Demi-Monde of Progress in IT

by Bhaskar Chakravorti

Progress must be measured where it counts: on the migratory path from the laboratory to the living room

Gordon Moore's Law predicts that the processing power of microchips will double every 18 months. Given the microchip's metaphorical and physical significance, the prophecy of Intel's cofounder has become the de facto mantra of the information age.

True progress, however, cannot be measured simply in terms of raw computing power; this is progress made in the labs and the fabs (wafer fabrication plants).

Progress must be measured where it counts: on the migratory path from the laboratory to the living room. Here the pace has been somewhat more deliberate.

For those of us who reside along this path, I propose a second mantra: information technology's value to the user progresses at about half the speed predicted by Moore's Law. I suggest we name this Demi Moore's Law.

So why the lag between progress in the lab and the product's arrival in the living room? It is the extreme interconnection of the information industry that slows the pace of progress. While revolutionary breakthroughs are predicted and produced by our engineers, it is a circle of interconnected decisions made by consumers and producers that acts as a brake on the engine of change.

As consumers and producers, we are all independent decisionmakers with our own criteria. Still, as a consumer, I need my software to be compatible with that of my coworkers; as a producer, I want to make commitments on a product or technology if it is likely to become the predominant industry standard—the next MS-DOS.

Therefore each of us ends up placing bets on which technology or software will be the most popular and used the most often. To a degree, we are betting on the bets being placed by some other strategic player in this game—either our coworkers or a technology maker with a compatible or competing product.

In highly interconnected situations, this other strategic player may, in turn, be betting on us. The resolution of this circularity—this continuous round of betting—becomes harder and more tortuous as:

Our decisions become more interconnected.

Companies understand the value of their product becoming the industry standard and plan for it.

The pace of technological innovation increases, making it more difficult to resolve the endless rounds of betting. The ultimate irony of the "information age" is that it has over-delivered on all three fronts, thereby extending even the "natural" delay between invention and adoption. The result: progress has slowed relative to our expectations.

This is the core logic of Demi Moore's Law. And it will become especially significant as the prevailing obsession with processor speed gives way to the next obsession: transmission speed.

Consider the example of the 56K modem which promised to double, for most of us, the speed of access to the Internet. Despite the introduction of a well-priced 56K modem in autumn 1996, less than 50 per cent of modems shipped in January 1998 were of this variety.

Why?

First, a 56K modem installed in a PC can download information at 56 kilo-bytes per second only when the 56K modem installed on the server operated by an Internet service provider conforms to the same standard.

Second, since 1996 there have been two competing 56K standards: the X2, sponsored by 3Com (US Robotics), and the K56Flex, sponsored by Rockwell Semiconductor. A PC maker or channel would back a particular 56K modem only if its users wanted it. But users would want it only if the ISPs conformed to the same modem.

Most of 1997 was a race between 3Com and Rockwell to align the incentives of everyone in this betting circle towards their version of the modem. However, the International Telecommunications Union announced a common standard for the 56K modem that ended the contest.

Interestingly, shares in both 3Com and Rockwell rose after the announcement of a standard, suggesting that both would have been better off giving in earlier.

The interconnection of decisions and the competition to own the standard caused more than a year's delay here. Not surprisingly, given the relentless pace of innovation, the 56K modem was obsolete before it was in the shops. New technologies—DSL and cable modems—promising speeds more than 50 times as fast will give industry participants yet another reason to pause.

We frequently express indignation and surprise (not to mention the jokes about the "world wide wait" and the "interminablenet") when the hyperbolic promises of the information revolution are not instantly fulfilled.

Yet, given the interconnection of decisions that must occur for a standard to be set and equilibrium to be reached, it is safe to assume that technological progress will be truly realized only half as fast as technology permits.

So expect the bark of Moore's Law always to be louder than its byte. Gordon's law that is, not Demi's.

— from Financial Times