Another One Bites the Silicon Dust

At imbX, Singapore's three-part mega tech expo last week, Apple co-founder
Steve Wozniak spoke of the pursuit of simplicity 'and fewer chips', further
shaped by severe cost constraints, that drove his 1970s game box and computer
designs...which later led up to the Mac.

Then he 'discovered' the microprocessor, which had been around for two
years by then. It represented the reduced component-count simplicity he
desperately sought. But that first Intel microprocessor cost $400. "There
was no way I could afford one," he said. He finally found a $40 chip from
Motorola, and the rest is history.

That history is now about to be changed. One of the last few frontiers for
the world's largest chipmaker (PDAs and big iron...Sun and game boxes...) is
breached: Apple announced recently that future Macs will run Intel chips. An
Intel Mac? What next? Windows on the G6?

In hindsight...it wasn't unexpected. Keeping processors ahead of the power
curve of Moore's Law has become so expensive that few have managed.

But hindsight can't make me happy about it. And that's not just because
an Apple is what I started my tech experience on, 22 years ago, or that the Mac
is a tech lifestyle, an innovation benchmark that Microsoft, IBM, et al,
struggled to follow for two decades.

No, my unhappiness is because there really is no alternative. And that's a
bad idea in the tech world (or in any world).

I've felt the same way about the entire handheld/PDA world, including the
once-iconic, pioneering, and 'different' Palms, moving to XScale chips three
years ago. My ancient Palm Vx runs a Motorola 20 MHz chip, but now everything's
Intel, from the Treos to the LifeDrive. So are the PocketPCs, and various
smartphones. And the big servers are moving to Itanium. And Sun uses AMD x86
chips in blade servers-can Intel be far away?

So where does Intel expand if it's hitting 90% share levels in the
processor world? To the other silicon around the processor. That's the 'platform',
Intel's new strategy for world domination. The platform is the processor,
chipset, and base software. As such, it's a good thing for performance when
the platform is made by one vendor: like the Mac itself, or the Centrino mobile
platform.

So what's the problem with a world where almost all the core silicon is
made by the folks who wrote the book on processors, and defined the tech curve
for others to follow? The same as any area with a monopoly. No competition.
Thus, one less motivator for innovation. And processor progress is not just
about Moore's Law. It's about cost, power management, form factor, and the
ecosystem.

That's the challenge Intel will have to tackle: innovation, even with near
zero competition.

Prasanto K Roy