclear energy, Lewis L. Strauss, chairman of the Atomic Energy Commission said ‘It is not too much to expect that our children will enjoy in their homes electrical energy too cheap to meter’ [New York Times, August 7, 1955] ‘Sadly, computers are stupid. They require extraordinary effort, from the smartest programmers, in order to perform the simplest tasks. Given their crude functionality, that computers can be used at all is a tribute to the genius of their human users, not their programmers. Simply put, all software sucks, because writing it is extremely hard, and beyond the capability of most humans. We may be near the limits of what computers can do, because of the limitations of their human programmers. Ray Kurzweil probably did not do much programming himself, spending his time managing, theorizing, writing white papers instead, or he might realize that. Maybe he believes other, better programmers (like me) can do it, but we can’t. [But can’t computers themselves take over much of the job of writing software? – A.T.] ‘PS – I am not the only one dubious about Kurzweil’s predictions. Google ‘Kurzweil incorrect predictions‘ and you will see many dissenters, including this one. ‘PPS – Regarding mapping the human genome, which you use as an example of modern technological wonders: note that all the mapping did was read the base-pairs; it did not decipher what they do. Since then scientists have learned that a given gene sequence does not even guarantee what protein will be produced. Also, the ‘garbage’ genes that were thought to not encode information, in fact do. Worse, apparently long sequences of ‘garbage’ genes combine with ‘regular’ genes to determine what proteins get produced. Because of that, we may never be able to decipher the genetic code. It may be unbreakable, like a 30,000 bit encryption scheme. ‘In other words, sequencing the human genome was also a grossly overrated achievement. But the hype accomplished its purpose: billions of dollars pour into genetic research, and the technologists get paid to play :-). ‘Sorry for this long mail, but it was cathartic.’ ☞ And interesting! But if I have my choice between believing in Biblical miracles and technological miracles, I’ll go with the latter, if only because I’ve seen so many of them. ‘We may be near the limits of what computers can do?’ Not bloody likely.