'How to Create a Mind': The matter of Ray Kurzweil's mind
The single-minded transhumanist offers his best insights -- and worst tendencies
December 23, 2012 5:00 AM
Ray Kurzweil is no slouch, but he "habitually oversteps his sphere of competence."
By Josh Raulerson
Ray Kurzweil only has one idea, but it's a pretty good one. Good enough, anyway, to fill out a half dozen best-selling books and pack lecture halls virtually everywhere he goes. It is essentially this: A simple concept, iterated to an extravagant degree, will produce astonishing outcomes.
In the books for which Mr. Kurzweil is best known -- "The Age of Intelligent Machines" (1990), "The Age of Spiritual Machines" (1998) and "The Singularity Is Near" (2005) -- the operative concept is Moore's Law, the Silicon Valley axiom that machine intelligence tends to accrue exponentially. The astonishing outcome, inherent in the logic of exponential growth, is singularity: the moment when a seemingly linear pattern of technological development suddenly goes off the rails. One day very soon, Mr. Kurzweil predicts, we will wake up in a world transformed from top to bottom by presently unthinkable technological marvels: sentient computers, probably, or human minds bootstrapped to godlike superintelligence.
An avowed transhumanist, Mr. Kurzweil is particularly keen on the latter scenario, and his latest book attempts to lend it some needed theoretical coherence.
"HOW TO CREATE A MIND: THE SECRET OF HUMAN THOUGHT REVEALED"
By Ray Kurzweil. Viking/Penguin ($27.95).
Its aim is ambitious, to say the least: a neurophysiological model of consciousness upon which to build artificial intelligences ontologically equivalent -- identical, as far as he's concerned -- to human minds. By extension, such a theory would also support the proposition that an extant human mind might be transcoded to, or merged with, software. The quest for this knowledge, Mr. Kurzweil writes with characteristic bombast, "may be regarded as the most important project in the universe."
Ray Kurzweil makes a magisterial case, bolstered by reams of impressive diagrams and off-the-cuff citations, that what we experience as consciousness is an epiphenomenon: a gestalt produced at the intersection of uncountable, discrete and hierarchically nested neurochemical events. Reduced to their most basic mechanics, it turns out, these structures of cognition are really quite simple -- enough to replicate, and therefore augment, at the level of code. We are already computers, in other words, so making the leap to posthuman transcendence should be a fairly straightforward matter of upgrading our hardware.
For all its technical granularity, the argument is quite accessible. Persistent readers will follow it easily enough, and many will find it persuasive.
• • •
Those already acquainted with Mr. Kurzweil's style of futurism may be less impressed. As elsewhere, the grandiosity of his overarching claims and the considerable authority he brings to bear on them belie a rhetorical strategy, and a corresponding worldview, that are borderline-autistic in their simplicity.
Briefly: Begin with a complex set of facts -- historical, economic, phenomenological, whatever -- and boil it down to a simple observation. Posit that observation as not merely a descriptive comment, but an immutable "law." Then, extrapolate it to its logical extreme. This is pretty much the only trick he knows.
It might make for a less problematic approach if Mr. Kurzweil had better instincts about how far to take it. Indeed, as a layperson's primer on neurocognition, "How To Create a Mind" is informative, refreshingly transparent and often engrossing. But Mr. Kurzweil has mistaken the concept of pattern recognition for a comprehensive theory of mind, encompassing everything from the nuances of aesthetic judgment to the sublime mysteries of love.
The result is philosophically naive, methodologically dubious, at times intellectually dishonest -- and, above all, monstrously reductive.
In seeking to "reverse-engineer" the mind -- demystifying its outward complexity and reducing subjectivity to a schematic -- Mr. Kurzweil habitually oversteps his sphere of competence, casually conflating unrelated concepts and leaping to bizarre conclusions.
His notion of mastery, for example, is a one-size-fits-all mathematical function: Garry Kasparov is the world's greatest chess player because he has learned 100,000 different board positions. In exactly the same way, William Shakespeare is history's greatest poet because his works make use of 100,000 different word senses. (Mr. Kasparov has already been beaten at chess by a computer, so presumably one will soon "beat" Shakespeare at poetry.) It soon becomes clear that by "mind" Mr. Kurzweil really means "brain" -- which might as well be floating in a jar, for all that the rest of the body figures in his account -- and by "consciousness" he means perception, memory, and abstract reasoning.
Hence, all you really need to understand about culture: music, art and literature are higher-order patternmaking functions, secondarily imbued with emotional content and lyrical qualities for mostly evolutionary (i.e., erotic) reasons. Love itself is dispensed with in three rambling pages on hormones, neurotransmitters and cliches mined from pop songs.
Odder still is Mr. Kurzweil's exasperated, Spock-like bewilderment at the vagaries of social interaction: our reliance on language corrupts and complicates what should be a simple, unmediated exchange of data, if only we could tap directly into our interlocutors' brains and "master" the patterns therein. "Of course," Mr. Kurzweil admits, "we don't yet have access to someone else's neocortex; we need instead to rely on her attempts to express her thoughts into language (as well as other means such as gestures) ... it is no wonder that we misunderstand each other as much as we do." Sorry, ladies -- he's taken.
• • •
Readers who have followed his career know there are really two Ray Kurzweils, and their response to this book will depend largely on which Kurzweil they perceive behind it.
One is a MIT-trained computer scientist, artificial intelligence pioneer and creator of the modern musical synthesizer. He's been honored with the National Medal of Technology and rightly commands the respect and admiration of his peers. Then there's the other Ray Kurzweil. That's the transhumanist self-help guru, co-author of books on how to forestall the aging process through relaxation techniques and dietary supplements. At "Ray and Terry's Longevity Products" (rayandterry.com), he and his partner -- a real actual doctor! -- will supply you with the nutrition bars, shakes and assorted pills and powders you'll need to stay alive until there's a cure for death.
The trouble with "How to Create a Mind" is that contains both Kurzweils -- the noted man of science and the enterprising crank. The latter is more than willing to cloak himself in the former's authority in service of claims neither is really qualified to assert. He has a condescending habit of noting the mind's intrinsic limitations -- we resist novelty, miss the big picture, rely on irrational coping mechanisms, etc. -- qualities that may make it difficult for some readers to buy what Mr. Kurzweil is selling. In deploying this defensive tactic Mr. Kurzweil implicitly compares himself, more than once, with Darwin and Einstein: iconoclastic thinkers whose ideas were at first ridiculed and rejected, but ultimately vindicated.
While it's possible Mr. Kurzweil will be similarly vindicated one day, we're not there yet. Until then, a better historical analogue might be found in someone like Nikola Tesla: a prolific inventor and legitimate genius, with quasi-mystical pretensions and a propensity for the theatrical. As with anyone of his stature, Ray Kurzweil is not easily dismissed. But he shouldn't be taken too seriously, either.
Josh Raulerson is a reporter and local host of NPR's "Morning Edition" on 90.5 WESA-FM (firstname.lastname@example.org). He also produces "Speaking Volumes," a weekly conversation with Pittsburghers on books and reading. His Ph.D. thesis in literature, "Singularities: Technoculture, Transhumanism and Science Fiction in the 21st Century," will be published next year by Liverpool University Press.