Gordon Moore was interviewed by Charlie Rose in the autumn of 2005.
The tape of that interview
was broadcast on November 14, 2005. I've excerpted sections dealing with
the origin of Moore's Law, Fairchild Semiconductor, the Intel name, and
a view on integrating development with manufacturing.
(About 28 minutes into the program)
CR: |
And where do you put Moore's Law in your own achievements? |
GM: |
Ah... a lucky guess that got a lot more publicity than it deserved. |
CR: |
It recently celebrated it's, what, 40th anniversary, was it? |
GM: |
The 40th annversary this year. A paper I published in 1965 was the
origin of what they call Moore's Law.
|
CR: |
For the benefit of an audience who is not techologically oriented, what is Moore's Law?
|
GM: |
Okay. Well, I have to give you a little history here. |
CR: |
Okay, do it, please. |
GM: |
In the early '60s the first integrated circuits were hitting the
market. This is where there was more than one component on the silicon
chip. Up until 1960 making a single transistor on a chip was all you
could do. But with the technology, particularly Bob Noyce's invention,
we saw how you could put a whole circuit on a chip, consisting of
transistors and resistors.
And we started doing that and the initial ones were used almost
completely by the military because they were expensive, and the military
was a cost-insensitive application. Small size and light weight were
important to them and they could pay more for them.
But by the time we got to 1965 it wasn't quite clear what was going to
happen. I was given the job of writing an article for the 35th
anniversary edition of Electronics magazine, a trade journal, to
predict what was going to happen over the next 10 years.
And I looked at the few years of history we had. We started out with
this transistor structure, the so-called planar transistor. It was kind
of the beginning of the technology. And then a couple of years later we
had the first integrated circuit, it had four components on it. And
then four, the next year we had 8, and then around 16.
And I could see in '65 that the next one coming had about 60. And I
looked and these and said gee, it's been about doubling every year,
starting with one in 1959. So I plotted... I just said, well we're
going to continue to do that and I just drew a straight line on
logarithmic paper, which was a straight line and so forth, and said
we're going to go from about 60 components to about 60,000 on a chip in
the next ten years.
And because of that it's going to be the route to low cost electronics.
That was really the message I wanted to get across. This was going to
make cheap electronics. I didn't expect my projection to be especially
accurate.
But over that 10 years, instead of ten doublings I think we only got
nine, but essentially we followed the curve for that period. One of my
colleages dubbed this Moore's Law. And it's been modified a bit since
then.
|
CR: |
But let's talk about that. It came to be defined, at least, I may
be wrong about this, is that the power of the microprocessor will double
every 18 months. Is that right?
|
GM: |
That's a corollary of it. My predictions were all on the
complexity. How many components you can put on a chip. And I had one
year doubling for the first ten years, and then I modified it saying
whoops, we're going to lose half of that, it's only going to double
every two years. Again one of my colleages changed it to computer power
and said every 18 months.
|
CR: |
Which colleage was that? |
GM: |
That was David House. |
CR: |
So then... so this has happened. It is said, and tell me if it's
right, that this was part of the assumptions built into the way Intel
made it's projections. And therefore, because Intel did that, everybody
else in the Silicon Valley, everybody else in the business did the same
thing. So it achieved a power that was pervasive.
|
GM: |
That's true. It happened fairly gradually. It was generally
recognized that these things were growing exponentially like that. Even
the Semiconductor Industry Association put out a roadmap for the
technology for the industry that took into account these exponential
growths to see what research had to be done to make sure we could stay
on that curve. So it's kind of become a self-fulfilling prophecy.
Semiconductor technology has the peculiar characteristic that the next
generation always makes things higher performance and cheaper - both.
So if you're a generation behind the leading edge technology, you have
both a cost disadvantage and a performance disadvantage. So it's a very
non-competitive situation. So the companies all recognize they have to
stay on this curve or get a little ahead of it.
Recently, in fact, we've actually accelerated that a bit. It used to be
kind of a three-year period between technology generations. Well, we
want a lead so we're going to pull that down to two years. So it's kind
of squeezed down to a two-year technology generation. Now we actually
accelerate the rate of at which we're increasing complexity.
|
CR: |
So you've accelerated. What does that mean in terms of... so what
do we say about sort of the longevity of Moore's Law then? It
continues?
|
GM: |
It continues for a while. The doubling time is actually shorter now
than the prediction we've been working on recently.
|
CR: |
Will there come a point where it's no longer applicable? |
GM: |
Well, one of the principle ways we achieve this is by making things
smaller and we're approaching the limit that materials are made of
atoms. We're not too far away from that. But talking to the Intel
technologists, they think they can still see reasonably clearly for the
next four generations. That's further than I've ever been able to see.
It's amazing how creative the people have been about getting around the
apparent barriers that are going to limit the rate at which the
technology can expand.
|
ORIGIN OF FAIRCHILD SEMICONDUCTOR
|
(about 35 minutes into the program)
CR: |
So you went to Fairchild to form your own company. You got the
money from Fairchild, who was into aviation.
|
GM: |
Sherman Fairchild had set up two companies, an aircraft company and
a camera instrument company. He wanted to do aerial photography. His
father was the biggest investor in IBM. So Sherman Fairchild had a
fortune in IBM stock and he loved technology.
The people who were trying to find us financing to set up a new company
caught up with Sherman Fairchild and he introduced us to Fairchild
Camera and Instrument and they supported the beginning of Fairchild
Semiconductor.
|
ORIGIN OF THE NAME 'INTEL'
|
(about 37 minutes into the program)
CR: |
How did you come up with the name 'Intel'? |
GM: |
Well that was a tough fight. You have to get a name that you can
clear through the corporation comissioners. Typically then we went
through California and New York. If you cleared both of those states
you were usually okay. We tried all kinds of combinations. I think it
was was about the fifth one that we tried to get cleared and finally
came up with Intel.
|
CR: |
Something, Intelligent, Integrated, something, what was it?
|
GM: |
Integrated and Electronics was really what it came from. You know,
in our logo we have that dropped 'e'. That was kind of the separation
of the two syllables, I guess.
|
MANAGEMENT STYLE AND INTEGRATING DEVELOPMENT
|
(About 46 minutes into the program)
CR: |
Who is responsible for this management style that they [Bill Hewlett
and David Packard] got credit for? It was said they created a different
style of management at HP, at Hewlett-Packard, when they were there.
You also were given responsibility for creating a different style - a
much more open, a much more collegial management style.
|
GM: |
Yes, and I think it must have been the normal trend out there.
|
CR: |
Silicon Valley was different?
|
GM: |
Well, yes, and the industry is different. The detailed technical
knowledge is so important in making the decisions you have to push them
down pretty far. Top-down management just doesn't work well at all. So
the net result is there is a lot more interaction between the top people
and the ones who really know what's going on and I think it results in a
more egalitarian environment.
|
CR: |
You are given credit for, at least, forcing Intel to make sure that
the timeline between an idea in the laborary and a product in a
customer's hand was much more streamlined, much more efficient and much
quicker.
|
GM: |
I guess I deserve credit for something like that. As I said, I had
run the laboratory at Fairchild and increasingly it was becoming
difficult to get new things from the laboratory into manufacturing. The
more technically competent the manufacturing area became, the less
willing they were to accept the laboratory's advice.
So when we set up Intel, very specifically we did not set up a separate
laboratory. We told the development people to do their work right in
the production facility. We'd take the inefficienty in production in
trade for being able to have the technology already existing in
manufacturing when it was developed.
So we eliminated a step, essentially.
In that respect, I'll take responsibility. I didn't want a separate
laboratory doing the development.
|
| |