Punk Rock & Moore’s Law

Fact: Intel’s CEO says “We have new processors that have 250 million more transistors, and yet are 25 percent smaller than today’s version and don’t require more electricity to run.” (Interview published 2/1/2008 )

Analysis: Moore’s Law: immortal, or destined to be broken?  Punk Rock: dead, or in revival?  And why were Johnny Rotten and one of the legendary Traitorous Eight in the Fairmont Hotel in San Jose on the same night back in 1989?

Once-and-future Sex Pistol John Lydon (Rotten) was there after playing a gig up the road, and I happened upon him in the bar, where he proceeded to buy round after round of Heinekens for me, him, and his roadies. He latched onto me because he wanted to talk about American politics, and to his delight I reminded him of some caustic things (surprise) he had had to say over the years about politicians like Reagan and Carter. That just got him started, and we wound up laughing pretty drunkenly into the night in the Fairmont’s swanky lobby bar.

But earlier that evening, I had shown up at the hotel with friends to attend the Silicon Valley Business Hall of Fame dinner, where Bob Noyce, co-founder of Intel, was being honored along with others….

Also in the crowd as I recall was Gordon Moore, and as a young tech junkie I shook his hand — as I have reminded my wife the lawyer, he’s the only person I’ve ever met with a Law named after him!

Well, Electronic Engineering Times of Asia has a solid, meaty interview this morning with Paul Otellini, CEO of Intel. The article’s titled “Saving Moore’s Law,” and Paul is quoted as saying “It’s not an exaggeration to say that we were heading toward a premature end to Moore’s Law,” before making several recent advances including implementing the high-k metal gate in Intel’s latest transistors, for their Penryn processor in the Core 2 Duo series. 

I’m threatening the limits of my knowledge of the tech, but Intel began mass-producing Penryns with the next-big-thing 45 nm fab process in November, and along with IBM and several traditional chipmakers, Intel sees great promise in 45 nm.  As others are noting, that dovetails with Microsoft’s serious interest and R&D work on manycore and multicore processing, and the astounding advances we expect for general purpose computing and software, as laid out in our white paper on “The Manycore Shift.”

That “end to Moore’s Law” line just caught my eye, because I’ve been in a running debate for several years with others who think Moore’s Law is always just on the verge of  obsolesence, or irrelevance. I disagree, and for a couple of decades that’s been the smart stance to take, and likely will continue to be.  The high-k metal gate is just the latest related advance that nudges the world back into rough alignment with Moore’s projections, and there’s always another around the corner.

The importance of this is the performance benefit to consumer and enteprise computing that spins off from constant advance; and one interesting race to watch is the international Olympic-scale sprint among nations for the best R&D to support processing advance. I often hear talk of, “We should do a Sematech” for this or that segment of the tech industry, as if it’s easy to convince a Bob-Noyce caliber leader to take on an entire foreign nation’s industry segment (as he did successfully with Sematech, taking the CEO role there in ’89 and by ’92 helping to recapture the lead for the United States in semiconductor sales revenue for the first time since ’84).

That’s the rub – the technical side of progress advances, not inevitably, but pretty darned inexorably. It’s the human front of progress that limits us, or fails us. When I met Johnny Rotten, he was adamantly swearing there’d never be a Sex Pistols reunion and that he was determined to make only new music. Time (and the lure of easy money) changed that, unfortunately.  No more PiL with its revolutionary sounds.

Moore’s Law, thank goodness, is not hostage to one man’s personality, relying instead on the ingenuity of crowds (well, crowds of Ph.D.’s) and the spark of creative genius represented in large-scale and determined R&D projects.

By the way, I have a neat reminder of that evening: a program for the “Silicon Valley Business Hall of Fame” event, its cover scribbled on by Johnny Rotten, decorating his signature (“John Rotten – a True Star”) with obscene topless drawings of my date for the evening🙂


AddThis Social Bookmark Button

4 Responses

  1. Lewis,

    You are so right about Moore’s law, and I especially like the way you put the most important point: “The importance of this is the performance benefit to consumer and enteprise computing that spins off from constant advance.” I would bet money that Moore’s law (“the number of transistors on a chip will double every two years”) will continue for at least another 10 years, which means a continuing exponential growth of transistors on chips. The neat thing is that even this transistor growth is not the best measure of the growth of computing power anymore. Actual computing power is increasing far faster than Moore’s law. With the ability to put multiple cores on chips and run multiple threads per core, significant enhancements in I/O and processing have occurred which move things forward faster than Moore’s law.

    Even when Moore’s law totally gives out, I think we can safely say the mega trend of computing power will continue. The way Kurzweil puts it is: “Information Technologies (of all kinds) double their power (price performance, capacity, bandwidth) every year.” He is a very modest man who does not call that Kurzweil’s law, but he is the first who wrote it that way and maybe it should bear his name.

    Cheers,
    Bob

  2. Hi Bob – great comments and naturally I agree with the thrust. One note of gentle contradiction: Ray Kurzweil actually does call it “Kurzweil’s Law,” and translates that simply as “the law of accelerating returns.” He makes a persuasive case, though, so he can & should be forgiven. Here’s his best description of, in his essay “Kurzweil’s Law” from 2003:
    http://www.kurzweilai.net/meme/frame.html?main=/articles/art0610.html?

  3. Thanks for that Lewis. I should have done my homework and I would have seen that there is something already called Kurzweil’s Law. I would point out that it is not the same point as the observation I wanted to call Kurzweil’s Law, but the two are related.

    What the paper he did for Edge calls “Kurzweil’s Law” is something he currently calls “The Law of Accelerating Returns,” which is a scientific theory he backs up with 25 years of research. “Moore’s Law” is part of this phenomenon, as is the observation I quoted above that “Information Technologies double their power every year”.

    But the key point I wanted to highlight, which makes me very optimistic for our collective future, is that the benefits to mankind should be increasing faster than Moore’s law’s doubling of transistor density rate. Both Kurzweil’s observation and Moore’s law are exponential, but Kurzweil’s observation is of a doubling every year and Moore’s is a doubling every two. And Kurzweil’s is not tied to transistors and chips, it applies to all things technical (software and hardware).

    We are living in interesting times and they are about to get more interesting.

    Cheers,
    Bob

  4. I stand corrected, and again agree with your points; I’m a bit dubious about a particular “singularity” but a big fan of Kurzweil in general and his enthusiastic pursuit of understanding future advances.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: