Moore's Law, Kurzweil and Telecosm Stocks

George Gilder
November 3, 2005
Print ArticleOriginal Article

GREAT BARRINGTON, MASS. - Overflowing the Resort at Squaw Creek in Lake Tahoe, Calif. in late September into hotels in the nearby ski village, this was the Telecosm of the “Singularity.” A singularity designates a point in the future beyond which the “event horizon” darkens, as the horizons of the past darken beyond the reaches of the Big Bang. In between--we are to believe--is the known universe.

But still in the dark remains the question of when and whether Broadwing will break out into profits using MPLS (multiprotocol label switching) on the intelligent edge and fast all-optical switching at the core of its still industry-leading network, while Cisco struggles to keep the smarts in the core.

Even the present is enigmatic. We have little assurance whether bandwidth prices are stabilizing--as Jay Adelson of Equinix reported in a fascinating speech--or whether they are continuing their downward plunge as confidently testified by Cogent CEO Dave Shaeffer. Nor, even after all of the earnest explanations of CEO Eli Fruchter and CTO Amir Eyal, do we know when EZchip will begin an explosive ascent of revenue for ten-gigabit Ethernet and line-card processors. And despite the presence of representatives of both Foveon and Synaptics , we still don’t know when these two kindred companies will burst into the huge markets for “tele-puter” sensors and imagers (though news from Foveon has been picking up since the conference, including the special “Progress Medal” from the Royal Photographic Society in London).

Introducing a dazzling new bestseller, The Singularity is Near, and generously giving a copy to each of the attendees, Ray Kurzweil acknowledged that macro-futurism, projecting Moore’s law in all directions, is much easier than the micro of predicting what will happen to specific companies and technologies.

Nonetheless, on stage the first night of Telecosm, Kurzweil faced a skeptical micro question from yours truly on the dismal failure of several teams of robotic engineers last year to create a device that could negotiate a Defense Advanced Research Projects Agency (DARPA) course through the Mojave desert without plunging off the road into a ditch or an infinite loop. In response, Kurzweil confidently asserted that teams from Carnegie Mellon and Stanford would succeed in this task in October. Sure enough, he was on the button with this prophecy.

So far, so good. At the heart of his larger prophecies is the continued exponential progress of all the arts and sciences of information technology on beyond machines into a biological “singularity.” Ray’s intriguing argument is that today’s exponential curves merely follow in the train of the original evolutionary curve--which also reveals an accelerating pace of advance. It only took 13 billion years from the exquisitely calibrated bang to the biosphere, with DNA processing in the eukaryotic (nucleated) cell, then the Cambrian explosion of life forms some 3 billion years ago, and then the rushed ascent of punctuated equilibrium to the emergence of man and Ray and the Telecosm list, after which things really start popping.

Discerned in all this heroic ascent is scant intelligence at all until the arrival of human technology, though the information processing underway in the some 300 trillions of cells in your body, each with some 6 billion base pairs of DNA programming, excels the output of all the world’s supercomputers with all their intricate software and firmware. As Ray points out (page 209), the ribosomes that translate DNA into amino acids accomplish 250 million billion “read” operations every second just in manufacturing the hemoglobin that carries oxygen from the lungs to body tissues.

While the genes are digital, much of the biocomputing is inscrutably analog. But in another four decades, so Kurzweil calculates, digital-machine intelligence will exceed human intelligence, precipitating the “singularity.”

Humans, he predicts, will use the machines massively to extend our lifespans and to project the reach of our learning by mastering the mysteries of consciousness within our own brains and out into space, with an imperial march of human intelligence incarnate in our machines and in our newly bionic bodies.

It is a grand and triumphant trajectory of thought on which Kurzweil is launched, and his argument is finely mounted and gracefully written, with much self-deprecating humor in artfully shaped “dialogs” at the end of each chapter. But as some attendees groused, it would be nice if by the time of the “singularity,” or even before, Microsoft (nasdaq: MSFT - news - people ) could get Windows to boot in less than four seconds and could avoid the darkened event horizons of its chronic blue screens.

And after many projects at Caltech attempting to use neuromorphic models as the basis of electronic simulations of brain functions, Carver Mead observed that we still have no idea of the workings of the brain and nervous systems of a common housefly. As I describe in The Silicon Eye, it goes about its business, eluding the swatter and garnering chemical sustenance in the air, all on microwatts of power using means that remain beyond the grasp of our most sophisticated neuroscience.

Oh, well, observed my colleague Nick Tredennick, all these exponential curves look flat to the engineer attempting to solve the immediate problems he faces. So back to work, folks.

Excerpted from the November edition of the Gilder Technology Report.