The age of Ray Kurzweil
KURZWEIL TECHNOLOGIES takes up two floors of a low office building in Wellesley Hills, near where the Charles River crosses and then recrosses Route 128. In the reception area are a vintage Thomas Edison dictation machine and a large flat-screen monitor on which a computer program draws angular, cartoon-like portraits. Across from the entrance sits an alarmingly lifelike man made of wax, bearded and brandishing a pipe as if in conversation.
Ray Kurzweil, the company’s founder, is an inventor, and has been one for as long as he can remember. ”When I was 7 or 8 my inventions actually began to work,” Kurzweil told me recently in his large, cluttered office. ”I’d build these robotic devices, like a theater that would move scenery and props and characters in and out of view by elaborate mechanical linkages.”
He was still a high school student when, in 1964, he created a computer that composed music in the style of Chopin, Mozart, and other great composers. In the early 1970s he invented the first flatbed scanner and the first practical character-recognition software, paving the way for everything from digital photography and graphic design to online newspaper archiving. Combining those two technologies with a text-to-speech synthesizer (another of his inventions), he made the Kurzweil Reading Machine. He sold the very first one to Stevie Wonder–for whom he then developed the first music synthesizer able to fool professional musicians into thinking they were listening to real instruments. In 1987 his company Kurzweil Applied Intelligence was the first to market large-vocabulary speech-recognition software.
By any measure, Kurzweil has had an exceptional career. Now, however, he has a new project: to be a god. And not just because he thinks he can live forever. Within decades, he predicts, he will be billions of times more intelligent than he is today, able to read minds, assume different forms, and reshape his physical environment at will. So will everyone. Today’s human beings, mere quintessences of dust, will be as outmoded as Homo Erectus.
All this, Kurzweil believes, will come about through something called The Singularity. Popularized more than a decade ago by the mathematician, computer scientist, and science fiction novelist Vernor Vinge, who borrowed the term from mathematics and astrophysics, it refers to the future point at which technological change, propelled by the explosive growth of artificial intelligence, will accelerate past the point of current human comprehension. In Vinge’s prevision, once artificial intelligence surpasses human intelligence there will be no turning back, as ever more intelligent computers create ever more superintelligent offspring.
Among the programmers, scientists, and philosophers concerned with the larger contours of technological evolution, the term quickly caught on. The Singularity became an axis around which debates on technology, human nature, genetic enhancement, and the future of consciousness all turned. Figures like Marvin Minsky and Hans Moravec, the artificial intelligence pioneers, and K. Eric Drexler, the father of nanotechnology, took it up.
Today Ray Kurzweil is the most radical and most visible prophet of The Singularity. In talks, public debates, articles, postings on his website, and in a series of increasingly provocative books–”The Age of Intelligent Machines” (1990), ”The Age of Spiritual Machines: When Computers Exceed Human Intelligence” (1999), ”Fantastic Voyage: Live Long Enough to Live Forever” (2005)–he has done more than any other thinker to make the case for both the desirability and the imminence of The Singularity. According to Doug Lenat, a leading expert on artificial intelligence, ”Ray is one of the few people who can step back and see the big picture for what it means for our species and for the planet.”
This week Kurzweil has a new book out, with the self-consciously millennial title ”The Singularity Is Near: When Humans Transcend Biology” (Viking). It is the most detailed brief he has yet written for the nearness of the unimaginably strange future, and it arrives with approving blurbs from Minsky and Bill Gates (”Ray Kurzweil is the best person I know at predicting the future of artificial intelligence,” writes the Microsoft founder.) At a time when political debates over the ethics of stem cell research, genetic modification, cloning and even nanotechnology are growing at once more fervent and more complicated, Kurzweil offers a vision of technology as destiny, of transformative change that has slipped the bonds of politics, culture, and–for many–credulity.
That his predictions make moot most of the cultural norms and physical limits of today’s world is, he believes, only a testament to the power of the forces he describes. To his many critics, however, Kurzweil is simply spinning fairy tales, preaching transcendence but propagating ignorance.
Arrayed around Kurzweil’s office and in the hallways outside are a few of his inventions. When I asked, he readily showed them off. He had an old Kurzweil Reading Machine flatly declaim the opening of the Gettysburg Address. He played the first few measures of a Beethoven piano sonata on an early-model Kurzweil synthesizer, stumbled, started over, stumbled again, then switched to Gershwin. He arranged a demonstration of a pocket reading machine for the blind that he plans to roll out in January. He told me about FatKat, his artificial-intelligence investment program: Over the past two years, he claims, it has brought in stock market returns of 80 to 100 percent.
Kurzweil is compact and trim, with full cheeks, a small smile, and a knot-like nose drooping toward a broad chin. The tone of his voice, deep and deliberate, is somewhat at odds with his eyes, which narrow and furiously blink as he talks. He is 57 years old, nearly the age at which his father died of a heart attack. According to a battery of controversial tests administered by Terry Grossman, the anti-aging expert who co-wrote ”Fantastic Voyage,” Kurzweil has not aged appreciably in the past 17 years.
Every day, Kurzweil takes hundreds of nutritional supplement pills, and once a week he takes several others intravenously. He is, as he puts it, ”reprogramming my biochemistry” and claims in so doing to have conquered his Type 2 diabetes. More importantly, he insists, he is stretching his natural lifespan until either genetic therapies, microscopic ”nanobots” (hypothetical robots on the scale of single atoms and molecules that Kurzweil believes will be able, among many other things, to take over some of the vital functions of the human body), or simply the ability to download one’s mind onto a computer make immortality a reality.
What links all of Kurzweil’s creations is the concept of pattern recognition: recreating the human ability to distinguish signal from noise. As he sees it, the predictions he’s making are simply pattern recognition applied to history.
The pattern he sees is a simple one: He calls it the law of accelerating returns. To explain, Kurzweil uses the example of Moore’s Law, the storied 1965 prediction by Intel cofounder Gordon Moore that the power of computer chips would double roughly every two years. In 1972 there were 2500 transistors in an Intel chip, in 1974, 4500, and by 2004 there were 592 million.
For Kurzweil, however, the explosive power of exponential growth goes far beyond transistors: Human technological advancement, the billions of years of terrestrial evolution, the entire history of the universe, all, he argues, follow the law of accelerating returns. He has put a team of researchers to work gathering technological, economic, historical, and paleontological data. All of it, he claims, graphs neatly onto an exponential plot, starting out slowly, then nosing sharply upward through the ”knee of the curve” into higher order and greater complexity, arcing toward infinity.
”Ultimately,” he promises in ”The Singularity Is Near,” ”the entire universe will become saturated with our intelligence. This is the destiny of the universe. We will determine our own fate rather than have it determined by the current ‘dumb,’ simple machinelike forces that rule celestial mechanics.” How he is not sure, but he trusts his math.
At such moments, Kurzweil’s predictions have the ring of eschatology, of half-cocked end-times rapture. For him, though, it’s surreal to hear people talk about the size of the Social Security shortfall in 2042–by then, he believes, advances in nanotechnology will allow us to ward off disease and senescence and to manufacture all the goods we want for a pittance. By then, in other words, aging and poverty may hardly exist and people may not retire or even work in a way that’s recognizable to us.
For Kurzweil, stubbornly linear habits of mind explain why, for example, so few neuroscientists share his conviction that we will soon be able to reverse-engineer the brain. ”A lot of scientists,” he told me, ”Nobel Prize-winners included, take a linear perspective. They just intuitively do the mental experiment of what will it take to achieve certain goals at today’s rate of progress, with today’s tools.” Kurzweil points to the skepticism that greeted his forecast, in 1990, that in as few as nine years a computer would beat the world chess champion. He was too conservative, as it turned out: Deep Blue beat Garry Kasparov in 1997.
Yet even among those like Vinge, Minsky, Drexler, and Lenat, for whom The Singularity is less a matter of if than when, Kurzweil is a figure of rare certainty. Nick Bostrom, a philosopher and the director of the Future of Humanity Institute at Oxford University, isn’t so sure the timing of The Singularity can be pinpointed. ”We should be thinking about it more as a probability distribution smeared out over a long period,” he says.
Then there are the many thinkers who find Kurzweil’s case less than compelling. Since his theories take in the whole history of the universe, there is no shortage of points at which to contest them. Some skeptics dispute Kurzweil’s computer science. They argue that even computers billions of times more powerful than today’s wouldn’t necessarily be meaningfully intelligent, much less spiritual. Any one of a number of hurdles–from the complexity of neural networks to the difficulty of recreating the brain’s analog processing with a computer’s digital circuitry to our continued inability to begin to articulate the essence of consciousness–might stand immovably in the way of human-level artificial intelligence.
As John Searle, a philosopher of mind and language at the University of California, Berkeley, wrote in a public exchange of letters with Kurzweil, ”the existing technological advances that are supposed to provide evidence in support of these predictions, wonderful though they are, offer no support whatever for these spectacular conclusions.”
Others, like the Harvard cognitive scientist Steven Pinker, take issue with Kurzweil’s teleological view of evolution. ”It’s the old idea that the process of evolution is some push in the direction of greater complexity–in particular greater intellectual complexity,” Pinker says. ”In one twig of the tree of life, namely ours, having a big brain happened to have advantages. But that’s just what worked for a particular species of primate 5 to 7 million years ago.”
Still others see something darker in Kurzweil’s visions of transformation. Bill Joy, the founder of Sun Microsystems, was so horrified by a conversation with Kurzweil that he wrote a now-famous Wired magazine cover story in 2000 entitled ”The Future Doesn’t Need Us,” describing a technological apocalypse, the earth chewed to pieces by out-of-control nanobots. Thinkers like the political scientist Francis Fukuyama of Johns Hopkins University foresee a subtler corrosion: The pursuit of biological perfection, Fukuyama warns, deprives us of qualities like compassion and courage that spring from an awareness of our vulnerability.
Jaron Lanier, a pioneer of virtual reality computing, and like Kurzweil and Joy somewhat of a tech-world guru, manages to combine the technological and the moral critiques of Kurzweil’s thought. In a 2000 essay entitled ”One Half of a Manifesto,” he argued that our ever-more-powerful computers were likely to be limited, for the foreseeable future, by the software running them. Lampooning Joy’s nightmare scenario, he wrote, ”Just as some newborn race of superintelligent robots are about to consume all humanity, our dear old species will likely be saved by a Windows crash.”
Still, Lanier finds Kurzweil’s ideas unsettling. ”Ray has incorporated in his little system of thought all of the elements of a religion that are selfish but none of the ones that are generous,” Lanier told me. ”His thing is purely, ‘Here’s how to live forever, here’s how to be uploaded into the machine.’ There’s no concern for other people since it’s assumed that everyone will be infinitely rich and happy in his future.” It’s a philosophy based on narcissism, Lanier charges, a dream of ultimate individual fulfillment.
The last chapter of Kurzweil’s new book is entitled ”Response to Critics,” and it is nearly 60 pages long. Kurzweil’s rejoinders are detailed and exhaustive, ranging across topics from software development and neural networks into quantum mechanics and the philosophy of consciousness. Nowhere, however, does he offer any apology for his promise of eternity or his focus on individual enhancement.
This individualistic, mechanistic ethos, his critics argue, also blurs Kurzweil’s predictive power, because it ignores all the ways in which technologies are bounded by social forces. As Harvard’s Pinker points out, ”the track record of technological predictions is laughable. I remember a prediction in my childhood that by now we’d be living in domed cities and commuting by jet pack and eating protein pills instead of meals. On the other hand a lot of revolutions are predicted by no one. My favorite is that in the movie ‘2001,’ you had space travel and human-level artificial intelligence, but people were still writing on clipboards. Arthur C. Clarke hadn’t predicted the laptop.” Autor: Drake Bennett