Oral-History:Bruce W. Everitt

About Bruce W. Everitt

Bruce W. Everitt graduated with a degree in electrical engineering, then went to Harvard Business School for his MBA, taking a leave of absence from the Motorola Communications Division. In 1959, Everitt became an analyst, working in the semiconductor industry. Everitt moved out to Boston in 1964 to form a company specializing in technology investments named G.S. Grumman and Associates, which later merged as Institutional Division of Cowen and Company. As of this interview in 1996, Everitt was a fund manager.

In this interview, Everitt discusses mostly the growth of the semiconductor industry itself from the 1950s onward. He talks about transistor production problems and the need for inspection requirements, and the entering of companies such as Texas Instruments, Fairchild and Intel into the industry. The issues of investment and innovation are spoken of at length throughout the interview – how both were arrived at and may have influenced each other. Everitt also speaks to the importance of flexibility within management and the decision making process, disagreeing with the Japanese model, and for companies to not ‘rest on their laurels.’ He also discusses the importance of the microprocessor in the growth of semiconductors, but also the future importance of digital signal processing and memory. Besides talking about particular companies in the industry, Everitt also talks about people such as Les Hogan and William Shockley.

About the Interview

BRUCE W. EVERITT: An Interview Conducted by Andrew Goldstein, IEEE History Center, 9 May 1996

Interview #268 for the IEEE History Center, The Institute of Electrical and Electronics Engineers, Inc.

Copyright Statement

This manuscript is being made available for research purposes only. All literary rights in the manuscript, including the right to publish, are reserved to the IEEE History Center. No part of the manuscript may be quoted for publication without the written permission of the Director of IEEE History Center.

Request for permission to quote for publication should be addressed to the IEEE History Center Oral History Program, IEEE History Center, 445 Hoes Lane, Piscataway, NJ 08854 USA or ieee-history@ieee.org. It should include identification of the specific passages to be quoted, anticipated use of the passages, and identification of the user.

It is recommended that this oral history be cited as follows:

Bruce W. Everitt, an oral history conducted in 1996 by Andrew Goldstein, IEEE History Center, Piscataway, NJ, USA.

Interview

Interview: Bruce W. Everitt

Interviewer: Andrew Goldstein

Date: 9 May 1996

Location: Wellesley, Massachusetts

Early Career and Microelectronics

Goldstein:

This is Andrew Goldstein and we are talking with Bruce Everitt at his office in Wellesley on May 9, 1996. I thought a good way to start might be to talk about your early career and how you became involved with microelectronics.

Everitt:

Initially, I was an electrical engineering graduate and then went on to Harvard Business School. While I was at the Business School I was on a leave of absence from Motorola, where I worked in the Communications Division that was at that time all of about four million dollars in 1950. When I came back and through 1959 I was with them, but I got quite involved in the early use of semiconductors when they designed the first pocket pager which was one of the really first industrial uses of transistors versus the military. At that time, trying to get one thousand transistors a week was really difficult. It was still a black art and we were finding reject rates of ninety percent when they got to us. That was partly because I had learned that by following the downward drift of the beta, you could bake transistors at 80 degrees centigrade for two weeks, it would stabilize the beta and then it would plateau out. But if you didn’t, why your equipment would go out in three or four months because [of] the downward drift from the gain of the transistor. I can remember having real fights with Texas Instruments and a few companies like that, that said, “You can’t do that.” These were germanium transistors. Anyway, working with the companies I felt that I had really got involved in establishing some of the first incoming inspection requirements. I handled the responsibility in that division while I was in production.

Each of us in production engineering had certain things that we were responsible for; the incoming inspections, and seeing that the specs were met. I had plated circuits and semiconductors, among several other things. So I got involved with that. Also I got involved because we were looking for suppliers for what we call the 2N111, which was one of the first basic germanium transistors. That was one where Motorola wasn’t able to supply us in quantity, and we had six weeks before we were to go in to production. We had to scramble around and had to actually redesign the unit from P-N-P to N-P-N transistors. At that time, there were only four companies we were looking at, Texas Instruments, RCA, Raytheon, and Motorola. Motorola couldn’t deliver it at that time. Raytheon and even RCA still had the problem that their salesmen were not technical people, and so you would ask them a question and they would have to go back to the plant. So it went back and forth. At that time Texas Instruments sent in an engineer to sit down with the circuits and work with you. It was pretty apparent to me early on in that stage, Texas Instruments was going to become a major factor in that field.

Goldstein:

What year was this?

Everitt:

This was about 1955. We ended up with Texas Instruments as the sole and only supplier who could supply what we wanted, and we still had the go-around with them in setting the specs and having as much as ten times the amount in the pipeline due to high reject rates. We would reject ninety percent and they would go back to their plant. It was working out what were the problems. You would find also that the production would be inconsistent in quality. That was when they were trying to put silicon grease on the junctions to help protect them, as they hadn’t learned about passivation yet. So, it was an interesting time, and as a result, I got to know a lot of people in the industry.

Semiconductor Industry: Fairchild, Intel and National

Then, before I left Motorola when Fairchild Semiconductor was formed, my college roommate (actually a Chinese fellow who had lived with our family at the University of Illinois) went with the Beckman-Shockley group. They were doing some of the most innovative work in semiconductors, but Shockley himself was a strange character; he just wouldn’t freeze anything for production. So they had all this great development but never got any products out. Frankly, the group began to get a little disgusted with him because they felt so throttled. Plus, he began to get illusions they were against him and wanted them to take lie detector tests when his secretary cut herself on a nail in a door jamb. So that sort of did it, and a group of them went out seeking a home. They looked for some time before they found Sherman Fairchild who became interested enough to say, “OK. I’ll stake to you the semiconductor operation and have options on Fairchild Semiconductor to be bought out by Fairchild Camera and Instruments completely. And it went well enough that they did. It became a subsidiary of Fairchild Camera and Instruments. Over a progression of time Fairchild Semiconductor did very, very well. It developed a real reputation in the field. Meanwhile, Sherman Fairchild was blowing the money in other parts of Fairchild on everything from sound movie cameras to who knows what, and the semiconductor group was making all the money for Fairchild Camera and Instrument. Stock was going nowhere and the semiconductor people were getting pretty frustrated that they were earning twice what Fairchild was reporting and as a result, their options weren’t going anywhere. Finally, Noyce, Moore and a group of them spun off and started Intel and that sort of started the exodus.

Charlie Sporck got picked off to head up National Semiconductor, and it was Peter Sprague that financed that one. Actually, National Semiconductor was a no-name discreet component company in Danbury, Connecticut, so they really just took the shell of that company and turned it into the new National Semiconductor. John Hughes, who was the VP of Finance of Perkin-Elmer right in the Norwalk area (because he knew Peter Sprague) ended up being the key financial officer for National Semiconductor. Charlie Sporck had really been the production manager at Fairchild and a little bit later Jerry Sanders, who had been the sales manager, spun off and started Advanced Micro Devices. Fairchild Semiconductor was thus the progenitor of the semiconductor industry. One way or the other it came out of that group. I can honestly say it started by a series of problems. First at Beckman-Shockley who wouldn’t move ahead, and then at Fairchild who was wastefully spending the money developed by the semiconductor group.

Luckily, Noyce and Moore had such good reputations at Fairchild, which was doing extremely well in semiconductors, and everybody knew of them. Thus, when they went to form Intel--let’s see that was 1968--they were inundated with people who wanted to finance the new company. They ended up with more than enough money to get started from out of that group. There was enough interest by then in semiconductors that money flowing into the industry from venture sources really began and it was with the start of Intel, I would say, that all of sudden you broke out with individuals forming or reforming new semiconductor companies. Now Peter Sprague was the money behind National as he was part of the Sprague family. He was the maverick of the Sprague family who had left the Sprague operation to do his own thing as he was really more of a venture capital type of a guy. The only other one back in the very early days of semiconductors, of course, Texas Instruments, developed with a good base that by today's standards looks small: about a two hundred million dollar company in the geophysical exploration field, but they developed and were one of the earliest ones in really doing things with semiconductors. They developed in the germanium field and their big win was in transistor-transistor logic, or TTL. This later became their albatross, as they did so well that during the transition period to CMOS technology, they had development and transition problems.

Intel started as the first company that quickly moved solely to work in silicon and they rapidly moved from the earliest silicon-based discreet units to integrated circuits right at the start. Jack Kilby of Texas Instruments was the first guy to piece together an integrated circuit while Noyce the next year was the one who began making the connections by putting it all on a piece of silicon and creating the circuitry by implantation and such. So, they really both get credit for the integrated circuit but as I understand it, Kilby was really first with the concept while Noyce was the one that really put it into effect.

Goldstein:

That’s the way the story is usually told.

Everitt:

In fact, in the twentieth anniversary issue of Intel they had a very good write-up on the development of their operations and then Jerry Sanders and AMD wrote one not too long ago, a treatise called “Silicon: The Super Highway of the Future.” He documented much of that history.

From a financial standpoint, money didn’t create any problems. There were enough people out there in Silicon Valley who knew what Fairchild had done that as soon as Intel, which was the first spin off of any significance, was organized, they had no trouble in getting financing. In fact Intel probably could have gotten ten times more than they got. They were fighting off investors and of course when Intel finally went public it was extremely successful.

G.S. Grumman and Associates and Venture Capital

I think the recognition in the financial field of the potential for semiconductors had already begun by about 1964, as that was the year I moved out here to Boston to help form a company that specialized in technology investments. It was called G.S. Grumman and Associates, and later merged to become the Institutional Division of Cowen and Company. That is still going today, so it was 1964; that year one of the first write-ups and recommendations that I did was on Fairchild Semiconductor. I can remember a lot of institutions thought I was crazy, but Fairchild Camera and Instruments ended up being the largest gainer on the New York Stock Exchange that next year. That also said that the financial field, from an investment standpoint, was starting to recognize the semiconductor industry as a pretty exciting place to invest. I am trying to think of any other progressions on the thing because it just started carrying itself from then on pretty well.

Goldstein:

You say that it wasn’t hard to get venture capital for companies in the wake of Intel. So does that effect the way a business plan is developed, organized and presented if the money is scratching at the door?

Everitt:

No, because I think venture capital people on the west coast, particularly, like Arthur Rock, were pretty sharp and pretty sophisticated. So they still required a pretty good business plan and wanted to know the people. I think more than anything it had already begun to develop by that time where they could check on or already knew those that had technical capabilities. They laughed in the early days, they said the problem was, people wanted ones with ten years experience in the industry and there weren’t people with ten years experience; they were all young thirty year olds when they started. They laughed about it in the financial field. At one time in venture capital you couldn’t meet one of the requirements which they had. People who had been in their field for at least ten years and who really know what the field was all about. They were creating the field at the time. But I still think that yes, you did have some times where money was kind of thrown at one.

I think the best example was the early days [of] that Transitron. Not from the investment because they got started with [the] Bachalar’s. Actually the one Bachalar was a junk dealer and he had the money that he put in and his brother David Bachalar, essentially took the gold bonded diode that he worked on at Bell Labs and appropriated for the first product, Transitron. That was a very successful product, but they didn’t really have the development capabilities. That was one that blew up in people’s faces. Not too many years down the road, but I can remember it was written up in Fortune magazine, that the big exciting company - I had forgotten what year that was - it came apart. But with that exception and that was not venture capital money, it was the Bachalar’s, almost all of the others were a lot of little companies that didn’t do well and probably got more funding than they should have. Any area where the product areas are getting exciting, you always have capital people that are pretty sharp and they look in great depth and then you go down several levels and you will have somebody that will still put some money into some things. So they had a lot of semiconductor companies that have disappeared from the scene. But more of those rather than go bankrupt, just sort of melded into other companies and disappeared from the scene. You have had a few.

There was a period where Solitron was big and it was sort of a semi-junk dealer that did a lot in obsolete “sunset” semiconductors. Then you had a lot of companies that were in discreets that didn’t make the transition to integrated circuits. It was a lot more sophisticated. That’s where it began to separate out who were really the ones with in-depth processing capabilities and design capabilities, and the jump from discreets was a tough one. Unitrode, a very outstanding company in the discreet business. They had developed the glass bonded diode or the metal seal and that made them very good. But then they had a tough time going to integrated circuits and their diode business was sort of fading away.

Military vs. Industry Impact and Development

I think one other thing to look at in the development of semiconductor industry is fairly important. For Texas Instruments, Motorola, Westinghouse, and some of those that were in it early including RCA, the indirect financing came from contracts for the Minuteman and some of the earliest missile systems. That gave the first volume requirements for the industry before industrial and then commercial use developed. That was where a lot of money came from, not by investing in the company but by giving them a source of some volume to run.

Goldstein:

And the common interpretation of that pattern is that the direction of the developments was shaped by military priorities.

Everitt:

Definitely.

Goldstein:

That is easy enough to accept, it seems well-documented. But then the question that I have is, once other sorts of investment money, other support for research came in, what impact did that have on the sort of developments that you saw? What impact did it have on the outcome of research and on the nature of the research?

Everitt:

It is a little hard because it was a transitional type of a thing that moved slowly, and yet I think again probably the defining point was the spread out from Fairchild. Intel directed very little basically at military. That didn’t say they refused military but they weren’t defined by the military. I would say the early days of Texas Instruments, Motorola, Westinghouse, RCA and even to an extent Fairchild were focused on the military as the first users and good money went into R&D contracts to develop things for them. Though the reliability of the industry got its fundamental base in the mil specs particularly for Minuteman. But then I think it is kind of hard to say that it was the money fall or whether it was sort of the timing of developments to get the cost of semiconductors down to a place where industrial applications started to make some sense. But also I stop and think [about] the pager in 1955, which was an early user of transistors.

Goldstein:

Or hearing aids.

Everitt:

Well the hearing aid was really the first one. But that was driven not by cost, but by size that allowed them to do things that they hadn’t done before.

Goldstein:

Right. There was no genuinely competitive product because you could make them small.

Everitt:

That’s right. It was a very different thing because they wanted to miniaturize. The first ones were put in eyeglasses, if you remember, you probably wouldn’t remember. But they were big thick pieces in the eyeglasses - they were trying to get to the point of it not being apparent in what you were doing. Those were again pretty simple semiconductors, they were just amplifiers.

Goldstein:

I want to contrast the research that was done during this military-driven age with the research that was done differently. Does a military application foster a different kind of development procedure than an industrial or commercial application?

Everitt:

Well I think several things: for the military, costs had nothing to do with it, it was just to get the job done. And so they didn’t have to focus on costs, but they also had to focus on the unbelievable amounts of paperwork for documentation and such. It almost became a thing of itself where the minute you moved towards the commercial, it was almost like comparing car radios where the costs right down to pennies became important. Well the cost at least in dollars became important. I think some of the first applications like the pocket pager were made possible in new product categories that w[eren]’t possible until semiconductors came along. And those first ones still were comparatively not - I think it was seven hundred and some dollars for a pager of that kind, but that is still for something that hadn’t been done, but it didn’t break into broad market to where you brought the costs down. But I think that it began to direct attention to mass productions of items at a very cost effective type of level.

Goldstein:

Is this an issue, if you are doing military research perhaps there are confidentiality restrictions, whereas if you are doing other kinds of research then it is easier to publish and function as part of a scientific community a little bit better? I know in the beginning of the century it was very important for General Electric and AT&T to make their research labs like small universities because those were the expectations that the staff had. So I am wondering if the changeover involved issues like that, whether commercial or industrial research could engage in an exchange of results and information that when you look at the research community as a system functioned differently than it would of if they were doing military research.

Everitt:

I think it did, but [for] the military I don’t think the secrecy thing on the component level was a problem, it was maybe for some of the circuits.

Goldstein:

The systems were secure.

Everitt:

But for developing a transistor that was the same kind of transistor you could use over on the industrial side, except that it often cost too much for what you were doing. But I think you just change the whole level of research which I probably better compare, you have military, you have industrial and you have consumer. Well if you are in the industrial area orienting yourself to the consumer areas is a step function jump. You saw that at Motorola where we were in industrial, where we were making quantities of fifty radios a week and our sister division across the street was making thousands of TVs. And they couldn’t use the same kind of things we could because they had been so cost effective they had to be watching every penny. The quality of the resistors and transistors, they could take chances on some things going wrong, not as much on the quality but maybe on the long term reliability and things like that. Industrial is another step up and then you probably are three steps up to the military on reliability: meantime-to-failure type of things. I think [that] probably drives it as much as anything. Your development has to be in a much shorter time frame, you have to look for a payback in industrial. Your products are potentially going to become obsolete faster and so I think it is a whole different orientation.

Most of the people who were working on military development whether semiconductors or other parts of the electronics just have a totally different orientation and that makes it much harder to be dynamic in the commercial and then [there’s]another type of person in the consumer areas. I think that is one of the things people talk about with the break-up of Bell and such. Could the guy from Bell Labs carry out the innovativeness that was necessary? And yet you took the three that invented the transistor. Shockley was the first to start a commercial enterprise, he couldn’t pull it off. And I think he was too imbued with the way Bell Labs did it, he sort of thought he could always be better. He was like a perfectionist and didn’t recognize it. Somewhere you have to say that’s good enough for now, we have to freeze it and go to production; he never could freeze anything for production, so it was a good example of that progression. He didn’t come out of the Bell Labs complex, and yet he developed under his tutelage a lot of good guys. So it wasn’t the technical as much as it was the product orientation. And when they went up to Fairchild they began to get into product orientations even though they did a lot of work for the military which helped pay their way on things then the industrial sort of grew out of that.

Consumer-Industrial Growth

I think the big step in the industrial came about the time Intel was formed. However, Texas Instruments was the earliest one in 1956 when they first developed the transistor radio.

Goldstein:

It was actually marketed in 1954.

Everitt:

In 1954, okay, that’s when it was and that was [a] pretty crude product, but still it developed the first consumer application of any size. Then about 1970, the electronic calculator, and that first one was actually done by Rockwell and I remember bringing one of those home to demonstrate. The thing looked like a great big box with Nixie tubes on it and everybody was just aghast at the small size of it. Because you have been using Marchant calculators that did their division by successive subtraction and they chunk along. They were electric not electronic. You realize it took a long time [until] you get up to the calculator in 1970. That I would say was starting to move into a borderline consumer-industrial. Really it was [a] business type of application, but then within just a few years, T.I. came out with another calculator that was quite a bit smaller, but still a fair size. Then it wasn’t until the integrated circuit started to really take a hold in applications there that you got down. There was a long period in between on the various things. My feeling was that the greatest progress in the industry came when the integrated circuit came along and after that when the microprocessor came. I think the microprocessor was a greater step up than going from discreet to integrated circuits because then all of a sudden you change the whole approach to the electronic industry because all of a sudden you started to get computing power down in price. And one of the first things Intel did was actually in the memory area. They are the ones who finally got the price down below a penny a bit and that knocked core memories out. Core memories kept hanging in there because no one could get below their price - they kept reducing their price - it was about a penny a bit. That rolled over, I was following Ampex at the time. They had moved their core memory assembly plants over to Hong Kong to get low labor [costs], they referred to their “Hong Kong core house.” It was just within a couple of years later that Intel came along with the 004?

Goldstein:

The 4004. It was the 256 DRAM or something.

Everitt:

Right. And off it went. You had step progressions and it was first getting memory down, then it was really getting microprocessors going with the 4004 and then pretty soon a lot of others. Then once you got memory and processing starting to come down, then all of a sudden the whole field of semiconductors broke out. This is a talk that I had given on the development of the semiconductor industry back as late as 1970. The first I.C. major factory sales came in 1962. Up until then it was all discreet industry and you could see it came along there and this sort of steady progression from there. It was a kind of interesting development of how those came along. Then I think really it was processing that took over, it was the key to a lot of further development of the industry for the improvement of processing and putting more and more semiconductors on a chip.

Goldstein:

LSI did that.

Everitt:

To where we’re at the .35 micron level and getting up to five and a half million on a chip and things like that. It was a long time to work those down. That is where you can look at between the Intel talk that Jerry Sanders gave [which] documents well the exact years in which each of these products came along and how they went from under a micron and so on. I can remember the industry whe[n] they were saying we can’t break this because we start to get the wavelength of light [which] is a problem. Then we were talking about X-ray probably all the way down to 0.35 microns now, and they still haven’t had to go to shorter waves to do it.

Freezing Production

Goldstein:

You mentioned the problem that Shockley had, he didn’t freeze production. When you are in an industry that changes as rapidly as semiconductors has, where there are so many incremental improvements and process improvements like you are describing now, can you think of any situations where it is a prudent business decision not to freeze production? Is that a tension for companies in this field where they are developing innovations but they have to sit on them or wait for a little while? How do you manage the timing of that?

Everitt:

I think more is changed to this extent. What they do by design principles they set objectives of a product level they want to be at and they have a project to go from 256 on up.

Goldstein:

So they determine what the market for [a] certain performance level is.

Everitt:

Or they know what the next level is going to be and so they set up a program to achieve that and they may hope to get it done in a year and it may take a year and a half. But I think it is more that type of thing where “what’s the next level?” We saw that with DRAMs particularly, you had one, four, sixteen and on up. It works by progressions. We saw though the speed of innovation which was partly being accelerated because if you look at the number of years, I think it was running for about four years in the DRAMs from one generation to the next, and pretty soon it went to three years between them and then two years between. And the microprocessors have done the same thing where they have been speeding up the rate of innovation. They have learned more and more over time of what the principles are that they have to apply so there is a lot better database of what needs to be done. And let’s face it, from roughly about 1960 you can count on a couple of hands all the really in-depth semiconductor people there were in the industry. Now it has proliferated so it is almost anybody coming out of engineering now is pretty well steeped in semiconductors and the application of them and in computer technology as well. I think part of the speed of innovation has come about by the fact that there are so many more people and so many more well-trained people in the field. I haven’t studied by year, but some of the progress in the auto industry and then the TV industry and various things that came along. Incidentally, stepping back away, one of the first large scale commercial uses of discreet semiconductors was to replace the alternator in car radios and that was really the base where Motorola started to build its semiconductor operation because they were big in car radios and they wanted to get rid of the vibrator.

Goldstein:

Right, the oscillator.

Everitt:

Right the vibrator which created the AC from a battery and they developed those early ones, but those were ones that you used in the hundreds. They have a terrible time of making those at first. In fact it was kind of interesting because they had over 50% hi-beta rejects in those that gave too much gain. And they wanted to find some use for that and so the communications division developed for the large handie-talkie, commercial handie-talkie, a transistor power supply that made use of these high betas right after they got the thing developed and got into production all of a sudden Phoenix called and said, “We found the problem with our high beta rejects, there won’t be anymore,” and they ended up having to redesign their handy talkies for their battery supply in the end. Because the economics to get high betas weren’t such that you could just make those when they were reject from a commercial or consumer level and sometimes you get fall outs like that that occur.

Finance, Management and the Innovation Process

Goldstein:

Well it occurred to me that this whole conversation is predicated on an assumption that may not hold. Let us explore this for a second. Does the financial community try to understand the process of innovation when venture capital or some aspect of the financial world was interested in a semiconductor company that has a promising idea or a product needs developing? Is the innovation process a black box in their calculations?

Everitt:

I think probably their biggest measure, let’s say from the generation after Intel, and then other semiconductor companies that came along the whole group that you have now of smaller niche companies that have come up. In most of the cases, I think that you will find that the venture people that had invested in those were investing as much in the people and the confidence of the people who now have become more known quantities. They were at some place and moved to that place. In other words they could be convinced that these people were able to handle innovation but would come with a business plan; “I see [a] niche in the market and I think we have got the capability to do this.” If Joe Doe who had been buried down in an engineering team somewhere came and said I want to start a semiconductor company and I am going to make XYZ component he would have a lot more difficult time getting money than if one of the senior guys of Intel decided to leave. So I think they do a lot of measuring of people.

Goldstein:

How about management consultants, if you have a small company and they bring in some management consultants, do they make an effort to define the way research should be done? Does anybody understand the innovation process?

Everitt:

There were IEEE papers written on that years ago. It was called “Managing Engineering” in those days. In a sense, “managing innovation” and it’s somewhat nebulous, it’s hard to tell exactly what makes innovation and how it is done. I think people for lack of being able to measure it, measure people more than anything, so their confidence gets put into people because it is not a thing that they can easily measure.

Goldstein:

So, if it is in people, what sort of track record are we looking for? History of developing a commercially successful products?

Everitt:

Yes. Having developed some successful products at other places.

Goldstein:

OK. The problem with that is, the commercial success of a product is frequently a function of a bunch of things that have nothing to do with the technical acumen of the people in charge.

Everitt:

Well yes and no. Somebody let’s say who was the head of the project on the 386 or the 486 of Intel, decides to go off on something. The feeling is “well he was the leader of that group therefore he must be good.”

Goldstein:

I am worried about the reverse situation: where there is a product that isn’t successful and perhaps that’s a problem of marketing or bad suppliers. For some reason the product isn’t successful and it tarnishes the track record of somebody who actually can develop new technology.

Everitt:

Yes, that is a problem and I don’t know specific places where that has happened. But I think because most often the guy that goes off trying to start his own company or a group of people that go off, step off because they have been successful on something. Or they have been successful which includes having been successfully marketed but most of the time I would say the definition of the product, even if you are the product manager particularly of larger semiconductor companies, can’t be blamed on the marketing if it doesn’t work well. Part of the product manager’s job is to understand the needs of the marketplace that he would direct or redirect the design of that product to fit that. Now that doesn’t say he is a marketing guy. I guess, the industry has gotten sophisticated enough now that when they set up a project a lot of them are sort of progressions from one stage to the next or such. Then you have ones that had a group of people and let’s say Linear Devices when they started. One of the key guys was Jay Hoerni who had been one of the early ones in the Intel group and that was a kind of a maverick personality but was given a lot of credit for being innovative. Then several of the other fellows went off with him and they put together a team that I think the investors looked at as having a guy that could manage. I don’t think Jay would have gotten money himself, he was one of these guys that would work thirty-six hours and then not show up for four days, he would be down in Mexico playing. He was a real genius type, a real character. Totally uncontrollable. I don’t think anyone would have invested in him alone but when you look at the team, now I can’t remember all of them--there were five of them that were put together--they were people that others had the confidence to say they can make a team of, and they can tap on some of the innovative capabilities of Jay Hoerni and that is a plus. But they were not investing in Jay alone. I think that is a good example where they were saying these guys defined what they wanted to do in the linear market with real concentration. And then they had on top of that somebody that was known as having been very innovative. I think as time goes by I’ve forgot, Jay did the 4009--the operational amplifier. I think he designed that at Intel, if I am remember[ing correctly]. This is one where time goes by and you forget who was where and what. Then you look at other companies that got started. AMI and some of those kind of never got a good definition of what they were doing and where they were going and they sort have been on and off one.

Signetics, I think, had a lot of very good people and I think that was almost like a Fairchild. Overview of Corning kept throttling what they could do on money flows and projects and things like that. Then they finally sold, just as Fairchild had. Schlumberger seems to do a beautiful job of acquiring companies that disappear into the woodwork and you never hear of them again. I don’t know, I think they often buy companies that are on the way down. In spite of the fact, Schlumberger has been a very important company but I can remember when Schlumberger bought Weston Instruments, who was one of the top ones in electronic instruments: meter movement type of things, and that disappeared under Schlumberger and was never to be heard of again. I think they bought Tung-sol too in the tube business. You have some of those where they become graveyards of different companies. So they were all companies that hadn’t quite gotten the definition as a specialty niche.

Goldstein:

Is there a process by which firms in microelectronics organize to rationalize the research or innovation process? I am trying to be as general I can. I don’t know if a certain company will buy the appropriate firms to develop a complementary suite of development niches or fields of expertise or whether it happens automatically to rationalize this process?

Everitt:

I think it has become a little bit more movement of people in which somebody will say, “We would like to broaden into this area” and they seek good people in the field.

Goldstein:

The brain drain.

Everitt:

Yes. This is the brain movement more than anything and then you get unusual things that happen in the industry.

Leadership and Innovation

And let me just go way back, [to] Motorola when I said they couldn’t deliver for the semiconductors, they had a fellow at that time [who] was heading up with the semiconductor division who was a financial man didn’t know anything about semiconductors, non-technical guy. In fact he was the one that when they found the high beta rejects on the car transistors, notified the head of the product division up at Motorola that “we found our problem that you won’t get them anymore.” The guy said “that son of a bitch, he doesn’t understand we designed the product for it.” Not to pass on too many names, but the product manager became the chairman of Motorola. And that fellow clearly had problems in semiconductors, he was sort of an interim holder of the reigns and Motorola’s semiconductor division particularly in integrated circuits was just going nowhere. They ended up getting Les Hogan who was a Harvard professor.

At the time he went to Motorola, boy, everybody in the industry said, “My God! Are they out of their minds to try to get a professor to straighten out a semiconductor division.” Well, how that confluence came together is hard to explain but Les Hogan did a wonderful job. I think what happened was there had been enough progress in the industry and Motorola had stumbled badly enough, but they were a big enough company that he had the opportunity “to restart the division” in the integrated circuit part without any albatross of past things done wrong. He knew enough about the industry perhaps from having been a professor to see what had been done right in the industry. And he restarted and made Motorola a major semiconductor company. It had been very strong in discreet but in integrated circuits it had failed totally. He did a wonderful job; no question. He was the guy that made the semiconductor division there.

Now interestingly enough when Fairchild got in trouble after Noyce and Moore and those left, they hired Les Hogan to go off to Fairchild, to straighten this out. There was a suit and everything, he took fifteen people with him, all the key people from Motorola. He was not able to fix Fairchild, because they said Fairchild deteriorated to the point it was sort of being held together with baling wire. They were having problems that each product sector was so focused on getting their product out that they didn’t give a damn about the other one. If acid was leaking from their [floor] onto the floor below on somebody’s other product they didn’t care. And Hogan could build from scratch but wasn’t a man who had the same capabilities to fix a sick operation. It very hardly defines what were the differences.

Now he brought Wolf Corrigan with him, who had made the discreet operation at Motorola and he came up to work with him at Fairchild. Unfortunately or fortunately as I understand it, politics came into it and very quickly Corrigan was out for Hogan’s job and he did everything that he could to undercut Hogan. Until finally Hogan got shot down and Corrigan took over. Then Corrigan later went over to LSI and did a great job and [is] still there today. When you talk about innovation, leadership, I don’t know what you call it, it is very hard to define. That is where there is innovation, there is timing , there is luck, all of those things mixed together and what made one work at one time doesn’t make it [always] work. Where there are a little bit of differences I still think it is a lot of people leadership type of thing.

I look back and not just on the semiconductor industry but in the electronics technology field. I think one of the things that made Hewlett-Packard great in the early days was that they were a one-two team. They were both very good engineers, but Hewlett was more the in-depth engineer. Packard was Mr. Outside, Hewlett could have never been Mr. Outside because personality-wise he wasn’t an introvert, but he wasn’t as outgoing and things like that. And they made a great team and they developed into great managers and Hewlett-Packard is the name of success on how you do it. Then you go over to Intel and what made Intel good, I think, was terrific technologists in Moore and Noyce and actually in Andy Grove, and a number of other people but as they developed I think Noyce was one of the outstanding visionaries, if you want to call him that. Moore was a very low key competent guy but all of them were highly respected. Andy Grove was equally good technically I think, but you need a son of a bitch within an operation to drive the troops and that was what Andy could do. I mean he was the whip cracker and things like that. They had totally different personalities but they were complementary personalities. And they put together a team that became very complementary. Then Craig Barrette came along and a bunch of others that have done a very, very good job and they seem to have been able to develop, lose people, replace them, sort of a constant renewal on what they did. In fact, Noyce had worked himself up where he was less active, so when he unfortunately died of a heart attack very suddenly, the company didn’t miss a beat because he had already pulled back enough. I am sure that there are a few products that aren’t there today that would have been if Bob had still been around but they have developed a real team of innovative people. But when you say, “how do you measure innovation?” Boy, that’s like saying “how do you grab a cloud?”

Predicting Innovation

Goldstein:

Well, when people are preparing their spread sheets, or considering their business plan, is innovation treated differently than other instabilities like the political situation or certain economic factors that aren’t easy to predict? Are all of these things equally intangible?

Everitt:

Well, not as much. I think that the innovation side doesn’t come into it, [rather] it almost doesn’t come into it. When you come with a business plan you are talking specifics. Now, whoever invests in the company has to decide: Do I have confidence they can pull it off? Do they have the technical people to do it? Do they have a good financial man? I do think people never tend to mention the financial people and yet I have felt one of the places you will look at right off is “have they got somebody good financially?” And the one I could look back to right off that I knew was Peter Sprague, who was a very sharp guy. When they formed National, the first people he got were Sporck, who was very good as a manufacturing man and John Hughes, whom Sprague knew through Perkin-Elmer as being a very capable financial man. It was interesting that those were the two key people he chose, if you want to call it that, to take over National Semiconductor and build it. He had the confidence that between the two of them they could build other things. But they were the ones that didn’t come with a great new product idea and they developed successfully for many years as being sort of the broad-based company that wasn’t on the leading edge of any one product but good in the five major product lines in semiconductors. Add them all together, and they were a pretty good semiconductor company. Basically, they were a production house because that was the nature of Sporck.

Advance Micro Devices developed primarily in its early phase based on the chutzpah of Jerry Sanders as being a hell of a marketing man and knowing what he was doing. His co-partner in sales at Fairchild was Pierre Lemond, who ended up going off into the investment field, first with Capital Research and now Sequoia. He, incidentally, has financed some of the semiconductor start ups. Again he was a knowledgeable guy in the semiconductor field. And you look at Kleiner Perkins who now has Floyd Kuamme, who was the marketing manager, and a very good one, at National. He was a key person because part of what made National good was knowing and understanding the markets, and then designing devices that fit into those markets. National began to break-up once Sporck left, and Kuamme left. Kuamme is now over at Kleiner Perkins, a major venture to capital firm. That’s again how you measure innovation.

I haven’t studied it, but I think you will probably find that Sequoia and Kleiner Perkins have done a lot more in the semiconductor industry because they had some people that know and understand it, and have the contacts to measure everybody in the industry. If somebody comes, they can find out even the color of the shorts they wear. So I think this becomes an important part of the future investments in the semiconductor industry. You have venture people that know and understand it. Interestingly enough, there are some of the most successful venture capital firms out in Silicon Valley who really haven’t done as much in semiconductors as they have in other things. Institutional Venture Partners is one of the most successful, but I don’t think [it] has any major investments in the semiconductor industry. And the fellow that headed that operation, Reid Dennis, came out of being an electronic analyst and later the head of Firemen’s Fund American Express. He got involved with Ampex in the early days and then another spin off that formed a vacuum pump company, that came out of Varian, and merged with Perkin-Elmer. He then became a large holder of Perkin-Elmer. He was a big one in Seagate and different ones like that. So there are a lot of venture capital people, but I think it just has been because he - and I am using him as an example not necessarily as a single person - he had not done much in semiconductors. He knew there were people who had done a lot in semiconductors, so he didn’t play around in that field.

‘Open Technology’ and Progress

Goldstein:

Silicon Valley is characterized by an “open technology” and you have touched on some of these things. There is this tremendous mobility among personnel, they bring know-how with them. There is also loose intellectual property protection in America. So the point is everybody, all the firms, can share results with each other. Everybody knows what anyone else knows. The economic analysis of that situation, a classic one I think done by Ken Arrow, is in an environment like that will lead to underinvestment. Why would anyone put up the money if the result can be shared by anyone? I wonder if you have observed that and if you think that?

Everitt:

I don’t think that has been a problem. I think what has happened in the field has been that there has been so much opportunity for markets and things like that that a lot of it gets down to investing to certain degrees in a process capability for a product that is just in time to market and too late for another guy to come in and duplicate it. In the early days there was some of this so-called reverse engineering. I don’t think that any reverse engineering of significance goes on now because they know by the time they reverse engineer, it will be obsolete. And so they are not as worried about that anymore.

Goldstein:

But wouldn’t that problem apply to the initial research program that if the life of a product is so limited then you know it won’t be worth developing it in the first place even on your own?

Everitt:

Audio File

MP3 Audio
(268 - everitt - clip 1.mp3)

Yes. Except that is what they have to look at all the time. Management has to look at a potential product and say “what’s it going to cost me to develop and can I recover the cost and make a profit.” Plus the fact that in most sectors, let’s take microprocessors or memory or other things, it’s generally one product that then is obsoleted by the next step up or refinement in that existing product.

Goldstein:

So you needed to do the base work.

Everitt:

The best example is Intel. Andy Grove said time and time again, “We are going to be the one, and our objective is to be the one, that obsoletes our own products.” I think that is the key to it. They should be able to because they are good enough. The only time when you see companies get in trouble is when they rest on their laurels. It is not an industry where you can rest on your laurels, it just keeps going too fast, and I think that’s been the key to most of it and the most successful companies. I think you can also have major new technologies or processes.

MOS and Texas Instruments

We haven’t had any developed for a long time, but a good example when you look back was MOS, as a new process technology. Texas Instruments could not seem to break out in MOS. They had six or seven managers, I think part of the problem there was Bucy (Texas Instruments President), who was heading it up at the time. He wanted progress so fast that he was too impatient. He turned over managers and there was no stability. The first people in MOS, the first initial MOS products came from General Instruments but they were never able to really successfully pull it off and develop a major position.

Goldstein:

It was developed at RCA, right?

Everitt:

Well actually the patent for MOS is at Fairchild.

Goldstein:

Really!

Everitt:

Yes. In fact, that happens to be my college roommate, Tang Sah, a Chinese fellow, who is now the head of solid-state physics at the University of Florida at Gainesville. He was at Fairchild, and was with the original Beckman-Shockley group: came up about six or nine months after the initial team formed Fairchild Semiconductor. So he was not officially a founding member of Fairchild Semiconductor but had worked with that crew. But anyway MOS took a long time to develop.

Goldstein:

In fact, I think it was patented in 1960.

Everitt:

And the interesting thing was that Texas Instruments broke their tail trying that because they knew that was the thing that needed to follow TTL. They stalled out on that. And when you look back at it, Bucy turned over some very good men. L.J. Seven, who started Mostek, was MOS product manager of TI for less than a year.

Goldstein:

I am sorry, who is that?

Everitt:

L.J. Seven, the founder of Mostek.

Goldstein:

Like S-E-V-E-N. I didn’t know that name.

Everitt:

He is now the Seven of Seven-Rosen, the venture capital firm. But interestingly enough he was the founder of Mostek, which was very successful and then went into United Technologies and disappeared. He started Mostek and in my view, he just hadn’t been given time to do a really good job with the MOS operation at TI. But Bucy went through about seven, eight or nine managers in a few years' time, trying to get the MOS process going in Texas Instruments. Here you have what at that time was one of the most successful ones in the industry couldn’t go ahead. And one of the problems that Texas Instruments had is that they never developed anything as successful as the TTL. They were [at] the right time and the right place and should have done much more with the integrated circuit than they did, and particularly the microprocessor; they never got a microprocessor really going that was a winner. So now they have regrouped and are doing pretty well, but they were one that their original orientation in TTL was so directed in the early days at Minuteman. That had been one of the problems. Although their product became the so-called “2N-111” that was in the pocket pagers. And so they did get a lot of commercial products and of course they designed the calculator and some of those things. That’s where again innovation is hard to define. I think maybe what happened was [that] they got into this being a little too muscle bound where everything became projects that had to be cleared by three different levels.

Goldstein:

They had a management initiative, I forget the title of it. It was “Techniques, Objectives and Plans” or something like that.

Everitt:

They got over planned and bogged down a la AT&T. I mean exactly the problem with AT&T was that if you wanted to change a screw on something you had to get it approved by three levels and write up new manuals and etc. I think this is also a problem, and that’s why I think you have been able to find a lot of these niche semiconductor companies like Linear Technologies and the like that do very well because they focused on a niche and have a flexibility in the decision making process. It is at a level where it doesn’t take any time to move and that’s where I think Intel, which I consider a unique semiconductor company, to me that is the Hewlett-Packard of the semiconductor industry. They have been able to get as big as they are and still seem to not bog down the way so many do. But they have had a fairly singular focus and that has been the microprocessor in recent years. Now we are seeing a few companies spinning off or starting to specialize in embedded microprocessors and things like that. But these are special instances.

Money and Innovation

Goldstein:

You were starting to respond to this idea that there really hasn’t been an underinvestment because of the open technology.

Everitt:

I haven’t ever sensed any group with a good idea and good background in recent years that hasn’t been able to get the money they needed.

Goldstein:

It is very natural to believe that there is a direct correspondence between the development of products or innovation and funds available for that development. Is it that simple? Is the rate of innovation a function of the money available?

Everitt:

Well, I think in any business, particularly in electronics and semiconductors, availability of money always limits you. But in controlling or focusing innovation it actually can be good. If there is too much money around, you [can] go in too many different directions. And that was exactly what was happening at Shockley, he wanted to kind of do a little of everything and not focus on single things. I think most of the time in the semiconductor, there is always a limitation of funds. There are always more things that you would want to do but it makes you pick and choose the best. It probably makes it healthier than if you just had limitless money flowing. But the big companies have gotten big enough that they can go to the market and get more money any time they want to but they also have been successful enough that retained earnings have been more than enough for them because most of them don’t pay any dividends and it is all plowed back in. And they don’t pay any dividends because they feel they can get higher return on what they have. Most of them haven’t been in a position where they wanted a lot more money. Almost all of the financing of the bigger semiconductor companies have mostly been European ones and that is more to establish a borrowing base, a stabilization against the dollar by having some. And they are always ventures, they are not equity. And they will borrow over in those countries and that’s part of hedging of the dollar. They will build a plant in France or something like that like down there in Cannes, T.I.’s plant. Well, they financed the physical plant by giving some bonds over there and that type of thing.

I think you would almost have to ask some of the managers in the semiconductor industry but I haven’t sensed that money has been a major problem for them. You had some cycles in the semiconductor industry that haven’t been favorable to bringing new offerings or secondary offerings but most of them have been able to modulate that by the fact that during the periods it is down is when the industry is not contracting but isn’t expanding as fast. So the cycles kind of go with their needs. Now from a development ahead standpoint, that creates some problems that you would like to be investing, you would like to be pouring on the steam at the right times. And now we are finding the problems that the industry has gotten into, that is it costs about one billion dollars for a new fab. And the lead times two to three years on a fab. Right now is an interesting time where you have this huge commitment for new plants out. They are dogging it a little bit, not that the industry is going slow but it is just not going quite the breakneck pace that it was. So they are throwing back. None of them have canceled a plant yet, they have just slid it off by a year or more. But that I don’t think it as much as the lack of money as it is saying, “We don’t want an unutilized plant, so we aren’t going to have another one until we are sure that we need it.” But it also is what creates some of the on-again, off-again in the industry that you go from shortages to surpluses because it’s off. With a lead time of three years on the plant, [it] is awfully hard to modulate your capacity.

Goldstein:

You mean more flexible production facilities.

Everitt:

That’s right.

Government and Japanese Model

Goldstein:

When you hear about the debate about bringing some structure to America’s high tech industries or micro electronics, in the model of MITI in Japan or something, I think the conventional wisdom is that the government has a dismal track record at picking winners and this is why this is not such a great idea. I wonder if you have any ideas about why that is? Is that endemic?

Everitt:

To the government you mean?

Goldstein:

Or to the micro electronics industry?

Everitt:

Well, I think the obvious thing you get no matter what it is, it seems the minute you get too much government in there, you start to get things too structured. People who don’t really understand start making the decisions on how the money will be spent. One of the things that has been good about semiconductor industry has been its own capitalistic nature that has kept it very flexible. On one side you saw MITI in Japan controlling a lot of this and they went through one cycle where they were very successful during which the U.S. cut their capital expenditures way back. The Japanese kept pouring it on and they then were awfully strong when the cycle picked back up. But they got a little over confident and the next time around they ended up deciding that they had to cut back and in this last cycle beginning three or four years ago, Japan cut way back on its rate of expansion and it’s now in a catch-up mode. So I think that’s why the government, the quasi government influence there was a won-one, lost-one type of a thing. But if you look at Japan and innovation, they have not been innovative in semiconductors, they have been production houses. Americans design, develop it, and then will run with the quantity production, that’s where they did so well with the DRAMS. They haven’t done anything with microprocessors to speak of other than clone. And partly cloning with acceptance. They have licensed them.

Goldstein:

Well they have achieved much better production yields, so it was a process innovation.

Everitt:

But production yields in what? If you can get something that’s got a high enough volume, anyone in time can get good yield. So they concentrated on the areas where you could get good yields if you had it. They’ve gotten into the high volume thing. Texas Instruments had extremely high yields on their garden variety germanium transistors (TTLs) in the old days where those things were popping out at 96-97% yields. Well I compare DRAMS for the Japanese in that kind of a category, but I don’t think that they were any better, I just think that it is just where they focused to go for very generic products.

Goldstein:

So they picked up products already ahead of their learning curve.

Everitt:

That’s right. And they became very good at doing that. But I can’t name an innovative product that the Japanese have had in semiconductors, or any process, I don’t know. Have you run into any?

Goldstein:

No, well actually nothing is coming to mind. But in the context of national competitiveness, does that disadvantage them?

Everitt:

I think so, terrifically. In other words, I think we have a major advantage. Microprocessors also are a very good example of where that has been innovation after innovation and they just haven’t done anything. They are always a generation behind in making them, and get some good volume in them but the other thing is beginning with the early days of the microprocessor the real key again I think was with the 386 for Intel was the micro cover. So all of a sudden you had a computer, a true computer, that was doing a lot of processing capabilities but it had a lot of hard wired software. The Japanese hadn’t been good in software, in anything. I think a lot of their problem in software is a little bit their language, their written language. They can’t sit down and program things in any of their language with kanji so they have to use ours and it is like asking the Japanese to speak foreign languages. They have learned how to do it but I don’t think they are good at it as we are. And so I again haven’t seen a software innovation come from them.

Things are kind of fat now because the balance of trade in semiconductors isn’t as skewed as it once was. But in the mid-1980s, I don’t know if it would have been, you probably would have gotten an argument when you said that the fact that the U.S. innovates is good enough. I am saying that when Japan had been down in market share and all these products that they were producing that was considered a serious risk.

I think that was a little bit political. A couple of things about it that I think was important and that was the companies that are producing semiconductors over there are multi-billion dollar companies. They have a lot of products that by our anti-trust laws you cannot do and I think that is one of the things that had to be factored into it. They had high market share, and let’s say DRAMs was the best example, but I’d have matched our profits against them any time. We can’t get to it, but we know in a lot of cases they sold below cost or on extremely narrow margins. They did back in the old TV days, to sell below cost because they viewed winning in the world market, in share of market as key rather than profits as the key. And for their big companies semiconductors were an important factor but not as a sole factor so they were able to accept losses for some long times. And we will never be able to prove it one way or the other, but Motorola proved it a little bit. I’ll tell you one. The guy that gave you my name, Glen Madland of Integrated Circuit Engineering, they did some analysis way back and [were] pretty clearly able to say that there is no way that a Japanese [company] at certain prices can be producing it at a profit.

And I think the other thing was that there was a terrific tendency to be counting share of market as related to dollars of revenue but I am not sure that they were always good measures. Because one of the problems, which clouds things for us too, are the number of semiconductors for which we ended up having major fabrication done abroad, and then it is brought over here to be finished off to get the right percentage to call it of domestic manufacture or foreign manufacture. So much playing with figures that way, some of which are what the government says is the allowable amount and some of them quite frankly I think are the companies play that to the maximum, to minimize their taxes. And so it distorts what’s an import and what is an export and sometimes you count things short on export if the economics is better to make it over there and ship it into Japan from Sri Lanka or something like that. And that therefore doesn’t get counted as a U.S. sale to Japan because it didn’t come out of the United States. I am not sure if some of those figures don’t get very distorted on what they are. But I think the bigger thing is more of a question of what were the profits at the time they had the bigger share of market. I don’t think they had the bigger share of profits.

Goldstein:

So America’s emphasis on innovation was a definite advantage. The counter advantage that people sometimes cite, the advanced market share of the Japanese, is sometimes exaggerated in importance.

Everitt:

Yes. I think exaggerated because it is only counting gross dollars and even those gross dollars may be distorted but I’d argue less that factor than it would be ‘what were the profit on those?’ But I also think in any intermediate term period where they have very large share of market it’s because I think our government has for political reasons not worked on a level playing field. I think for years one of the things that we should have done is tit-for-tat, if you won’t let us sell into your market, well then you can’t sell into ours--dollar for dollar. But we did not do it, well we have never done it and we still haven’t done it. The Japanese do not play fair ball on that and I think that is wrong. Now some of the companies have done a pretty good job over a period of time by establishing operations. Texas Instruments was one of the earliest ones to establish a joint venture in Japan and that got them in much better. But MITI still really allocates what parts of markets will go to the Japanese versus us. That said, I don’t think it has anything to do with innovation, I guess, is the best way to say it. It is totally contrived by how the Japanese decide how to play ball and how hard our government will oppose that.

Now Motorola did a very good job in some of its two way radio and pagers by really facing up to the Japanese, but part of it was they had a strong base in Japan to work their muscle to some extent. But also they were gutsy enough, and I give Bob Galvin credit for that one. They really faced up against the Japanese without really all that much help from our government and they won on [a] few of the things. But I am not sure on some of the semiconductors they have been able to do the same. The Japanese will buy from you if they can’t get as good as a product from their own people - that was the case in pagers. Motorola had better head page, they didn’t have pagers, and so they [went] up into that market and the Japanese would buy it and they are the biggest suppliers to NTT, the Japanese Telephone Company, originally on cellular radios. Now they are coming up. NEC is selling more of the hand-sets over there and I don’t know what the percentages are. In fact that is what they did with DRAMs. When they needed DRAMs in the early days, they bought them pretty well from the U.S. sources, until they could build them and then all of a sudden down it went, but not by any edict you could see. It is just they began buying more from themselves. So always they stay with themselves. They’re the most nationalist country you could find.

Overseas Plants, Globalization and Financing

Goldstein:

You mentioned just before that there are some accounting tricks that people do to skew the numbers for imports and exports. Before that you talked about financing plants overseas. I am wondering whether there are any financial arrangements that you can point to that are new inventions themselves connected with the semiconductor industry that helps explain a success or a failure or even a direction that has been taken?

Everitt:

I don’t think so. Any company plays the tax laws within the law to their best advantage.

Goldstein:

And the industry is irrelevant.

Everitt:

Yes, the industry is irrelevant. Every company does that. For a long time there were a huge number of companies that had plants in Puerto Rico for volume on it. Not so much semiconductor companies but there were a lot of relay companies. In fact Beckman Instruments made a number of their instruments there because they got a favorable tax treatment. That was the sole reason they were down there. It was crappy labor. But they lived with it. Now in foreign countries we do ones of sending things down for assembly work in Mexico where they will have some plants. They have to have a certain amount of U.S. content to be able to get back into the country, in other words send them down, bring them back without paying an import duty. And then they can send them out. Other things that they can send down and if they have too much assembly done, they can be sent directly to customers abroad. That won’t get counted as an U.S. import or export on the thing. And the best example for quite a while you were having a lot of semiconductor processing being done in the States, you would send the wafer over to Kuala Lumpur and Singapore, where a number of them made semiconductors. They would slice apart the wafer, and assemble them and that was the biggest part that went into it. Assemble them into the cans and then generally they were bringing them back over here to do testing on the things. Some of the how-much-they-did-where, depended on trying to minimize the labor but also minimize the taxes and if you had some tax advantages by earning enough of your profit over there or a tax advantage by completing it over there. Let’s say ship it to Japan or somewhere else, maybe even to Europe - those all come into the thinking, so tax laws and import-export laws create a lot of processing systems that you probably wouldn’t do if you had none of those barriers.

Goldstein:

You know it is strange though because there is this vast globalization of the semiconductor industry, in particular, maybe industry in general. I wonder if every aspect of the business is keeping pace, maybe the corporate culture isn’t globalizing at the same rate which effects the development of a company.

Everitt:

I don’t know. I don’t think any more than any other type of thing because generally the corporate culture is pretty much U.S. based to an extent whether they are dealing in Europe or things like that. The product, even if it’s a process, is done according to U.S. standards whether you do it in Cannes, France or in Frankfurt, Germany. A lot of that is located for national purposes to be in the Common Market or to be in Mexico to get the labor or Kuala Lumpur for labor. You can tell they are in the Far East and in Mexico for labor advantage. They are in Germany for nationalistic [reasons] because then that becomes not an import to Germany but you get treated like a national. And yet the culture is pretty much the same with the exception of the sales. And the sales and the way you work with your employees, it has to be different because they are [a] different culture in Germany so they almost always have a national heading up the operation wherever they are. They have found that shipping an American over to run the plant doesn’t work too well. They’ll ship an American to teach, and the other fellow we want to have it done but they let him work with the people because over there you can pat a girl on the fanny and you won’t get in trouble. Now all of a sudden over here everybody works differently. I saw that once when I was on a trip over to Europe where we were looking at a lot of electronic companies and in fact visited T I’s Cannes plant. But we heard talk that they used Paris as the base. I remember hearing presentations by the manager of the plant operation of Beckman Instruments in Italy, and in Germany and in France. And what a difference in hearing those people, the German guy was “click-click-click” “it is done this way,” and the French and the Italian guys were real laid back, and you thought how could this be out of the same company? But it was because they were having to fit with the cultures of where they were. I don’t think semiconductor industry is any different in that sense.

The other thing is the semiconductor industry is becoming less and less dependent over time on the labor content factor, because they are getting more and more automated all the time. It wasn’t too many years ago where on an integrated circuit you had forty wires that were wire-bonded across to the outside package and now they are solder-blobs that are flipped over, heated in an oven, and made a connection to the header or the outside package. So it has taken a lot of labor out of things that it used to have. I think this is probably where the manufacturing culture changes because the needs have changed. They don’t have to send it to the Far East as much for that final assembly type of work where it used to be a lot of labor was put into the package. It was really the big thing. They tended historically to keep the testing over here because they wanted to make sure that it all worked. But now the testers are getting so automatically programmed there is no reason they can’t have a Terradyne tester on site in Kuala Lumpur. As long as they have good management supervision to make sure people are doing it right, they don’t have the same needs to bring the things back.

Military Systems

However, if it is going to be built for the military over here you can’t ever go out of the country. Again our big government says that nothing built for the military is going to [be] built [with foreign-made components].

Goldstein:

That means that military systems can’t be built with off the shelf components?

Everitt:

Exactly. In fact, fifteen years ago there was a big brou-ha-ha and when you read in the newspapers it sounded like Texas Instruments had done something awful in not documenting some of its components. But what had turned out was they were replacement components for the Minuteman or one of the military systems in which that component had progressed to the point that it had become commercial. They made five version revisions to improve it. Not electrically, but process-wise and it hadn’t been adequately documented and the government said, “You put a non-approved, non-documented transistor in.” Well, the problem that came up was that there weren’t any of the old ones available anymore, they don’t make them that way anymore and to go through the government equivalent approval thing over a period of time and the government had done this, had sort of looked the other way and said, “Well the 2N-111 and the military export 35 are really the same product. So you can substitute one for the other.” But then later on someone from the GAO looked into it and said, “Hey, they aren’t equivalent, that 2N-111 has had five revisions on it. So what you said was equivalent is no longer equivalent.” And so that’s where the documentation was kind of nobody’s fault but that’s the way it happened.

Goldstein:

I have seen interesting situations, just to be hypothetical if it ever happens. If a company develops a device or some unit that they believe has a strong commercial market, they might lock it up by using it in a military system. I am suggesting a situation where they are not able to transfer it out to the broader market?

Everitt:

I don’t think there are too many devices that the military says you cannot have a commercial equivalent. Generally what they do is they give them two part numbers. They have learned now if they do some improvements in processing on the one they carefully document it, the military says, “That’s equivalent, that’s equivalent,” or else it is not equivalent and will always have to go this way.

Goldstein:

That was just a digression.

Everitt:

Those were just examples that I gave that are the problems with government and again you had the GAO, I think was the one who blew the whistle on this thing. That made it sound like Texas Instruments was cheating but they were not at all. The GAO guys are accountants, they push pencils - they do not know anything about semiconductors and that’s the thing. Documentation is not there, these aren’t equivalent, they are trying to cheat the government by not giving them the ten year old device. I think they have gotten away from this, in fact there is more moves of late [in] the military. You can say, half the cost of the military is that we are insisting on putting mil spec things in, in a non-mil spec unit to pass all those mil specs, why don’t we do this on the thing? It has just been, I think in the last year or two, some moves to allow plastic encapsulated semiconductors in the military system, up until then everything had to be in metal-welded cans, which is a terrible expense. Not necessarily as resistant to aging and corrosion and things like that. That’s just getting recognition for those things. I think one of the things that has been unique to the semiconductor industry versus anything that we had before in electronics is you are down to so much in so small a place that it is hard to separate processing. It used to be you could count the components and you could tell what exactly you had and now it gets pretty difficult. So defining the thing gets very different. Before semiconductors came along, and particularly integrated circuits, I think you could take any military piece of hardware and you could define it by a parts list and now there is no parts list other than to say it’s an Intel 90XXX microprocessor in there, and you can’t really define the parts on it.

Judging Company Prospects

Goldstein:

In the work that you do now, what is the foundation for differences of opinion? What are the bases for judgment about the prospects of companies?

Everitt:

Well, I think obviously looking at the technology. Up close, it is pretty hard for a person in the financial field to define that this technology is better than that technology. Again we use people who know the technology and understood it better. So we talked to them about it. Back when I was very active and fully following the semiconductor industry, I was constantly visiting the companies and talking to the people. You learned as much about who had what that was innovative, and who the people were that were good and such by talking to them. When I was concentrating on the semiconductor industry as an analyst this was not easy but a lot easier. Now as a fund manager where I am investing in all sorts of other things, I don’t have the time to do that. So I have to turn a lot to the so-called street research of people that are doing what I used to do and see what they have to say, and then taking several I think are good and interpolating between them. But then if you get down to a very new company I do the same old thing of looking at who are the people, where they come from. Then before I make a major investment in them I want to get some conviction by talking to a few people in the field, so I do some testing of the water. It is not a seat-of-the-pants analysis, it is a much more simplistic analysis. But an awful lot of it comes down to looking at the markets and the markets served.

Of the larger companies, I have been very high on Intel for quite a few years here, because I have really felt that the microprocessor was the key to success and that they have really gotten far enough ahead. I don’t think anyone is going to catch them. They could stumble a little bit here or there but they are still moving ahead at a very pretty steady pace. And then I look at Motorola. It has a good broad line of semiconductors but investing in Motorola I look at as that’s a good strong growth area and they will participate at the rate of the industry but I am investing in Motorola a lot more because of its two-way and cellular capabilities and then I am on the semiconductors. That has now become less than one-third of the business, it is now a quarter of their business. On National Semiconductor they are broad line, I sort of lost interest in it after Sporck left because I didn’t see them going in any direction. AMD I have been less interested in because I have seen...

Goldstein:

Advanced Micro Devices?

Everitt:

Advanced Micro Devices. I still have a great respect for Jerry Sanders, for his savvy and his understanding of markets, but he has fallen into following Intel and he is one generation behind all the time in microprocessing. I think he made some strategic errors by trying to go after that part of the market so much, I think he burned some bridges himself by getting a licensing arrangement with a 386, and then trying to underprice to get a position in the marketplace. He got Intel mad, and Intel says, OK, fine, we have the agreement on the 386, but we aren’t giving you the micro code on the 486. He spent all his time arguing in court and things like that. Even though they have finally cleared up some of this, I just think AMD who was pretty innovative in the early stages, turned into more of a follower than a leader. Then I think in the other semiconductor companies it has been more the niches and who are good in that. Because I am doing broader investments I don’t feel as capable as I used to be to evaluate those well, so the only ones I will invest in are the ones where I see some unique difference in what they are doing and some signs of success in the market place.

I have been looking but haven’t invested yet in the some of the companies that are doing DSP (Digital Signal Processing) because I think that is a major sector for the market growth mostly in telecommunications and such. I think that is a good market and so I am looking. I have been following the DSP communications and other companies like that but I haven’t invested in it yet. I have found that it has been more difficult to come up with new exciting ones. In certain ways there is a maturity developing in the semiconductor industry that you have less new guys coming out of the woodwork who are dynamic and exciting, I think you can still invest in a lot of companies in the semiconductor industry per se. I think it was one of the above average growth sectors of the electronics industry. So from that standpoint it is good, but I think it is harder and harder to pull a new one out of a hat. Someone may come along [with] something that is very unique, but I just don’t see it. It is kind of interesting. I was [a] very strong investor or promoter, I shouldn’t say promoter, but I was very enthusiastic as an analyst when Mostek got started. DRAMs was their big thing and they really focused on that and did very well. But then you could see that the industry reached a point where all of a sudden it was becoming a commodity product. And the Japanese were starting to get into it, and I said I am not going to invest with anybody in DRAMs because I think it is going to be a commodity business. And that’s exactly what it has become. And it doesn’t say that people can’t get volume but the one that did some in that has been Texas Instruments because of some of their plant relationships in Japan. Still right at the moment, if you look at their last quarter, what they point [at] as being the weak spot, it was pricing of DRAMs. I was always extremely cautious about Micron because their total concentration was in DRAMS.

Now, flash memory has been an interesting area and that is where Intel started to do some good things. I think it is a good product line for them but you haven’t had any company that’s come way out in front. I have been sort of watching San Disk, as an interesting one in the flash memory field. I think that is a market that is going to be big.

Goldstein:

Now when you say watching them, what are you watching? Do they have products at market already?

Everitt:

Yes, they do have them. And I have been watching their figures and I got on their mailing lists, and I started following them from that standpoint. Now I am not doing as much investing in those areas because I am mostly working with pension funds for institutions and managing their accounts. It is not an aggressive growth area. Plus the fact that quite frankly there are a lot of areas where there is aggressive growth in telecommunications with some areas that are now more exciting to me than semiconductors. But if I look at the portfolio of companies I have had in a couple of accounts that I have handled that are aggressive growth, I have Intel in there, I had Motorola, but backed out of it because they have had a short period of transition. My guess is that I will be back in there by fall. That was just sort of an adjustment in the cellular business that sort of overran itself for a short time, now that wasn’t because of the semiconductors. Those are the only semiconductor companies that I have at the moment because I can find a zillion in the telecommunications area all the way from what used to be little, the Sysco Systems that are now billion dollar companies, to ones that are a lot smaller than that, and then applications of semiconductors or microprocessors. I see companies. You get to sort one through another, Cognex Systems is visions systems and they are doing extremely well for production line measurement of everything from the color of a label to the position of a component going in. They have acquired another outfit, which is doing some surface measurements in the semiconductor industry. I should mention I have had Applied Materials a lot, but that is one which is serving the semiconductor industry in the processing area and they are the leader in that. And then KLA Instruments has done a lot in semiconductors. So, peripheral to the semiconductor industry there have been some interesting ones too.

Analyst Changes

Goldstein:

Has the way that you conduct your business changed over time? You say that you used to do some research and now you hire people to do that. But that has more to do with your career trajectory than the industry in general?

Everitt:

Yes.

Goldstein:

Have the analysts kept the same position? Is there more outsourcing of technical expertise?

Everitt:

Well a couple of things have occurred. There are a lot more people in the industry but I also think the street researchers have not done as much of the in-depth analysis of technologies as they used to. It has gotten to be a little bit too much in my view, what to buy and what to sell. Rather, I used to follow an industry and then out of that came buys and sells. Now they look for buys and sells. Then of course I would have to admit that my view is different. When I started as an analyst in 1959, there were three of us in the whole country that had industry experience, engineering and an M.B.A. as a combination. Now they are a dime a dozen and so I would say we had some unique capabilities for over a decade. There just weren’t that many of us that could go in and talk engineering terminology with people and understand what they were doing. One of the big advantages that I had in the semiconductor industry when I began following it as an analyst was that I had worked with a lot of people as suppliers, so I knew the regional managers of Texas Instruments and a lot of the people down at T.I. All of a sudden you were kind of one of the family.

Goldstein:

Maybe the reason why you put such stock in your take of the people’s capabilities is because you are connected in that way.

Everitt:

Yes. That has been my connection of knowing. It was the way I followed the electronics industry too. I found that I got an acceptance in those early days when I went into a Hewlett-Packard or a Tektronix or an Ampex because I knew a lot of the terminology of the industry. Where in the early days their biggest problem was that they have an analyst come in from a bank and they were happy to spend most of the day explaining to him what a semiconductor was. And in fact as an analyst that was doing so-called street research. Basically our customers in those days were the large institutional investors, all the big banks, all the big mutual funds, the insurance companies, who had their own research departments but they looked to us. A lot of my visits to those companies or to those customers was giving a little bit of a tutorial on what a semiconductor was. I had charts and I explained a wafer and how it is built and how the mask overlays. That is where I ended up with some of those things on the wall. I had a whole drawer full of the glass plates on how you put a semiconductor on and different parts to show people how it worked inside because so few people knew about it. Now I think most people who would follow the technology part, they understand all pretty well. So I think there may be less need for it now.

By the same token I turned to the analysts in the networking field to explain all the nuances in networking because I am not that close to it to understand. I am luckily, even though I have been away from engineering for a long time, I am close enough to it that I can understand if they get any technical jargon. I can understand it but I still had to learn all about the difference between frame relay and ATM and this and that. So I think that is always the way in an industry where you follow it. When you are the guy, specializing in following it, you get to a lot more depth. Plus the fact, I visited on-site the semiconductor companies a couple of times a year. And I got to know everybody from the president on down of those semiconductor companies in the early days because they were small enough. They were actually happy to have somebody that could understand the industry enough that hoped to tell other people about it. So they almost looked at you as a friend not an enemy. I actually found when they came out with new products and such I would get one of the earliest call responses. If I called them to find out about a new announcement they would call me because they knew I would be able to transmit that to fifty clients who wouldn’t then have to call them. And so you became part of their pipeline to help get it out. And yet I was always in a position where I was not promoting their product, I would explain the product.

Investor Perspective

Goldstein:

When I read histories of microelectronics the story usually focuses on what’s happening in the laboratory, sometimes on the larger corporate context but rarely does it focus on the side of things that you might have seen. I wonder if you believe that’s the appropriate emphasis or whether the story really can’t be understood without knowing what the investor community thought of a product or development.

Everitt:

I think, too many people think you invest in a company because of a product or something like that. It is a new whiz bang thing they create: it isn’t. It is a continuing flow. My feeling always from day one was to evaluate the people, evaluate the research they seem to be having and then try to make a judgment. Have they the capability ongoing to keep reinventing new things that are going to make this into a big company? Or as I saw in Transitron the gold bonded-diode they were good at, but if you wanted to call it, David Bachalar stole it. He walked out with the process from Bell Labs and they knew how to build it but he never developed a group that could come up with anything else. And I could see that and I wasn’t that brilliant a guy, but it was because I got there and I saw, I talked with other people in the field, and when I talked to some of their engineers when I asked, “where are you going next?” You kind of got broad blabber. They didn’t say “we are going from here to here.” When you talked to other semiconductor companies they didn’t necessarily tell you what products exactly they were doing but they had a road map of where they were going. I think people tend to think that the financial community does it by just looking at products. There is a little of that [going] on, [but] I think that is more hype by the newspapers or a Netscape comes out.

Goldstein:

That adjusts quickly.

Everitt:

My feeling, I have never bought a company for a single product.

Lag, Diversions and Financing

Goldstein:

You mentioned that there was a serious lag in introducing MOS technology. CCDs sat for a long time before they found widespread commercial use, I wonder if there were any cases where a lag like that is explained [by] something on the financial side.

Everitt:

I think at MOS it was just plain tough technology to learn. It was a step function for the industry. There were a lot of products that wanted it. I know of several companies that failed because they thought they had it. Big companies like T.I. wouldn’t deliver because they were a little too early in the stage of being able to make everyone’s product. So I think that was an easy sector to point to that had more of them that had been a little more evolutionary than that. Now what was the other question that you had? CCD never found its right place for what it could do. They are still used in places but not in big markets. There are others limited markets like night vision which is mostly for military and things like that--very expensive. It has to be super cool, have restrictions, but I think that most of the time it is again perceiving where our market needs are. And a good example [of[ where you have seen the industry get off, get diverted if you want to call it that. Intel got off into watches when they were real whiz bang and kind of exciting looking. They came very close to kind of getting a big focus and emphasis over in some consumer products but they quickly learned it wasn’t the technology as much as it was that was going to become a commodity and that wasn’t their cup of tea. They were wise enough to write it off and get out. T.I. did the same thing in watches, they reached a certain point, it was a commodity that wasn’t their cup of tea. They have continued in calculators and it’s an OK business but nothing exciting. My guess is if they had their druthers they would look over the years and say they would rather not have been in it. Hewlett-Packard did a marvelous job in calculators in a particular niche, the scientific calculator. They have continued to make a very nice business out of that. National Semiconductor got into clocks and watches. Fairchild had the very first TV game called Pong. I found that the semiconductor industry had to go through this, each one of them. They didn’t give up the other things they were doing but they started to focus some of their development talent and management talent for some consumer products. Luckily, fairly early on after a couple of years each began to conclude these were going to become commodity things, that it was a marketing game and not marketing the semiconductors guys were good at. Each one of those companies, T.I., National, A&B never got in, Intel never got in, but that was because they were beyond that stage. Well, Intel did. So even the best of companies have their little periods of... At that time they thought they were whiz banging things and they quickly said, “Oh, I know was on the watches.” Really the killer on the watch was the LCD. It turned out that the Japanese were a lot better at LCDs than we were. And the LCD became the expensive thing. At first they had the Nixie diodes. They realized that they just ate up batteries too fast. So that was one they hoped to accelerate in getting out or in getting the technology from anybody else. It is just interesting; none of these are perfect they all have their getting off into some various. Luckily it didn’t let them kill the company.

Management Style

Goldstein:

It’s common to associate some of the microelectronic companies with a unique management style. You mentioned before the way Hewlett-Packard and Intel are run, they are really stripped down, really basic there is not a lot of management overhead. Could that be called an invention of the industry?

Everitt:

Yes. Maybe Hewlett-Packard was the one that was the model in the Silicon Valley before it was called Silicon Valley on a management style. Hewlett-Packard never had big closed offices, they worked in open areas, a little of the California relaxed atmosphere of things. I think maybe that is one of the reasons the whole semiconductor industry developed that way. One of the problems with Fairchild was that corporate that was in Syosset, New York, was putting a little of this management overhead “we’ll dole out the money to you for this project and that project,” that was affecting their flexibility. And when Intel spun off there’s that free flow thing. To this day Andy Grove doesn’t have an office much bigger than this. The industry has not been one of pomp and circumstance. Probably the biggest one would be Jerry Sanders because he is a salesman at heart. I think it has created their flexibility.

I have a nephew that spent time over in Japan as a visiting professor at Fujitsu for a year. He has a group in computer science and LSI designs at Carnegie-Mellon. He had a very interesting observation also on the innovation of the Japanese. Things were so structured and they took the most innovative people and if you didn’t move up by the time you were thirty to a management level you were never going to go anywhere. And so they would push to move up that way so they were taking the guys to what he thinks are the most innovative years, in the 30-35-38 period. They were being pulled out of research all of the time. And also the Japanese had another thing, you got on the track in a Japanese company so you never got out of track to go and get an advanced degree. By the nature of their structure they discouraged advanced degrees. He feels this is a major shortcoming of the Japanese. Incidentally, on this innovation idea, you ought to interview him at Carnegie because he has some of the most first-hand comparisons of anybody that has been able to go over there and live for a year and see how the Japanese think. I think he has gotten some pretty interesting insights that way.

Goldstein:

Do you think the myth exists in an innovative semiconductor company needs this sort of a new California management style? Does that myth have any power?

Everitt:

I think it does. It keeps this from getting too much structuring. Look at Intel now, that’s [a] multi-billion dollar company but it [has] still kept that informal flow and I think Advanced Micro Devices has [it], Texas Instruments stumbled because of that problem of getting a little too over planned.

Goldstein:

I remember now. That was OST (Operations Strategies Tactics).

Everitt:

And it started being much more like a big company .The guys at Mostek all came out of T.I. but they had that simplistic free-flow management. And they developed some awfully good people. In fact Bob Palmer who is now the president of Digital Equipment was the manufacturing manager of Mostek, but out of that came T.I. I think there is a lot to be said for that. I don’t know if I want to put the label California on it because a number of companies have gone down to parts of Texas but that is not their headquarters. Their headquarters is Silicon Valley where the greater group is located.

Goldstein:

It is probably a difficult question to figure out whether the companies’ liveliness depends on a management style that way, but it is a different question and equally interesting whether the belief that that is true influences the kind of support that a company can receive.

Everitt:

It could be, but I don’t think it is much a belief as it is a fact. It is like a new generation of technology companies versus the old--the GE’s and the Westinghouses. It has been kind of interesting that for what reasons it has been hard to define but you haven’t had semiconductor company in the east match one in the west. Plus it is also affected by one other thing and that is once you got the big enough group out there in California, they get embued with that lifestyle and you can’t get them to move east. That may have some effect on it. But I think it has been fairly important. But you have seen other companies do equally well with this casual atmosphere. Look at Seagate that now almost owns the hard disk field. They are out there in Scotts Valley, and have the same kind of relaxed management style. So it is a little bit a west coast versus east coast type of thing. However, the mini computer industry developed very well here in the East, up around the Boston area. It got awfully structured and I think [that’s the cause of] some of its problems. Plus the microprocessor ate away at them. And I think that is the biggest problem they had, the microprocessors began to get to them and it may have been the flexibility of Intel were greater than the flexibility of others. That’s why DEC is doing well now – they’ve got a west coast semiconductor guy.

Goldstein:

It is an interesting thought - there could have been a critical moment the future could have gone either way.

Everitt:

I think that is one where Ken Olsen was just too structured, he couldn’t change, he couldn’t think differently. The very guy who made the company almost ruined it. And they have gone through a real metamorphosis. And it is interesting that Bob Palmer who was heading up the semiconductor group of DEC when he was made president came out of the culture of the semiconductor industry. While he was never in Silicon Valley a lot of that culture had gotten down to Mostek and that type of thing.

Coasts and Unionization

I think the other thing that is important is in the old electronics industry, the big companies developed on the east coast--the Westinghouses and the GE’s--and they became highly unionized companies. The consumer product industry--the TVs--were a mix, the Admirals and those were highly unionized, that is not a professional but the level below. But it made them have to manage everything in [a] much more structured and much more defined way because you couldn’t make a change without getting the approval of the union. The whole west coast electronics group had grown up without unionization and I think that has been very important to them, and I think that’s why. Hughes was an early one in the semiconductors but I think because Hughes was one partly from a government thing, was very unionized in structure and things like that. They were the one big west coast one [that] looked good in the early days and sort of faded away.

Goldstein:

Is that sort of a pet theory of yours or is that widely understood?

Everitt:

That’s my pet theory, I guess. It is hard to say whether it is widely understood. I became more conscious of this perhaps because I started at Motorola in Chicago and in the communications group, Motorola was non union all the way. My first rub with the other part of it was that we were the main supplier of mobile telephones for the phone company, so I had to go over to the Hawthorne works in Chicago where other parts of AT&T operations were. And I saw the way they operated and it was so structured it was unbelievable, if they changed a screw on a unit they had to rewrite pages of manuals, get it approved by the union for a change. They had wire harnessing for other things but it was the same as we had. We were getting three and four times the productivity. So the flexibility was important that it builds up in a non union atmosphere because there were a lot of changes that had to be made just continually in the semiconductor industry. If they had been unionized, I think they would have had a terrible time. Now I don’t know about the Japanese. There are a lot of unions over there. Whether what level they are at Fujitsu, but I still think that they are terribly structured because they were great big companies that got into semiconductors just as T.I. for its day with a 200 million dollar company went in and got into semiconductors. So it started out being a little too structured and that might have been one of the reasons T.I. went through some metamorphosis and some problems because they started with that, where the other guys all started from a gang of guys growing up. Sort of top down versus bottom up. Does that cover most of it?

Goldstein:

Yes.