Predicting the Wintel Dynasty's Fall
search.nytimes.com
By ROBERT E. CALEM
It will be several weeks before the supermarket tabloids drag out their predictable assortment of seers and clairvoyants for their traditional year-ahead prophesies.
But George Gilder doesn't need the tabloids or their New Year's traditions to put himself in a frame of mind to prognosticate. Nor does he limit himself to predictions of events in 1997. Gilder, an influential computer industry pundit, has his own newsletter, and the prophesies just come rolling out like bugs from a box of Windows 95 drivers:
A revolution is at hand in the computer industry that will not only unseat the PC at corporate offices everywhere but change the very definition of network computing, overthrow the "Wintel" dynasty and anoint Sun Microsystem's Java programming language as the new monarch.
Even your home won't offer sanctuary from the technology revolution afoot; both your television and
your Wintel PC will give way to new kinds of personal computing devices based on Java and on new
chip technologies -- from vendors other than Intel -- and these will permeate all aspects of our lives.
Gilder made these predictions in the November issue of the Gilder Technology Report, the new monthly newsletter published jointly by his Gilder Technology Group Inc. and Forbes Inc., as well as in a recent interview. His predictions, he wrote, form a "new paradigm" that will see the old one -- the current state of the industry -- shrivel away "into triviality."
Gilder is no backwater oracle. His 1981 book, "Wealth and Poverty," became the bible for the first Reagan White House and sold more than a million copies worldwide.
A well-known futurist in computer industry circles, he is also a columnist for Forbes ASAP, a technology magazine and, as of Nov. 15, a fellow of the International Engineering Consortium. He formed the Gilder Technology Group last April to spread his views, and Forbes began backing his newsletter with the November issue, his fourth.
Next year, the Gilder Technology Report will add a Web site linked to the Forbes magazine site, and the two companies are planning Gilder/Forbes technology conferences next fall.
All of which would seem to be positioning Gilder as the industry's guru du jour. But not all pundits are falling in line.
"Gilder is falling into the trap, stating everything as a binary result," declared Stewart Alsop, a partner at New Enterprise Associates in Menlo Park, Calif., and a not-insignificant prognosticator in his own right. For one thing, Alsop protests this whole paradigm-shift thing.
A paradigm shift, Alsop asserts, is a discontinuous event that "doesn't actually corrupt the one that came before." And he further argues: "Anytime you have a discontinuous event, you can't use the past to look at what's happening going forward."
While Alsop agrees with Gilder that Java is an important development in the computer industry, he denies Gilder's assertion that it will eventually bury the Wintel machine.
And Gilder, like any good prophet, has left the door open just a crack to the possibility that his predictions might be wrong. First, he attaches no time frame to his visions. In addition, he concedes that Intel and Microsoft won't go down without a fight and might even triumph -- but only if they both "become fully committed, new-paradigm firms."
"History is against them," he says of Intel and Microsoft. "But both firms have created history before, and defied it."
So just what does Gilder think we have to look forward to?
"Under the new paradigm," he writes, "the old platforms will gradually wither in the face of network-oriented designs. Leading in unit sales, the most common personal computer of the new era is likely to be a digital cellular phone."
But this won't much resemble today's pocket phone -- nor does it sound like it will be easy on the pocketbook.
"It will recognize speech, navigate streets, collect mail, conduct transactions, contain a Java runtime engine, and command an Internet address," Gilder promised. "It will link to a variety of displays and collapsible keyboards through infrared or radio frequency connectors."
In addition, he predicted, new kinds of Java workstations will be found in homes and offices, and they'll communicate with the outside world through satellite feeds, fiber-optic cables and broadband modems.
"Just as the mobile teleputer will displace telephony under the new paradigm," Gilder wrote, "the fixed teleputer will displace the TV. Both new devices will use Java and the Internet. Neither of the teleputers are likely to be Wintel devices."
Or maybe they will. "You never can tell," Gilder conceded. "Microsoft and Intel are both determined to succeed in the new era."
In the interview, Gilder said of the Wintel dynasty: "The Java movement will emancipate the industry from that anomaly. We're right at the point of a radical change." The two companies' pinnacle of success, he said, is "right now."
While admitting that Intel would continue to be a dominant microprocessor manufacturer, he predicted that the company would find it necessary to sell chips optimized for Java. At the same time, he said, the ascendancy of Java would free the industry of both Microsoft's operating systems and "the legacy X.86 processors," thus opening the market to other chip makers.
What that means, though, is essentially a planet full of network computers. Gilder defends that prediction by asserting that even today "people spend a lot more time accessing remote memories than local memories" -- which is to say that the average PC user spends more time pulling data from the Internet than from the computer's hard drive.
Among the underpinnings of his new paradigm, Gilder listed the following:
1.The Internet affords an immense increase in the power of the average computer by vastly increasing the resources available to it.
2.The cost of liquid crystal displays, or LCD's, has fallen, resulting in lower-cost portable devices.
3.Basic PC architecture is meeting internal contradictions. While processor speeds have increased at a rate of 60 percent a year for the last 10 years, memory access speeds have risen only 7 percent a year. Already, he said, most computer microprocessors spend most of their working time waiting on information to flow into them from memory. And the problem will only worsen unless applications and operating systems manage memory access better.
4.Java is the first truly robust, platform-independent programming language and "model" oriented toward remote access that reduces reliance on a local microprocessor. Therefore, Gilder concludes, in the new paradigm, the definition of network computing changes. Instead of a powerful computer acting as the server to a PC client, the network is the server and all computers hooked to it are the clients. Moreover, he added, with increasing bandwidth to homes, the network computer ultimately will displace the TV as the dominant leisure-time toy. This year, he noted, PC sales dramatically dwarfed TV sales, and, he asserted, "There's no turning back."
Where will homes get all this bandwidth? Through satellite services like DirectPC from Hughes, through cable modem service and from other linkups, he said.
Alsop pooh-poohed Gilder's prediction that Intel and Microsoft would need to fight Java to maintain the PC's dominance.
"It seems to me that people like PCs," Alsop said, noting that 250 million PCs are already in use and that millions more are purchased every year. Thus, for the PC to be overthrown by the network computer, Alsop said, at some point 250 million people would have to decide not to replace the one they have. Moreover, while Alsop agreed that "Java is the killer app of the World Wide Web" and "operates in a different paradigm" because of the way it manages memory, he downplayed the notion that it would alter the definition of client-server computing.
Alsop also painted Gilder's Death-of-TV scenario as a bad prophesy. People will never stop watching TV, or buying standalone TV's, even if Internet and TV functionality are built into a new form of teleputer, he declared. "It's like saying that because you buy a clock with a radio in it that you'll stop listening to your stereo." |
