This first one reprints a story I wrote for Canada's National Post newspaper back in 1999 when Intel has launched it's new "Pentium III" chip - and he reflected on some of the company's earlier mistakes (including a calculation bug in the original Pentium processor that Intel initially brushed off as something which wouldn't impact most people - and then had to urgently fix after a consumer and press backlash).
Here's the 1999 story:
Andy Grove doesn't admit easily his mistakes. Intel Corp.'s chairman took the company from near extinction in the mid-1980s to semiconductor super-star status today by knowing right from wrong decision.
But on the occasion of Intel's Pentium III launch, Grove is in a reflective mood. Perhaps it is because he has pulled back from day-to-day involvement in the company, ceding the CEO and president's job to Craig Barrett. Perhaps it is because he recently fought and won a very public battle with prostate cancer.
Mr. Grove now looks back on the famous "Pentium bug" fiasco in late 1994 and early 1995 with a self-critical eye and admits that he has learned from his mistakes.
Back then, the newly-introduced Intel Pentium processor was being heavily promoted to consumers as the chip that would really take their computers into the multimedia world - allowing them to use graphics, sound and video as never before. At the height of the 1994 Christmas sales season, a "bug" was discovered in the chip - a bug that would cause it to inaccurately complete a certain complex series of calculations.
For the majority of users, this bug would never intrude on the use of enjoyment of computers with Intel Pentium processors inside them. Intel said so - and offered only to help out those few consumers that it felt really needed help.
Consumer groups in the United States, however, took a different view and were shocked that Intel should presume to tell them which bugs were and were not a problem. "The customer is always right!," they told the company. Chastened, Intel eventually did move to widely distribute bug-fixes to the problem and Intel Pentium processors have enjoyed strong popularity ever since.
At the time, however, it was no sure bet that the Pentium would succeed. Intel competitors such as Advanced Micro Devices (AMD) had successfully produced “clones” of Intel’s earlier generations of microprocessors including the 80386 and 80486 and stood to gain if consumers backed away from the Pentium (which, at that point, had not been copied by anyone). Any perception that Intel’s future technology path was flawed in some way could have had a strong impact on the company’s market share and sent consumers looking for PCs with other manufacturer’s processors.
This was a very real fear – and one that Intel knew well. Back in the mid-1980s, after Intel enjoyed its first flush of success when its 8086 family of processors were used in the original IBM PC, the company came very close to going under.
Arch-competitor Motorola was enjoying huge popularity in Apple’s Macintosh, the Atari ST home computer, the Commodore Amiga and the Sinclair QL. And the notion of having desktop computers with the same company’s processor in them – generation after generation – had not yet taken hold.
Intel also had no guarantee that IBM would stick with its processor family and faced heavily competition from much larger Japanese manufacturers. The release of the 80386 in 1985 was really the turning point, after which Intel knew that it – not IBM – was in the driver’s seat. This was because the 80386 was a huge success and Compaq, not IBM, was the first manufacturer to ship systems based on the 80386 processor. And it enjoyed huge success as a result. IBM thereafter felt compelled to follow.
But we digress. The lesson from the Pentium bug experience came home to Mr. Grove in January when reports started to appear that Intel would be featuring a processor serial number (or PSN) on its new Pentium III computer processor. This electronic serial number would allow computer companies, corporate users of computers and operators of Web sites to use software to call up the serial number of the processor and keep track of which computers were doing what.
This time the outcry came from privacy and civil liberties groups concerned that processor serial numbers made it easier Web site operators to more easily track the buying patterns and interests of consumers.
This time, Intel reacted differently - working with both interest groups and the rest of the computer industry to come up with a solution to the problem even before the Pentium III came to market. So when Pentium III systems were ready to ship at the end of February, Intel was able to announce that it had developed a utility that would allow users to switch the processor serial number feature "off" if consumers did not want to have their systems automatically identified.
"We are a lot less righteous about things than we were four years ago," says Grove, admitting that Intel redoubled its quality control efforts after the original Pentium bug problem and has become much more responsive to consumer-level concerns. "Having said that, no-one will ever built a perfect microprocessor," he adds. "In retrospect, the problem in 1994 was not the problem (with the chip) but the way we handled it and reacted to it."
He is even philosophical about the company's current legal battle with the United States Federal Trade Commission (FTC), although he strongly denies that there is any merit to the notion that Intel has engaged in any unfair trade practices. The FTC has charged that by withholding products and products plans from Intergraph, Digital Equipment Corporation and Compaq Computer, Intel illegally used its "monopoly power" to force those companies to cease pursuing intellectual property claims they were then making against Intel. Intel denies any wrongdoing and says it was only taking appropriate steps to protect its intellectual property.
No comments:
Post a Comment