Transcendent Man

Kurweil and many others have been chattering about their "Singularity" since at least the mid- 1990s. This is not the astronomical phenomenon, but is similarly dense. Basically, the notion that machines, specifically computers, will someday soon exceed the intelligence, cognitive, perceptual, analytical, and other mental powers of humans, and become "self-aware" and achieve consciousness. As seen in the Terminator movies, The Matrix, this has become one of the basic, stock tropes of science fiction, though it has been around in fundamentally the same form as present since the early 1980's, at least.

Interesting stuff, and not only for entertainment purposes. And clearly machines will (and have already) become more able than humans at a wide range of tasks. From Big Blue beating Gary Kasparov at chess more than a dozen years ago, to welding robots in auto plants, machines do many things faster, and ultimately better than man. Persons under 20 have much less "data" in their heads, having come to rely so heavily on Wikipedia and Google (having been taught to do so by parents and teachers, in fact), and the online fact is up-to- the-minute and dead-accurate, isn't it? More reliable than what lies in your mind? No doubt computers will continue to increase in power, and in the influence they have on our lives. And we will come to rely even more on them than we do now.

However, Kurweil, like almost all the other exponents of the Singularity, including those offering a much darker version of the future than Kurweil's, fails to understand and take account of a number of critical points. First, the notion of consciousness/awareness, and even of intelligence itself is poorly defined. Neurologists, brain scientists, psychologists, who study the human mind as a profession disagree fundamentally as to what these qualities even are, how they work, or why one individual human has them in abundance, another hardly at all. The human mind and brain may truly be the last frontier of science. We know less, understand less, about the brain and mind than we did of infectious disease in the early 19th century, and psychologists and brain scientists would agree. We are only making a bare beginning at understanding the brain and nervous system, and how it works, and how we actually use it.

For Kurweil, or anyone else, to predict machines of any sort will attain human levels of consciousness, intelligence, or thought by 2040, or even by 2080 (two frequently cited dates) is a patent absurdity which takes no account of the state of neuro-science and psychology. Who will design or program this machine that emulates, then exceeds, the human brain/mind? In order to make a copy, you have to understand the original in every nuance. Not the kludgy, narrow silliness of "Eliza" or "Racter", but something that can reliably pass a Turing test and also learn. And as for self-awareness, that would be a trick, wouldn't it? You'd need some sort of reverse Turing test to apply to the machine. That's Kurweil's problem, and that of his colleagues. They are, none of them, professionals in medicine, psychology, brain science, or neurology, nor do any of them (that I am aware of) have ANY training in these disciplines. And until brain science advances a GREAT deal, I fear there is little hope of a machine that even approaches human consciousness, nor general ability and ADAPTABILITY and the ability to LEARN. For focused, targeted tasks there are super-human machines, and will be more and better every year. Futurists would do well to understand the question before giving out answers, especially extraordinary 'predictions.' There are computer scientists working with research physicians and brain scientists, each learning the others fields (a very healthy activity for progress into such a brave new world), and the most optimistic among them might predict a computer that you can actually have a real, spoken conversation with (on LIMITED subject matter) in another 20 years. As for a whole mind, anything coming anywhere NEAR to the overall human capacity for language, learning, problem solving, changing one's self to suit the environment (and the environment to suit one's self), and both analytical/logical thought as well as creative/lateral/syncretic thought, that will have to wait until we first understand what it is.

Oh, of course. I forgot. What about the machine that improves and modifies itself when the lights are turned off? It'd first have to have a motive or some sort of imperative to do that, and more importantly, some model of what it was modifying itself INTO, and would have to understand that model at its essence, which is the whole problem and main barrier to the human endeavor towards this end. It's a catch-22: you need consciousness and human intelligence to build it, and to WANT to build it. So sorry Ray, it ain't gonna happen in your lifetime (and you should look elsewhere for the talents and ideas that will eventually get us there. They aren't to be found is CSE programs or Silicon Valley).

"Transcendent Man" was a fun docu to watch. Kurweil is an articulate spokesman for his ideas, and a likable fellow. This is a thoughtful, well-made non-fiction film, and should spark a great deal of thought in those interested in the subject.