Interview with Gerry Parker
October 6, 2003
Los Gatos, California

Gerry Parker joined Intel as employ number ninety-nine in 1969. After thirty-two years with the company he retired in 2001 as Executive Vice President of Intel Corporation and General Manager of the new business group. In his career at Intel he defined manufacturing processes and organization, including the all-important introduction of new semiconductor processes in the manufacturing until moving to the new business group in 1999. In this 2003 interview Gerry describes the explosive growth of Intel and the unending challenges of transferring technology from R&D to production.


RW: Well, Gerry, tell us about your early days, your family, and so on.

GP: My early days. Well, do you want real early? I was born in '43 in Germany and we came to the U.S. in 1952. Didn't speak a word of English and my stepfather threw me into a Richmond, Virginia Elementary School, me speaking German and everybody else speaking English, and so that took a while, like two weeks to learn English. And then we actually went back to Germany for a while as part of the Army, and I wound up in California in 1960 and finished high school - Monterey High School, and then went on to Cal Tech for eight years. At the end of that it was off to Intel where I stayed thirty-two years.

RW: Well, let's -

GP: That's the - that's the snapshot.

RW: Let's - let's go back. You were - you were born in Germany. Now what were - what were your mom and dad like?

GP: Never knew my dad, so my mother married a GI in 1950, so I had a stepfather from then on.

RW: And you spoke German because you lived in Germany.

GP: I was born in Germany, lived in Germany and spoke German. You got that.

RW: So did you have brothers and sisters?

GP: No. Only child. So - my mom is here still. My stepfather died some time ago. And so she lives in Napa and we've of course been in California since the early '60s.

RW: So what were - what were your dreams as a - as a child?

GP: Oh, I always enjoyed engineering. I mean, the nice thing about Germany is they had a lot of trains and streetcars and things like that and I remember always thinking, "Boy, I'd sure like to be an engineer or drive a train or design a train or build tracks," or something, so I kind of loved engineering from ever since I can remember.

RW: So at Cal Tech, did you know Carver Mead there?

GP: Yeah. I started Cal Tech in '61 and - and was pretty much through the meat grinder with the rest of their freshman and sophomores, and I didn't - didn't meat Carver - I - I think I met him in - in senior year, which would have been about 1965, and at that time I applied for Grad School to get an MS at Cal Tech and - and Carver was - was my advisor - was going to be my advisor, and so I really started working with Carver during that Masters year. And then of course I applied for and stayed on for a Ph.D. and Carver was my thesis advisor.

RW: So what - what did you think of the guy?

GP: Oh, I loved Carver. Carver was - was great. I mean Carver was a young professor that was working in an exciting new field - semiconductors or diodes actually at the time, and Carver was a great teacher and really an inspiration. You know, quite frankly, because Cal Tech is an old institution and so you have the math and the physics and you have a lot of really wonderful, great people, Nobel Laureats and so forth, but Carver was kind of a - a - a fresh, young face that really knew how to motivate and was interested in students. He was interested in teaching. He was interested in - in doing research, and so you didn't have to stand in line to try to become a student at one of the - one of the big name guys. And I really wasn't interested in - in math or physics per say. I was interested in electronics and semiconductors. Carver really got me interested in semiconductor. I was more interested in electricity and electronics early on. And so I - I stayed with Carver, and Carver was just a great mentor for me.

RW: Now this was - was this before his VLSI days where he came up with silicon compilers and -

GP: Oh yeah.

RW: - that sort of thing?

GP: Carver - Carver started out basically looking at Shockley diodes. I mean Carver was in materials. In fact most of my thesis work was - was on Shockley diodes. We were putting metal on zinc sulfide on - I was trying to put it on sulfur. We were trying to put it on all kinds of different compounds to - to measure conductivity and to measure, you know, time of flight and - and barrier heights and so forth. So that was most of the work, and so during the time I was with Carver, we were working on tunneling Shockley diodes, very much diode type things, and it wasn't until after I left Cal Tech that Carver got interested in - in doing really integrated circuits. He started off with his - his main clock chips I think trying to make clocks by using essentially, gate arrays.

RW: So you were recruited to Intel. What - what was the story there?

GP: Well, I - during my years at Grad School I tried to get summer jobs. And so really the first - the first summer job I got was with Mobil Oil in Texas fixing their - their sound recording - or their recording instruments out in - in Pecos, where you put all these geophones out on the ground and that were the old TI analogue recording systems, and I was doing a repair job just fixing transistors and - and unsoldering components. Then the following year, I think the following year I got a job at Hewlett-Packard and I worked at HP Associates that summer. And then the following year I worked at Barium and the - the last year I think I worked at Fairchild. And that's where I got to meet Andy Grove and Gordon Moore, and so then about the time I was getting ready to graduate, Intel was founded. And so - so I think I still owe most of this to Carver because Carver kind of made the introductions, and I got interviewed by Fairchild, and I got interviewed by Intel, by Gordon, and was offered jobs and, of course, I took the Intel job. But what's - what's a bit funny about that is I was - was also offered a job at Fairchild and - and I think the Fairchild HR manager was Bob Palmer?

RW: Mm-mmm.

GP: That sounds - Palmer. Don Palmer, I think. And I remember I got the offer. I decided to walk next door to see Palmer, and his first question was "did they offer you a job." And I said, "Yes." He said, "Take it." And so - I said, "Okay. Well, that's good advice." And so I took the Intel job.

RW: Now what - you - you were very early on. What year was that?

GP: 1969. So my employee number was ninety-nine. So I got in just under the hundred cut-off. And actually was working - Carver - Carver Mead and Jim McCaldon had worked on a light emitter process that they - they had sold to Intel basically. But Gordon - Gordon bought the process, and they had set up a laboratory in Pasadena. And so I was actually hired by Intel, Gordon, Carver, to work in that laboratory. And I started there in March of '69, and while I was still actually working on my Ph.D., and stayed there until the end of - end of '69. And when you talk to Ted Jenkins, Jenkins was also involved, so Ted and I were actually in that lab. Ted started it and then I took it over when Ted came up. And then at the end of '69 we moved everything up to Mountain View at the time and shut down the Pasadena lab. And shortly thereafter Gordon and Bob decided to sell the light emitter process to Monsanto. And at the time I actually had a choice of either staying with Intel or going with the light emitter process to Monsanto, and I think I made the right choice. I stayed with Intel. But - so that was really the first year - almost the first year of work for Intel was working on light emitters.

RW: Employee ninety-nine at Intel. Now did you get stock options at that time?

GP: Oh yeah.

RW: Oh, that's nice.

GP: Yeah. Yeah.

RW: So what - what did you do then in Mountain View?

GP: Well, at the time they were having major problems with reliability. We had designed - Intel had designed Shockley diodes into the bipolar circuits and the Shockleys - I remember they were called collapsing Shockleys because they would work for a little while and then all of a sudden their reverse current wouldn't - wouldn't hold up and it - it would collapse. And so Gordon assigned me to really just fix that problem. And it turned out the problem was there was some cleaning in the - in the wafer fab and it took a while to track that down but - so I kind of got into more trouble shooting type stuff, working on reliability problems; why the MLS wasn't working, why the bipolar wasn't working properly. And I stayed with that for quite a few years and eventually wound up running quality assurance, but then by 1977 I decided I really wanted to get back into development and, you know, R&D type of stuff. And so then I - then Gordon reassigned me to basically run the technology development - operation of '77.

RW: Well, because of the Fairchild experience of - of having an R&D separate from manufacturing which caused all sorts of transfer problems, I think Gordon had a philosophy that R&D would be imbedded. Could you describe that?

GP: Well, that was actually, as you say, that - that was very much the thought process, is that having a separate R&D created a lot of problems in terms of transferring technology, and so the - the concept was we would just do the R&D on the manufacturing line. And there are some real benefits and some real detriments of that. And so for many years, the early years it actually worked - it worked quite well, because if you think - think about it that you - you really had everything already in manufacturing and you just modified it and so forth. And about the time we had the transition from N-MOS to C-MOS is when I think that model fell apart, because the C-MOS was much more sensitive to contamination, much more sensitive. And I remember arguing with the fab managers that your - your furnaces are too dirty. We can't do the process. And the fab manager would say, "Too bad. I'm running manufacturing. Get out." And so our development I think almost ground to a halt in terms of transition to C-MOS because you just couldn't make that dramatic a change on the manufacturing line without somebody getting into a fight. And so we struggled through that. And then the bureaucracy kind of set in to where you had to negotiate every step with the manufacturing manager and you had to then transfer - create documents, specs, and so forth, and to transfer the - I think it was the one-and-a-half micron technology so that was quite a while ago, to transfer that into manufacturing, we had to collect fourteen thousand signatures. Because every spec, every part of the process had to be specified. Every one of those specs needed about twenty signatures from the development manager to the fab manager to the training manager to the safety - you know, on and on and on. And it took six months. And at that point I said, "Okay, guys. This is not going to work much longer." And so Dick Pashley actually came up with the idea. He said, "Let's - let's just set up a separate module. In other words, we'll hire our own operators and have our own equipment and start to build it. But we'll do it in the fab. Not and R&D lab but just do it in the fab." And we actually got the opportunity to practice that because Intel decided to get out of the D-RAM business. And serendipity when Fab Five in Oregon was making D-RAMS, Andy decided to shut down the D-RAM business, and I lobbied very hard, "Can I have Fab Five?" for doing development, this module. And of course the initial answer was "Hell no. You're setting up an R&D lab and it's not going to work." And I said, "No, let's run it as a fab. We'll let the development organization run it as a fab and we'll turn it over to the manufacturing when it's running," because we - because we just were really having too much trouble using - when you made major technology changes you just couldn't do it on the production line. And so that was really the - the first of it. And it worked out great. I mean we were - we actually I think had R&D profitable because we were producing three to six shrinks out of that - that hundred - hundred megahertz parts or something, and - and - not quite a hundred, maybe fifty megahertz - and Intel was selling them so I could charge - charge manufacturing for producing it, and so I cut the R&D bill. And so that worked really well. So we - we kind of got the R&D guys to really prove that you could make it. And so the first phase of that was major improvements in yield. Then - then we pushed this whole copy exactly kind of concept where, okay, if it was run this way it had to be run the same in manufacturing, because otherwise you'd have a situation where manufacturing would re-engineer the process. That was just standard procedure. Of course those R&D guys don't know what they're doing. And so it evolved over a number of years to where I think Intel really has just a wonderful capability to develop and transfer and rapid manufacturing. And there are really a number of key points. One is that you have the R&D guys really do the work in terms of yields and output of the equipment, and then you copy that exactly in the manufacturing and it - or replicate is really - you - you know, you replicate what works and we're just fanatic about that. And so that - that - that has helped a lot. Now the other - the other concept which is very similar to this "do it on the manufacturing line" is I would only let the engineers change a maximum of thirty percent dollar weighted of the equipment or process capability so that when you went from a one micron technology to a point eight micron technology they could only - seventy percent had to be the same which guaranteed that at least seventy percent was working and only thirty percent was new. And it turned out that that was relatively easy to do so you kind of got the benefit of - of effectively doing it on the manufacturing flow without just throwing everything out and starting over. So there were a whole bunch of things in there that - that sort of made it work.

RW: Well, Intel stayed with - with N-MOS quite while and the Japanese really became the - the C-MOS experts.

GP: Well, and you know, I - I'm not so sure that part of our development methodology didn't create that. Remember I told you I had a hell of a time getting C-MOS developed in the manufacturing line because the manufacturing manager just didn't want to cooperate. But on the other hand at the same time we were convinced that - that C-MOS was going to be so much more expensive because of the added processing flexity that we like to stall as long as possible, and ultimately we went to, but - but I - for our products I - I would still submit that staying with N-MOS as long as we did was the right - the right choice from a cost effectiveness point of view.

RW: Mm-mmm. Did you ever read the book Inside Intel?

GP: Yeah.

RW: Well, they make - they make - the author their charges that Intel was always weak on quality assurance. And that - and that Grove personally didn't believe that - that the Japanese were that much better.

GP: I'm not sure I would - I would agree with that. I - no, I - I would disagree with that strongly. Because, you know, we were blindsided by the Japanese. There's no doubt about that. And let me - let me give you a little anecdotal joke, if you want to, that in the old days you ship products to a one, you know, a one percent AQL. Now what that roughly means that out of every hundred parts, two of them are bad. And that was acceptable. Customer said, "I'll take it" knowing that ninety-eight were good, two were bad. Now I'll give you an ethical dilemma. If you're inspecting it you find two bad ones. What do you do? Do you throw them away or do you put them back in the box and ship it to the customer? If you throw them away you've only shipped ninety-eight and the customer's unhappy because you shorted them two parts, but if you give them the two parts - two bad parts they - they presumably are happy because you shipped them what you said. Now this - that's how quality evolved and people are happy with that. But then - then, quite frankly, the Japanese surprised us by saying, "Wait a minute. We're not going to ship two bad parts with every carton." And then, okay, that became the new standard. Now Motorola got totally carried away with Six Sigma as a big slogan that they couldn't possibly meet or deal with, and that kind of got the whole quality thing. But - but the real - the real breakthrough, and I actually would credit Craig Barrett with that. Craig kind of - Andy would just get a baseball bat or whatever and threaten people to make it better. Craig actually brought in some content experts and - and really said, "Hey, we're going to improve our quality throughout dummying and through - through real methodology." And so it was probably in the early '80's that the - that all this happened anyway. And so then we went, you know, we went through a major quality revolution. And Andy was behind that. I mean there - you know, for that. But Craig was the real driver on it and it was a - in response to the changing world. The stuff that was perfectly acceptable in the early '70s was no long acceptable, so, you know, you do what the customer expects. And I think Intel's always tried to do what the customer expects. So I - I would dispute that we were shipping crap.

RW: Well now, you - you brought up some personalities. You were there very early and you knew personally Bob Noyce and Gordon Moore and Andy Grove. So what - what's your assessment of them and their personalities?

GP: Well, you know, my - my strong feeling has always been that the three of them combined are what made the company what it is. I mean Bob was clearly the visionary - kind of it - the symbol of the company, the outside man, you know, looked at by Wall Street investors and so forth. Gordon was - was kind of the strategist and the technologist that - that really in my mind set the course for specific technology directions and strategies. And Andy was - was the ramrod. I mean he basically set the tone for discipline and operational excellence. And no one of those things by themselves would have done it. You needed all three. And I think they complimented each other very well to where it was very clear that that's how the thing was run. And that's really what drove the culture and I think what made it - made it successful.

RW: Now - but that evolved over time, right? As - as first Bob stepped back and then - and then Gordon and Andy was free to do his thing.

GP: It didn't change much. I mean after Bob obviously left and then after - you know, left - he left for Semitech and so - then Gordon kind of took on more of the statesman role for the company, and - and Andy stayed pretty much on the operational side, although he - he then - after Andy became CEO he started being a little bit more of a strategist. But - but I - I still used to always quibble with Andy because he would - he would make a strategic long-range plan presentation that he had clippings from that morning's Wall Street Journal. I'd say, "What the hell kind of strategy is that? Why don't we wait until tomorrow? There'll be a different strategy." But - but Andy in general kind of took over that role of strategist. But Gordon was still - still kind of the thinker behind it. And now of course as Gordon phased out of it Andy truly did pay attention to strategies and - and business, where we were going and that type of stuff. And Craig took over the role of being kind of the operations honcho of - of pushing manufacturing and pushing, you know, pushing operations.

RW: But a lot of - a lot of the senior people that left Intel didn't get on with Andy, or he didn't get on with them, one of the two.

GP: Well, Andy is fairly digital. He either likes you or he doesn't. And that isn't necessarily forevermore. He - he likes some people and doesn't like them later, but he's usually - there's - there's really not much equivocation on whether he likes you or not. And - but I don't think that was necessarily it. I think Intel evolved, Intel grew, and, you know, if some of the people had different aspirations. They either wanted to run the company and that wasn't in the cards, or they wanted to something else, or they're obviously quite confident to go do other things. So I don't - I've never thought that was negative or a big personality issue. I think it was just a growing up kind of a problem.

RW: Well, how about with Will Kauffman? What was the problem there? He and Craig evidently didn't get - didn't get on.

GP: Oh, I'm not sure I would again agree with that. I mean I don't - I don't know - I don't know, and I usually choose not to care as to who gets along with whom, but I think the - the real issue was that Will probably took some of the let's say the - the blame for Intel's manufacturing not being up - up to the Japanese role model. And so - so Craig actually took over from Will when Will had been running manufacturing. And Craig kind of came in with a bullwhip and - and, you know, got everybody really thinking. And - but you know, that's - that's almost the classic situation where you have somebody that's been running an organization for a long time and it has to undergo major change, sometimes it's better to bring an outside change agent. That's what IBM had to do. The old IBM guard couldn't make the changes that were necessary for - that Gerstner did. And I think we were kind of in the same mode over in manufacturing, so it was probably appropriate to bring Craig in to, you know, really put a new day - a new face on manufacturing and - and so Will probably got blamed for some of that and, you know me, that's that way it is. But beyond that I'm not sure there was any - any real vindictiveness there.

RW: Now how about the Intel history of litigation? I understand they've got over twenty lawyers now on staff.

GP: Well - but you mean history of being sued or -

RW: No, suing.

GP: - suing? Well, there are - there are two - two sides to that, okay? One - one is that you're a big target and so people are going to sue you. And - and - I mean Lemelson was a classic case. There are people come out of the woodwork. We were being sued by Hughes for the Bauer patent which I had to - I wound up testifying in that, and it ultimately got settled, but you know, people dredged up old patents that probably should have never been issued, and they see that Intel's making them a lot of money on 486s or Pentiums so let's sue them and we'll take our chances in court. And so - so there's that aspect where you have just a lot of - of lawsuits coming at you. And then there's the other aspect of - of protecting you brand and protecting your intellectual property. And we'd - we've actually got the DA in Santa Clara to bring criminal charges against a couple that had left and stolen the - the test specs and the specs and the process information on the old math coprocessor. And I wound up testifying in them and they actually got off. And I - I couldn't believe that they got of but the - the defense attorney was a, you know, a Melvin Belli to be, and he - he got off to well, Intel was behind the technology of the Japanese. Yeah? And you're really careless but you gave your process specs to two hundred engineers? How could - how could that be valuable? And so, you know, my feeling was, "Oh, okay. It's okay to steal an American car because they're not as good as Japanese cars, but it's not okay to steal a Japanese car." And - and they got off. The jury said, "Oh yeah. This obviously wasn't valuable." So, you know, is that - is that litigious or is it trying to protect your property? The other interesting case in point is, you know, your trademarks. You have to protect them. And so Intel inside has been a - a trademark for a long time. And every once in a while you get other people trying to use that or, you know, even - even use it facetiously, and you have to kind rattle the saber and say, "Hey, you can't do that," because if you don't then you lose your rights to it. So - so a lot of what Intel was doing is really just to protect the trademarks and the proprietary things, and if - and it was really more in response in my mind that if you don't do it you lose them. Not that - not that we want to be jerks and go after everybody and - and - but, you know, that's my feeling. I don't know. There may be other opinions on that one.

RW: Well, Gerry, Silicon Valley has been a huge success in spite - we've had a few bumps along the way, especially recently. But looking at it overall it's been a huge success. What - what do you think some of the factors are that have made it so?

GP: Well, I - I mean I personally think its – it's been technology driven as much as anything. You've got Berkeley, Stanford and, you know, we were - I think we were fortunate to have the originally, the - the Shockley Semiconductor come here. If it had gone to the forest up in Oregon, that probably would be the - well, maybe not because of the university connection, so - so there was kind of the - the new technology emerged, and then for what ever reason venture capital was located here and so you had the convergence of both the ideas and venture capitalists and the - the right talent to start these things off, and then once it - once it got going, as you know, Fairchild was started and then everybody went from Fairchild to start new companies, and it's kind of gone from that.

RW: But, for - for example, you can fail here. It's okay, as long as you don't - you're not crooked, but it's okay to give it your try and - doesn't work out.

GP: Yeah, I - I suppose. I mean I - I'm not sure what to make of - of -

RW: Yeah, half of that hasn't happened.

GP: Well, no it isn't that so much. I think that the "you can fail" part of it is probably a correct assessment in the sense of the entrepreneurial system. And if you think about it, what, ninety-nine out - ninety-nine out of a hundred are not terribly successful and - and not that many survive, and so the whole entrepreneurial thing is just to keep trying. It's - it's almost a Darwinian system of just "keep trying" where as - to your point about failure, when you have a larger company people are expected to - everything to succeed. And actually the - the last job I had at Intel was starting new business. And that was a tough one from the standpoint that the culture was that everything we did worked. And I - I used to joke that I - I never built two factories hoping one of them would work. You know, they had to work. You didn't design a Pentium I or II and hope it works, whereas when you're starting new business there's a little bit of a hope involved. And so we actually had a lot of trouble with that because I - I would literally start at ten, twenty, thirty new little businesses within Intel, and there's kind of the expectation that they're going to work. And the reality is ninety-nine percent aren't going to work. And so then - so within a big corporation that fear of failure is actually quite high and it's not as acceptable to fail. Whereas to your earlier point, if you - if you're in the entrepreneurial world where a venture capitalists - and the venture capitalists know that not all of them are going to succeed or all their companies are going to succeed, and that's okay. I mean they - they'll still take a pound of flesh out of the poor start-up guys before they die. But it - that was okay. And so I - I - I do like your analogy there. And so perhaps the East Coast companies - bigger companies outside of this valley have that same issue that, you know, everything that you did had to work.

RW: Well, like Japan.

GP: Japan, failure is - failure is - is - is a horrible thing, so everything has to work. And - not that fear, a lack of acceptance failure in a society or a big corporation does inhibit the entrepreneurial spirit and so, you know, this valley absolutely had the right chemistry to allow those experiments to happen.

RW: Well, what was Intel's rationale for starting little companies?

GP: Well, it's the same as every other company. As you say, well how are we going to reinvent ourselves? What happens when our current bread and butter starts to run out? And, oh by the way, we're losing all these people to start-ups because they want to do something and exciting so let's give them an outlet inside the company. And that part actually worked pretty well, but there were a lot of people that just wanted to - to try new things and didn't really want to leave Intel, they just wanted to try a new thing. So we gave them an outlet for doing that. And - and that worked out great. And, you know, there were new technologies, internet, you name it, that are always coming out, and so as a large company you want to have - have some access to those technologies, and so starting them inside of Intel in an R&D mode or in a small business mode made a lot of sense. I still think it's - it makes a lot of sense, but - but the thing - the learning that we got out of it - that I got out of it is that, you know, this fail - failure or not failure thing is a different dynamic inside of a company that's used to having everything done right and working, whereas this going out in - in the venture world where it's okay to fail.

RW: Now how would you compensate someone that went off and started this little company within Intel? How would you compensate them if - if it went well?

GP: Well, the compensation was always an issue. It - it really was an issue. And quite frankly it drove me nuts. From the standpoint you had two types of people. So - so we being the inside venture capitalist that's going to fund these. You have two types of people that come up to you. One type of person just wants to do something new and they - and they're perfectly happy with Intel compensation, Intel stock options, and they just want to try it. They just want to do something. And then they get on with what they're doing. The other type has kind of a "get rich" mentality. And so a couple of things we started, we literally spent half the time, half the air time talking to these people about what their compensation was going to be as opposed to what the hell the business was going to be. You know, you try to bring the conversation back around. Let's talk about your business plan. No, let's talk about my compensation. And I said, I don't give a damn about your compensation; I want a business plan. And those people, quite frankly, you're never going to satisfy because if that - if that's the modus operandi you're better off letting them go to a venture capitalist and let the venture capitalist give them twenty percent ownership, and then the first time they need more money get them five percent, then cut them to five percent the next time they need money was then out to nothing. I mean that's - that's life. But - but I was - I - I got fairly annoyed at - at the people that were just trying to figure how to, you know, how to do a big con package, because the room - the people that you wanted were the ones that were motivated to do - do a good job and come out with new technology and do something. That's inside a company. When you're - when you're - when you're out in the venture world I have no qualms about the - the entrepreneurs asking the VC's for half, you know, half the company. And - but within Intel it didn't work.

RW: But - but Intel also took positions in - in start-up companies outside, right?

GP: There's a totally different approach to life. And - another little learning as you go along, if you own less than twenty percent, in other words, you own nineteen percent of a start-up, then you're an investor. So you don't have to consolidate with your own earning statements. It's just an investment. It's just like putting money in the bank. Once you get above twenty percent, beyond twenty-one percent up to fifty you actually have to consolidate the net on your income statement. And if you own fifty-one percent you consolidate it line by line, operations, costs, earnings and so forth. And so the problem you actually get into, and so Vadasz and Intel Venture Capital actually tried to stay below that twenty percent because they were truly just investments. And there were investments in technology that - that we felt would help the industry, it would help sell Intel product would basically bring out broadband, you name it. So it was that kind of investment. The stuff that I was trying to do that was internal was to really grow businesses. And there you actually got into some very interesting dynamics. For example, if I were to set it up as an entity and Intel owned ninety percent and employees owned ten percent, all of a sudden you - you have effectively minority shareholders and all the rules of how you run a corporation are different. And so it - so I finally came to the conclusion there are only two stable states in this world. One is you own a hundred percent, and the other is you own nineteen percent. Any combination in between ultimately cause problems from a just SCC ownership shareholder point of view. And we kind of went through that learning and we got in the middle on a few of them and I really wish I was at either end.

RW: Yup. Well, at - at Intel - at LSI Logic, excuse me, we also had some people that wanted to spin out, and the problem I had with it was paying royalties or supporting one group and then your loyal employees that are with the major company, they see this going on and you're - you're - you're pay - you're - you're giving incentives to these other people.

GP: There - there's no easy answer to that unless they - I - I think the simple thing was always to just create the environment. Let the people that want to try something new, give them - give them the opportunity to do that. Set it up as a separable unit and so forth. But the minute you start to get into the - the venture-type compensation, ownership situations, you're probably working on the wrong problem. Those people really just need to go off and do it on their own because any other scheme has – is fraught with problems. Either they're using the intellectual property or the hard work of Intel or LSI Logic to then make money on their own, or you're distorting your compensation systems, so - maybe I'm being too negative on it, but it - it really hasn't - it - it hasn't worked out very well.

RW: Well, also part of it is that all the Intel people share in the success of Intel -

GP: Right.

RW: - with - with stock purchase plans, with stock options. How many millionaires have been made at Intel?

GP: I don't know. We've probably got a few of them.

RW: More than a thousand?

GP: Yeah, probably. Yeah.

RW: Yeah. Lots.

GP: Well, you know, compensation is always one of the - the most difficult things at any corporation. And people really pay attention to equity and fairness and so forth. And even at Intel for a while there we got into situations where, you know, you have the microprocessor division which really makes the bulk of the earnings and the profits, and then you have new businesses trying to get into networking, trying to get into those, and occasionally you get into a situation, well, why - why should the guys working in the networking group get a big bonus when in fact all the money was made by the processor guys? And - and occasionally it worked both ways. Sometimes these guys would say we want our own bonus plan based upon what we're doing. Then they - then they look at it and it's going to be zero. Oh, no, we like the processor bonus plan. So there's never - there's never been a real, you know, easy answer to that but - but always to err on the side of just being as equitable as possible because ultimately that's what people care about is that you're being fair. But the compensation schemes, especially when you have start-ups and you have divisions with major earning differences, they're always difficult.

RW: Well, I do think though that one of the - talking about Silicon Valley and one of the advantages that we've had is that stock options have been a way of life almost from the beginning. And - and there has been a sharing of the wealth that you haven't seen in so many of the traditional companies around the world, like in Japan. You don't get rich in Japan even though you may be president of the company. So -

GP: Well, but that's more of an artifact of the - the size of the companies and the growth. I mean the - when you have a large established company, General Motors, offering stock options in General Motors to their five hundred thousand employees probably is not nearly as big an incentive as offering stock - stock options to a start-up that goes from zero to billion - billions of dollars in revenue. And - and in Japan you don't have the start-ups, you don't - you have just large corporations. So - so it's an artifact of the fact that these start-ups grow to be very large corporations that makes the stock options so attractive. It - it doesn't have the same benefit in a large corporation that's growing slowly.

RW: Well, thank you Gerry. This has been very interesting.

GP: My pleasure. It's great to talk about the old times.