James McGroddy

Thomas:

It is August 14, 2024. I’m Will Thomas with the American Institute of Physics, and it’s my pleasure to be sitting with James C. McGroddy, former director of research at IBM. We’ve done a couple of interviews with people at IBM that are in our collections—with John Armstrong, Ralph Gomory, Marshall Nathan…—and we’re very pleased to add this one. So, Jim, to get started, could you just tell us a little bit about your family background?

McGroddy:

Yeah, so if we go back, that’s a picture of my mother’s father on the mantlepiece there. He came from Ireland about 1890 at age 14 with a fourth-grade education, from Tipperary. And unfortunately, and it really frustrates me, we didn’t do these kinds of interviews with him and the rest of our family. I knew him well later, when he had become a very successful plumbing contractor in Manhattan.

On my father’s side, his family had come, I think, two generations earlier. We can’t find a connection back to Ireland, which is puzzling because we have found people there with the same name, but looking at least at the genealogical records, we can’t find a connection. We haven’t done the DNA, which we should. I’ve tried to get them to do it. I’ve done it. They were from up in the northwest in Donegal. And my father’s father was, by the time I knew him, retired. I think he had some sort of a functional job with the city, maybe a building inspector or something like that. And he lived in the Bronx.

My mother’s father had bought a house on East 17th Street in Manhattan about 1925, and I don’t know how he got from being a 14-year-old kid with not much in his pocket to having that brownstone house. He married an Irish woman who had spent much of her life in England, as many people did, to be house servants, and nurses, and things like that. I know very little about her. She died very early, as did my father’s mother. But I knew both of my grandfathers. My father was one of two kids, so he was raised in the Bronx, went to Xavier High School in Manhattan, a Jesuit high school, and then to Fordham University and Fordham Law School. He graduated there about 1930.

He had one sister who, I believe, also went to college and never had a profession. But I knew her well, she visited here many times over the years. On my mother’s side, she had one sister who mostly got involved in my grandfather’s plumbing contracting business, sort of ran the office and whatnot. The office was actually in the house, a brownstone, where you had one level that’s down a couple of steps from the street, and that’s where the office was. And then, they had a couple levels below that with pipes, and wrenches, and some of the tools of the trade. So, one of the remarkable things about that point was, these people, particularly on my mother’s side, a kid comes over 14 years old, fourth-grade education, has two daughters, insists in the 1920s that they both go to college.

Both of my mother and her sister, who was older, went to what were women’s colleges, which was pretty much the only option until the 1970s, in fact, and studied things that were appropriate for women to study. They got a liberal arts education and I’m sure learned some other things that were preparation for being a wife and raising a family. My mother and father married about 1931 or ’32—Depression time—after my father finished law school. Big fancy wedding. That clock over there was, I’m quite sure, a wedding gift to my grandfather when he got married, and was in that house where they lived until my mother’s sister died. I got it and got it functioning again. It’s been here for 25 years or more now.

After they were married, they lived in the brownstone where my mother’s father, and mother, and sister were living. A typical brownstone, that sort where, as I said, the level that’s a couple steps down was used as office space. Originally, it had been built as a single-family house, and that’s where the help would live, and the kitchen was there. And then, on the next floor up and the floor above that, there were sort of serving kitchens, tiny kitchens, and a dumbwaiter, or an elevator, essentially, for food. So, the kitchen was in the basement. The first floor up was the parlors, where entertaining would’ve been done and fancy meals served.

And then, the next level up had the dining room and a couple of bedrooms, and then there was a unit above that, which I don’t know at the time what that was used for. But what my grandfather did was offer to my parents that they could live in the parlor level, and he put a stairway into the back of the downstairs, where there was another bedroom and a connection to the garden. It’s a 25-foot-wide brownstone, so you had a little garden, 25 by 40 feet or so. So, we lived there. When I was born, that’s where my parents lived. I have an older brother, born a year-and-a-half before me, and as good Catholic folks, there were seven of us altogether. And my first sister was born, like, 14 months after I was.

At that point, my parents, presumably to have more space, and more open air, and all that sort of thing, moved to a house in Upper Nyack, which is directly across the river from here, which they rented. We lived in that house for many, many years, eventually bought it, and my father commuted. He was in a law firm in lower Broadway, and he commuted on the equivalent of this railroad on the other side of the river, which had passenger service back then. It was called the West Shore Line that ran New York Central. Took a while, take a train from West Nyack to Weehawken, New Jersey, and then they had ferries that, one went up straight across to 42nd Street, and the other went down to Cortland Street to connect with Wall Street. But a lot of people commuted, and I went to high school that way for four years, the same route.

Thomas:

I should probably note for the recording that we’re in Briarcliff Manor, New York, right now, at your home.

McGroddy:

Yes, right along the Hudson here. And in those days, there was no bridge across the Hudson here. The original Tappan Zee Bridge was built in 1955, so it was a different world on the two sides of the river. So, I was raised in Upper Nyack, went to the local Catholic school run by the nuns, and I got kicked out in sixth grade for making trouble. And I think boredom was always a problem, particularly in the schools, because the classes were huge. It was segregated, both by race and religion. The Catholic school had no people of color, and there weren’t any Hispanics around, there were no Asians.

The white Catholic kids went there, and it was sort of a religious rule that if there was a Catholic school, you could go to the Catholic school to get taught properly. But in sixth grade, I, with another guy… it must’ve been a warm spring day. We went out with, as I remember, a mayonnaise jar and captured, like, 100 grasshoppers in a field and took them back and let them loose in the classroom in sixth grade. And everybody went nuts. I still remember the poor nun telling the girls, “Don’t kill the poor creatures of God.” [laughs] But the next week, I found out I was in a different school in Haverstraw, which is about 10 miles north, a school run by Marist Brothers, which had much smaller classes.

At the grade-school level, it had the boys and the girls on different sides of this brick building that was built with good isolation. And they were very good teachers. And particularly on things like the structure of the English language and grammar. So, later on, when you study other languages, I was well prepared for that. And the classes were small. I think in the Catholic school in Nyack, St. Ann’s, I would guess there were 60 kids in the class with a young nun trying to teach, and you’d have one reader book per semester. It was totally boring, because most of the kids were way out in front of it. With the Brothers, they actually challenged you to a good degree, and I was lucky that the nuns kicked me out, probably.

For high school, I went to Regis High School in Manhattan, a Jesuit high school that was founded in 1911 by some woman who, after her husband died, gave a couple of million dollars to build this school and presumably to run it forever as a tuition-free boys’ high school. It still exists on 84th Street. I commuted on this West Shore Railroad, two hours each way.

Thomas:

Each day?

McGroddy:

Yes. It seems crazy when you think about it now. And it was a school that focused very much on classical education. Everybody took four years of Latin, and then you had a choice of taking three years of Attic Greek and two years of French, or two years of German. And I took the German track, which then allowed me to study physics and chemistry. They didn’t have biology, interestingly enough.

Thomas:

The language was actually connected to the scientific aspect of it?

McGroddy:

No, the focus was on the languages, and to some degree, on literature, although we didn’t read a lot of books. Remarkable school. Remarkable for what they didn’t do, in many ways. We were two blocks away from the Metropolitan Museum of Art—never went there. Very self-contained in this granite temple. And I’ve argued with them now for more than 40 years about how they run the school, and actually eventually withdrew my support of the place to support a different school, which I’ll tell you about, that, I think, meets what this woman, when she founded it, intended to do. So, science was weak, mathematics was okay. As I said, there was no biology. And an interesting point about that is that a graduate four years behind me was this guy, Anthony Fauci. He went to that high school, and there was no biology. Look where he ended up. He then went to Holy Cross, which is a Jesuit college up in Worcester, Massachusetts, and somehow connected with biology there!

But the physics teacher… chemistry, I didn’t like. I took chemistry in third year and physics in fourth year. And I struggled with chemistry because I wanted to go a level below what chemistry teaches you. Why do the atoms have these electrons spinning around, and what keeps them going? But no: “learn how to balance the chemical equation.” It was very rule-driven, as a lot of chemistry is, memory-driven.

Thomas:

Oxidation numbers and all that?

McGroddy:

Yes, valances. So, I didn’t do well, and I didn’t like it. But the physics teacher, a Jesuit priest, Albert Rooney was terrific. The physics course was good. And there was a key thing, I think, career-wise for me, he sponsored what they called the radio club. And what we did in the radio club, which only had a few members, less than ten, is you got yourself on a path to get an amateur radio license, which meant you had to learn Morse code, and there was a whole bunch of stuff about vacuum-tube electronics that you had to learn. There was a test to get your license.

He would take us down to Cortland Street, where there were a ton of stores selling war surplus electronic equipment, vacuum tube-based. He had a design for a primitive Ham radio setup, a set of receivers and transmitters that were built for airplanes and transmitters for air-to-air communication mostly, called ARC-5s, which you could buy and convert, build a power supply that used 110-volt AC instead of… the airplanes were on 40-volt DC. But the point is, we learned a little bit about tubes, and electronics, and how a power supply worked, and this whole set of things, and that fascinated me. Because before that, when I was a kid, I was always interested, like, in clocks. I was really into clocks.

By the time I was 12, I was into cars. Near our house, there was a car scrapyard with a whole bunch of old cars in it. And with a big family, your parents don’t keep track of you every minute of the day. I spent a lot of time playing in the junkyard. It was wonderful. So, by the time I was finishing high school, and I’d done very well in physics and in math, in terms of going to college, that school, even for some years after I was there, would not send your transcript to a college that wasn’t a Catholic college. They claimed that that was implicit in the contract that you agreed to when you accepted the scholarship to the school.

So, I knew a couple of people who resented that. They had an opportunity to go to other colleges, on scholarship even, and they couldn’t get a transcript. The one guy was a year or two after me who complained about it, he ended up becoming the CEO of Mobil Oil, so he didn’t do too badly. Another guy ended up becoming the head of McKinsey. About two years ago, I finally got a transcript from the school, which I had never seen. I don’t remember much discussion about career or where you’re going to go to college or major, but it must’ve taken place in some limited way. There was no such thing as guidance counselors in those days. But I ended up at this place, Saint Joe’s College at the time, in Philadelphia, in this course with this major that this guy, Father O’Conor, had created after the war, Electronic Physics. As I said earlier, in this major you’d take the full spectrum of the physics courses and learn to be a semi-competent engineer using vacuum tubes.

Thomas:

What were you sort of envisioning for yourself at this point when you went off to college?

McGroddy:

I don’t know that I envisioned anything beyond. I think life was, as best I recall it, more day-to-day, month-to-month. But I ended up in this program. I don’t remember a lot of discussion about alternatives such as, “Which college do you want to go to?” for example. When I look now at my transcript, I can see that it was sent to Villanova, which I think had an engineering school at that time, and Manhattan College in Manhattan, which is a very engineering-oriented school, and Saint Joe’s, and that was it. And I remember taking some tests about what your preferences might be, and the only one I remember hearing about was from this guy who was the school chaplain, but he was as close as it came to a guidance counselor, Father Burke. That test concluded that I really should be a farmer.

I’d worked on farms locally as my summer jobs when I was maybe… back then, when you were 14, you got working papers. You worked. You wanted to work, and you did work. You didn’t go to camp, or off to Europe, or anything like that. I worked on an apple farm, bicycle distance from our home, and there were a lot of them at that time. The farms eventually disappeared as the suburban growth drove the value of the land up to the point where the owner could sell the land, and put the money in the bank, and make more money than doing all the work of raising the apples. It’s the same sort of thing that’s still going on in large parts of the country. We see it in New Hampshire where we have a family vacation place, with corn and the like.

But I didn’t have a vision that, “In 20 years, I want to be doing X.” I had unarticulated but built in, that I didn’t want to follow the path that my father had followed and that my older brother was following. My older brother also went to the same high school I did a year ahead of me, and in those days, they admitted I think it was 160 kids, boys, each year. By the time they graduated, 60 of them would’ve been failed out. The rule was, if you failed any course for two quarters, you were asked to leave, so people just disappeared around you. My brother was one of those. He took the Greek-French track, and ended up at the same high school my father had gone to, Xavier, downtown on 16th Street. Regis is on 84th Street, but we still commuted together. He then went to Holy Cross, went to Fordham Law School and became a lawyer. It was a very similar to the path that my father had followed.

I, somehow instinctively, knew I didn’t want to do that, not in any analytical way. And I was fascinated by things and how they worked, cars. I used to go to rummage sales and buy mechanical clocks, pendulum clocks, and whatnot. Eventually, my mother threw them all out. The only pendulum clock I have left is this one, which has the risk that the pendulum is two tubes of mercury. You don’t want to drop it. When I went to Saint Joe’s, I viewed it I was sort of on a programmed course. I didn’t decide where I would go to grammar school, and when I switched grammar schools, I was told. And I think in terms of the college thing, I must’ve had some input into it, but not in terms of a whole bunch of options, partly because you had to go to a Catholic college.

Regis in those days was dominated by the Catholic view, in its most simplistic form that you’re on this Earth for some period of 50 or 100 years, and your objective is to not go to hell. And as a Catholic, you’ve got a chance to not do that. If you’re not a Catholic, you’re really in big trouble. But it’s very easy to do something, violate any one of these rules—and for teenage boys, it would be mostly about girls and sex. But rules about, you didn’t go to church last Sunday—that was enough, you’d go to hell, unless you went to confession and got that wiped out. So, it’s a very good power structure that they had built. The most dangerous kinds, they buy all that. Some people still buy that.

Thomas:

I was reading in the biography that you prepared for your family that you have remained a practicing Catholic.

McGroddy:

Yes. Although I’m not too deeply into it, and I don’t go to church now because of covid and my limited mobility. But I’m in touch with—and he comes over once in a while—the local pastor. There was a long period where I was totally separated from it, but when I got married when I was finishing graduate school, I married a Catholic woman, and we had four kids and pretty much followed “the rules”. I had big fights with the Catholic Church during that period, mostly over civil rights issues. I one time got involved when there was one of these marches on Washington. I was involved with a bunch of other religions in the area. I lived 15 miles north here. I said, “We can use the rectory porch as a place to collect sleeping bags, and food,” and whatever else it was they were collecting. And when I told the pastor, he told me, “You can’t do that!”

And I’d started a small organization, I don’t remember how we got to it, some years before that, which was not a unique idea, something I read about that I think was going on in England, called FISH. This organization, what we did—which as I said, I read about it being done in England—people would volunteer to be available to drive people to do errands they couldn’t do themselves, one day a week. We had some telephone-based system to organize that, to take some old lady to the doctor’s, and the like. It was an interfaith thing. I also got involved with a place in Peekskill, which was providing healthcare for poor, mostly Black people, called Aunt Bessie’s Open Door, which has now evolved into Open Door generally, which is a big thing in this county, still provides healthcare, dentistry, and such to people who can’t afford it, have no insurance.

Each of these things, I would get in trouble with the pastor. But on the other side of it, I was a lector, a leader of song on the parish council, etc. At a certain point, I decided that when you go to St. Patrick’s, and you look at the church, and you mostly go inside, it’s beautiful inside, you really like it. But if you really want to appreciate a church—at St. Patrick’s, you can’t do this—you go outside, and you go a distance away, and you look at it. St. Patrick’s is buried in Rockefeller Center and whatnot. You can’t see it in a larger perspective. I remember consciously thinking, “I’m going to step way outside it and look at it,” so I dropped out, stopped going to the church. I would still do the interfaith things, but not as part of that parish.

My wife stayed Catholic. Eventually, and that was probably one of the issues, she divorced me, in 1979. Beyond that point, I just didn’t pay much attention to the Catholic Church for many years. Before that, it had played a big role in my life. I remember when Pope Pius XII died, I was in the beginning of graduate school, and I couldn’t imagine that. He had always been the Pope in my life. “How can he die? There’s no other Pope.” I eventually remarried, to a Chinese woman, had a daughter with her, who was here yesterday for dinner. She still has a room here. She was born after I bought this house. And she just bought a brownstone in Brooklyn.

When this daughter, Aileen, got to be around 12 or so, and she was in public schools, many of her friends were preparing for a bat mitzvah, or a bar mitzvah, or a confirmation at a Catholic or a Lutheran church, and she wanted to do that. So, I set her up with people at Saint Ignatius in New York. She got a sort of quick instruction, and I got very involved myself, and connected again, now many years ago. That would’ve been about 1998. She was born in 1988, so she would’ve been 10 or 12. So, it was around the year 2000. So, now, I’m a member, but I don’t go.

Thomas:

It’s very different from the school environment that you came from.

McGroddy:

Oh, totally. Of course, the schools are all gone, and the nuns are gone. But there are still lots of these conservative priests around. Like, in high school, part of going to school was… I think it was every week, they marched the entire school over to Saint Ignatius Church, and everybody went to confession. [laughs] And you had a Friday morning Mass for the whole school. And they’d keep track of who went to communion, and who couldn’t because they had had impure thoughts or something. It was a world in which you avoided sin. You were here for a short time. What does it profit a man if he gained the whole world and loses his soul? That this is just a brief thing, so you don’t focus on it.

So, it’s then we get back to the issue of career and long-term vision. Didn’t have it. You’re on this process, and the next step is college. And then, in college, I got involved in various sorts of things, on the electronic front and otherwise. One of the good things about the college program that Father O’Conor had created, which gave you this bachelor’s in electronic physics, it was a co-op program. And there were a lot of electronics companies in and around Philadelphia at that time. And the way the program worked, after your second year, you went on a trimester program, and every other trimester, you worked full-time in a company and took one course at night. And on the others you would do is go to school full-time.

But you had these three blocks of semester-long, 16 weeks or so, where you worked in a company. I worked at RCA on a project where they were developing the fire control radar for the F-104 fighter plane. And one thing I’d learned there, which affected my career a lot, was that the engineers who worked there mostly didn’t understand very deeply. They knew how to use a tube manual and the textbooks. I remember the first project they had me work on there was, they built this prototype of some circuit for the radar, which was a pulsed circuit. It didn’t work. The circuits just didn’t work right. And the reason was—and I figured this out pretty quickly—you think more like a physicist than an engineer, that they were using… do you have an engineering background?

Thomas:

I did undergraduate physics before switching to history, actually.

McGroddy:

Some vacuum tubes, particularly ones used in pulse circuits, are very linear because you only want one or zero. Analog stuff that you had classically, before that everything used small signals, and you could look at things differentially. But basically, they had not put big enough decoupling capacitors into this thing to power the charge through, and I could understand that without having studied that. I do remember I bought a book—I think I still have it, hate to throw textbooks out—Pulse and Digital Circuits by Millman and Taub.

And then, there was another time where I was working on a unit that had the microwave generator, a tube called a magnetron, which runs at very high voltage. Microwave ovens use these, which also run at very high voltage, couple of kilovolts. And it normally would be in a pressurized case when it was in the airplane, but we were working on it in the open. I don’t remember exactly what we were doing. And there was this blue glow around the wires that went on. And I remember asking the guy I was working with—his name was, I still remember him, McCue—what that was. He said, “Don’t worry about it, kid, it doesn’t use any power.” So, I got the idea these people don’t understand deeply, and I knew transistors were coming. We had one course in college on transistor physics, so we were exposed to that, and I used to fool around with circuits. I even built a little pre-computer gadget.

I still have an article from the college newspaper. We had a whole bunch of philosophy classes that were not very well taught and very narrowly focused, very Catholic. I remember in the logic course, getting in an argument with the professor, who was a Jesuit, about how computers were going to be able to do some form of logic, and he said, “No, absolutely not.” And I actually built a little… it was hardly a computer, I think it was just using diodes and neon bulbs to do something. There was a chapter in the textbook about some piece of logic, and I had this little cardboard thing with a wire, and you could touch it on different things, and other things would light up, and then you could see what that meant.

So, I was into that sort of stuff at that time. I concluded at the end of college that I didn’t want to be an engineer. I had a wonderful offer from RCA, $6,500 a year—big money, and my family struggled with why I wasn’t going to accept that. But I went to the University of Maryland, which was somewhat local. Again, I didn’t do a very good thorough job of looking at graduate schools. I did apply to Stanford, but I applied to physics. I didn’t know at the time, they had a separate applied physics department. And I didn’t apply to any engineering departments.

Thomas:

Just to put some dates on things, you were born in 1937?

McGroddy:

Yes

Thomas:

And then you graduated in 1958?

McGroddy:

’58 from college, yes.

Thomas:

And then you went directly on to the University of Maryland?

McGroddy:

Yes. And that was a very exciting time in science generally and in the United States. Sputnik was October ’57, and when that news came out, I was with a classmate; we went to visit… I don’t know what it was called at the time, but it was before Carnegie Tech and I think the University of Pittsburgh, but the thing that became Carnegie Mellon, one of the pieces of that, as a possible place to go to graduate school. I remember we’d driven out there, and “wow.” And then, we went back, at the college, and tried to detect the signals from that satellite, managed to do that.

But I wanted to get into something where I was doing stuff that I understood, and I could see that the transistor was coming, and I used to push them in the job I was in at RCA. They had these little cards with some of the electronics on it, but they weren’t printed circuit cards. They had wires on them and had little, tiny vacuum tubes, but no transistors. And the military’s always behind on that stuff anyway. But in the Philadelphia area then, there was a huge amount of technology. The first real big computers were built. Penn was involved, companies like Burroughs. There was a big Navy lab there, RCA where I worked, and half a dozen other companies for these co-op jobs.

So, a lot of the people that I went to school with in that program ended up in the digital world, or in academic physics in some cases. They did very well.

At the University of Maryland, physics was very pure. The head of the department at the time, a very good physicist, John Toll, was a theoretician who was on the fringes of the particle physics space. I remember a discussion one time, I don’t know what triggered it, but it was some seminar where somebody made a list, “What are the really fundamental questions in physics that we should be working on?” And there were things like, “Why do the particles that are out there have the charges and the masses that they do?” Still a very unresolved question. There were a few other such fundamental areas considered to be “true physics,” few things in that kind of category. There was only one guy I remember who worked in lasers. He used lasers to do something—he wasn’t developing lasers. There was no semiconductor physics or anything like that.

So, I did my graduate work there and did a thesis, a basic physics-type thesis. People were trying to understand the electronic structure of solids in those days, which is much better understood now. There’s a thing called the Fermi surface that people studied that relates to what the distribution of electrons in a metal looks like in momentum space. So, people were trying to develop tools to study the Fermi surface, and the guy I worked for, a guy named Ed Stern, had this idea that if you took plane-polarized light, reflected it off a metal, and had a big magnetic field there, that the plane of polarization would rotate a little bit, a very microscopic amount. And if you could measure that, you could learn something about the Fermi surface.

So, when it came time for me to pick a thesis advisor, I ended up with him. The alternative I had was another very good guy, Joe Weber, gravity-wave Joe Weber, who was actually an electrical engineer. He was trying to detect gravity waves. He hung these big aluminum bars up. He had one in Maryland on the golf course and one in Illinois someplace, and at one point he thought he had gotten synchronous hits on them, and he was a hero for a while. That was about 1968 or ’69. I spent a year in Denmark that I’ll tell you about later. I went to a talk he gave at the Bohr Institute. And he was a big hero. And then, other people questioned his results, redid the experiment, and failed to see any gravitons!

Stern had a graduate student try and do this magneto-optic experiment; he and Stern built some equipment in a laboratory, in a space that was in sort of the back rooms of a shopping center in College Park. You know the main College Park shopping center on Route 1? Probably not there anymore, but there was a hallway you’d go down that said “Ballroom” at the end of it. There was a big space back there that the National Security Agency rented, a predecessor to a larger lab they later set up in College Park, called the Laboratory for Physical Sciences, which is located not far from where you at AIP are.

Thomas:

It’s very close by, yeah.

McGroddy:

Yes. This was a predecessor of that, and it had a dozen or so physicists working there on a bunch of different things. And they used University of Maryland as a way to do business anonymously. It was basically a laundering arrangement. So, this guy was doing this experiment there, a guy named Bill Harte, and he couldn’t get it to work. He had worked on it for two or three years by the time I came along and eventually gave up. So, I took that over, and I got it to work. In the course of this work I also found a basic flaw in the theory, and redid that. For the latter period of my thesis work, Dr. Stern was on sabbatical in Israel at Technion, and we had very limited communication.

The experimental apparatus was a gadget where you’d have the light bounce back and forth between two parallel mirrors a lot of times to increase the total rotation on the light beam as it passed through the magnetic field in a 12-inch Varian electromagnet. And the things I did that made it work were mostly experimental things. I took the magnet, and instead of using as it was and just having it modulate its magnetic field, which it was capable of doing by a small percentage… It was a magnet with poles this size, magnetic field of about a tesla or so. I disconnected all the electronics and hooked it up so that the whole magnet went from zero to full power, up and down, very slow, with a bunch of electronics. And we used what was intended to be servo equipment to do that.

So, the signal got much bigger. And then, we built a different design of this thing to try and get even more signal. We were supposed to get, like, three times more in different geometry after we thought we knew how much rotation there was each time. It turned out, that we observed was not three times what we had ben measuring, but nearly zero! And for a couple months, that was a huge mystery and frustrating. And it kept me up at night, and I would communicate with Stern by mail in those days. You couldn’t afford a telephone call, no email.

Thomas:

Because he was in Israel.

McGroddy:

He was in Israel. And you’ll appreciate this as a Catholic. One night, I was going to the church that we belonged to there. Once a month, they had what they called Exposition of the Blessed Sacrament at night, and people would go spend an hour there, and you’d meditate. I still remember walking toward the church that night and suddenly seeing the answer and the theory, all laid out. And the point was that the quantity that we were measuring was not just a scalar number—but a vector, and the gadget that we had built, when you add the vectors up one after the other, it went around and around and came back almost to zero.

And once you knew it was zero, of course, you could use that, so we redid the theory in vector form. But it was very basic physics, Physical Review-type stuff. That’s where it was published. After I had finished my experiments that same equipment was used by me with another student of Dr. Stern to do measurements on other metals and alloys. I stayed on at Maryland for a year as postdoc, but I wanted to do something much more applied. I had a vague idea that I’d like to be in a university, but I also wanted it to be connected to some kind of applied physics or engineering.

Thomas:

If I can pause for just a second, I found the title of your thesis. It’s “Polar Reflection Faraday Effect in Gold and Silver.” Was there any significance…

McGroddy:

Well, that’s after. The thesis is in aluminum.

Thomas:

Oh, it was in aluminum?

McGroddy:

Yes.

Thomas:

I was going to ask if there was any significance in the metals that you were working with.

McGroddy:

Aluminum was believed to have a Fermi surface that looked pretty much like a sphere. And the question was to test that, and the aluminum was consistent with that. Silver and gold had Fermi surfaces that were spheres with sort of things sticking out of them. If you built a dodecahedron, you’d put them on a little square facet, and that would have a big effect on this Faraday effect. So, I think we did gold, and we did silver. And the question then… these skinny things that stuck out would have a big effect, and they varied between silver and gold, so we did silver and gold alloys to measure something about that. And that was another guy’s thesis, Archie McAlister.

So, we wrote a bunch of papers about that kind of stuff. But I didn’t view that as something that was going to open a new field or anything. So, when I was near finishing graduate school and looking for jobs, I interviewed at a number of what were the obvious places. There were all these government labs around that were mostly defense labs that had a piece that was relatively basic research. I guess the best of those, in many ways, was the Naval Research Laboratory. But the other ones had some jobs that were interesting jobs to do, and I also interviewed at a couple universities. University of Delaware, I liked a lot.

I didn’t do an interview with Bell Labs. Bell Labs was, in the industrial world, sort of number one by far in terms of reputation, and certainly their own view of themselves. And I knew some people from there and had run into people at physics meetings and whatnot, and I found them quite arrogant. So, I didn’t even interview there. They asked me to, and I didn’t. The way I connected to IBM was, there was a guy who was an adjunct at Maryland when I was in my middle graduate school years, a guy named Frank Stern who had worked at the Naval Ordnance Laboratory in White Oak. And he’d gone to IBM. And they were recruiting like crazy then because the company was growing rapidly after the introduction of the System 360 mainframes.

Thomas:

I take it he was no relation to Ed Stern.

McGroddy:

No relation, right. But that was the time at which the Watson Research Center had been built, was finished in about ‘61 or so. So, we’re talking about, say, ’64. They were staffing that up. They had the Research group kind of spread around in some buildings in Poughkeepsie, and there were people at Columbia who’d been in sort of a research division for a long time. But now, they had this grand new palace. It’s still a very good building. And a thing that I feel extremely good about is that it is still there and it’s still very much IBM Research. It’s not a shopping center, as opposed to, say, Holmdel [a former Bell Labs research facility], which is sad. We’ll get to that.

So, I interviewed at IBM, and I loved it. First of all, it was near my family. My parents were still living in Nyack, nearby. And the area was familiar. My wife’s parents lived in New Jersey, less than an hour away. And I liked the things I saw there. I had a choice of two groups. (And I saw in the interview that John Armstrong had, he had one of the groups in common as one of his choices of two when he joined. He joined a couple years before me.) One was in a laser and optics group, which was in the middle of all kinds of exciting things. But at that time, the injection laser had just been discovered and developed by Marshall Nathan, clearly the key guy in that. But there was also Peter Sorokin, who invented the dye laser and the first four-level laser. And there were people doing NMR and those sorts of things.

And the other option was to work for Marshall Nathan, and it was less obvious what we would do. But I liked him personally, and I liked what he had done, and that’s where I went. And he had a bunch of things that he had, in the back of his mind, noticed in the lab that he’d never followed up on. And I think I describe some of those in the paper that you have. He got me off to an enormously good start. I think the first thing I worked on there was, he said, “There’s a mystery about this compound, indium antimonide.” If you think about the important parts of the periodic table for semiconductor physics, there’s silicon there. So, it’s carbon (four electrons in the outermost shell; carbon, not a semiconductor), then silicon, then germanium, then tin and lead. But if you took one on each side and made a compound which had the same crystal structure: gallium arsenide, germanium’s between them. Indium antimonide, one level lower. So, indium antimonide’s a semiconductor. It has a very small band gap, like two-tenths of a volt. Silicon is like 1.1 volts. So, it was historically used for low-temperature sorts of things, for infrared detectors, where you need the band gap to match the photon energy. So, that would be around eight microns or so.

People’s infrared emission peaks at about… with our body temperature being what it is, around eight or nine microns. But there was a mystery about the following: He said, “With a diode, you can put a voltage on it such that the electric field in the junction part, where the p part and the n part meet, like, 10,000 volts per centimeter. And yet, if you take a bulk piece of material and put a couple hundred volts per centimeter on it, it blows up. That’s a mystery.” So, I ended up setting up a bunch of equipment… I actually have one piece of it still down in the garage in a box. IBM does not preserve its archives at all itself. They have only notebooks and calendars, which is frustrating.

I ended up doing what amounted to sort of plasma physics in a solid, a bulk piece of indium antimonide. Put an electric field on it, and it would generate a plasma of electrons and holes. And I studied the plasma generation rate and stuff like that. But it got off to an incredibly fast start. I joined in August of ’65 at IBM, and by the time papers were due for the next International Semiconductor Physics meeting, which is an every-two-years thing, I submitted a paper, and it was accepted. They didn’t let me go. Marshall was obviously my co-author on it. The meeting was in Kyoto, I think, in Japan. And then, he pointed me in other directions. There just turned out to be all kinds of interesting stuff there. [laughs] And it was sort of adjacent to what was then the germanium integrated circuits program, silicon, so it’s getting close.

And I think a thing that drove me—I know it drove me—from very early on was, I started to think about, “Why does IBM invest so much in this institution? What do they get out of it?” And it was clear at that point… when I joined, the IBM 360 series had been announced a year before, and there was just infinite demand and all kind of problems that Research was not working on to be able to produce what they call solid logic modules, which had discrete chip-level transistors put on a ceramic substrate. This is before any real integrated circuits. For reasons of performance, the transistors were bipolar, not MOS. My insecurity as driven by the fear that, “Someday, they’re going to look at it”—and I thought about myself— “they don’t need what I’m doing.”

Even though at one point early on, there was a board of directors meeting, which was held in the Research building, partly because the new corporate headquarters had not been finished, and some of the corporate headquarters people took over part of the Research building when it was first built. And they wanted to take them around, show them stuff, and they picked me as one of the people. I got to talk to them, show them, and made kind of a model to explain it with little ball bearings or something. But I really was into this idea, “Why are they doing this? I could end up out in the cold someday. They’ll say, ‘We don’t need this guy.’” Nobody ever told me that or gave any indications. It was not an issue in the culture at all. But it drove me through my whole career in one way or another, in that, if I go to the other end of my career, when I was head of Research, by then I had built a very clear model of how a properly operating research division was of enormous value to this company. That model pretty much didn’t exist, or if it did, was not visible to me! I think the beginnings of it certainly must have existed in the key top leaders… Mannie Piore—I never talked to him about it, so I don’t know.

I think it was more a blind faith kind of thing at that point, and the company was… we weren’t a monopoly, but we were dominant in our field. In that period, companies which were dominant in their fields looked at Bell Labs, which made a lot of hay with the laser and other things they’d done. “So, we should have one.” US Steel did it at one time, North American Aviation. In fact, I had interviewed at the North American Aviation one out in California, Thousand Oaks. Xerox.

Thomas:

Essentially, the attitude that, academics at least, tend to think of as the golden age of industrial research.

McGroddy:

It was, in many ways, yes, in that the company did not expect very much. And in fact—and I don’t know how much it showed, I haven’t read in detail what you dealt with John [Armstrong, in his AIP interview], and I think there may be one interview that’s missing from the sequence—but there’s this competition between the product people, development and manufacturing, and research that is built into the system that you have to do things to overcome in order to be effective. And it’s a thing that Bell Labs never succeeded at. Or PARC. There’s that famous book, what was it called, Fumbling the Future?

Thomas:

I actually don’t know if I know that one.

McGroddy:

They [Xerox PARC] invented lots of stuff, the mouse and the local area network, and largely failed to capitalize on those.

Thomas:

You’d mentioned earlier before we started recording that you were not working on silicon at all.

McGroddy:

At that time, I’d been told—I’m quite sure, I mean, I was told this, I don’t have any hard evidence—that, “Look, Fishkill and Burlington, they’ll take care of silicon. You guys look at an alternative.” So, work on III-V compounds, in germanium, that was okay. But that somehow changed not long after I joined, and they got in, in a limited way, into the silicon business. And the germanium integrated circuit program just disappeared completely. And ultimately—this took quite a few years, under Ralph Gomory, who was the director of Research for a long time—in his era, the idea was, we do some stuff that relates to what’s the mainstream technology now, reaching out for, in other ways affecting… He’s a mathematician, so he doesn’t have really strong instincts on technology, but he understands how a research organization should be run. And: “We should do something that’s a radical alternative.” And that was driven partly by the fact that there were a couple of cases in which the people at the top of IBM were surprised by something some other company did. And they wanted to know, “Why didn’t we do that?” Like, the first supercomputers. But I think there were other things, too, and they also realized it wasn’t research, but Research got involved quickly.

At one point, somebody had the idea about a disk drive. IBM surprised everybody with that, built the first disk drive, which was the size of two refrigerators. One had the electronics, and the others had the disks, which were 24 inches, and a half-horsepower motor to spin them. And that was five megabytes. That was a huge deal because before that memory was magnetic cores, for example, where you threaded wires through donut-shaped pieces of magnetic material. And memory bits were worth a dollar apiece or something like that. And then in 1968, at Yorktown, Bob Dennard—who died just a couple of months ago, he lived right up here in Croton, we were very close—he invented this silicon circuit. It didn’t have to be silicon, but a certain circuit in which you could store a memory bit with one transistor and one capacitor. Before that, there had been circuits developed that were six or eight transistors, you could store a bit. And that was DRAM for a while, people used that term, dynamic RAM—dynamic in the sense that if you turned the power off, or even turned the clock off, you’ll lose all the information very quickly.

IBM didn’t follow up aggressively on that. IBM got in the integrated circuits business, mostly on the bipolar side because that’s the technology that was fast enough to drive the mainframes. That was Fishkill. And then, they got into the field effect transistor, NMOS and then CMOS, in Burlington, a site with a laboratory, and a factory there, which built memory chips, not originally DRAM, but had built other FET chips for different systems. The industry at that time—

Thomas:

Fishkill and Burlington were both development centers?

McGroddy:

And factories.

Thomas:

And factories.

McGroddy:

Yeah. They were part of a division called the General Technology Division. It was IBM’s semiconductor company. The industry at that time was very vertically integrated. If you wanted to be in the computer business—I think still at that point, it was true—you had pretty much to be able to pull silicon crystals, and slice wafers, make transistors. There were companies… IBM had worked with Texas Instruments, and it would supply some of that stuff. But IBM pulled crystals, and sliced and polished wafers, and did the whole thing for many, many years, before Intel came along.

Intel had a huge impact on the industry. We’ll come to that. So, anyway, there I was, doing this quite basic research in an area that’s connected by being semiconductor and surrounded by people doing somewhat more practical things. Also doing some much more basic things. Having started in ’65, I was viewed as doing very well because I was publishing a lot of papers, and that was a key metric.

Thomas:

That was one thing I wanted to ask about is the emphasis on papers, versus patents, versus some other metric.

McGroddy:

We’ll get to that. It’s a very important point. They would file patents on these things, too, but it was somewhat of a side activity. The real metric in terms of what your peers thought of you, what management thought of you, what your potential was seen in you, at least in the basic research areas, was papers. So, I used to write a bunch of papers every year. By the time I was at IBM a couple of years, I’d written a bunch of papers. And the next time there was this big International Semiconductor Physics meeting, which had been going on for I don’t know how many years—a long time—I had, I think, two papers accepted at the next one, and I also got asked by Ian Gunn, who was a wonderful and different guy at IBM, to present a paper that he’d written, which had been accepted. He didn’t want to go.

That was the first time I ever went and visited Europe. The director of the Physical Science Department in Yorktown, which I was in, Seymour Keller, organized the trip. And I think it was—if I remember correctly who traveled together—me, and Seymour, and Leo Esaki, and somebody else, maybe Marshall Nathan. But Seymour arranged the tour, so we first went to England. I’d never been overseas before then. We did some classical touring, went to Stonehenge, checked out London, but also visited some institutions there in the field that I was in, a place that was very active was a place which was then called the Royal Radar Establishment in Great Malvern, a guy named Cyril Hilsum, who’s still well as far as I can tell, and we exchange correspondence occasionally—we drove out to the Cotswolds where RRE is located. Wonderful trip.

Then, we went to Paris. The student riots were on, Dany the Red, etc. And we stayed in the Left Bank somewhere near Saint-Germain and Saint-Michel. We had time to walk with the crowds, and smell some tear gas, and eat some great food. And I think the only institution we visited in France was the École Normale physics department, where I made some long-lasting connections. There were people doing things there related to my work at the time. And then, we went to Copenhagen and visited the Technical University there, which was sort of the applied research university there founded by Ørsted in the 1830s or so—met a bunch of people there and checked out some of the local area. And then, we went to Helsinki, and on to Leningrad—St. Petersburg—we came into Russia that way. And it was all very primitive then in terms of how customs worked and whatnot. That was still the Soviet Union. I remember them checking every little thing you had in your suitcase. Somebody had a Life magazine, and they took it away from them because it had advertisements for stuff that American people could buy. They didn’t want that in the country. There, I think I gave a talk at the Ioffe Institute. Again, met people who were doing… the field I was in by then had gotten to be a hot field, hot electrons.

And the meeting was in Moscow. So, I spent a week or so in Moscow, which was very interesting. At that point, you could get a car to take you any place in Moscow for a Kennedy 50-cent piece, which was silver then, because they wanted dollars. There were shops then that you could only shop in if you had dollars, Beriozka shops. China also had such stores until 20, 30 years ago. I met a lot of interesting people. And it opened my eyes to a lot of things. I had long been interested, before I’d gone to IBM… “Academia’s a nice life. Maybe I want to end up there.” And I’d actually had a feeler from Stanford at one point, but the American universities were in turmoil. What happened sometime after…

Thomas:

By that, you mean the political turmoil of ’68, ’69?

McGroddy:

Yes, the student turmoil. The students were rioting on pretty much every campus, certainly a place like Berkeley and Stanford, having started in Paris and elsewhere. So, I had no interest in joining an American university then. But about a year later, I got… I don’t know what formal correspondence it was, probably a letter—there was no email then—from the guy who was… I guess he was the chancellor of the place then, maybe later to be chancellor, Niels Meyer, inviting me to come as a visiting professor for a year. I talked to Seymour Keller, who was my boss’s boss at that time. I think by then, I had been promoted to a first-level manager, which meant you had four or five people working for you—

Thomas:

Were you still working for Marshall Nathan, or had that changed?

McGroddy:

I think either he was in parallel with me, or maybe he then worked for me, nominally. Anyway, Seymour said, “Why don’t you go?” I talked to my wife and family, I had little kids then, my oldest girl was going into second grade, and we decided to go for a sabbatical, one year. Seymour liked the arrangement very well, and—I only learned some of this later, in that the university paid me more than IBM was paying me. The deal I had was, I would send my check to him, and IBM would keep paying me. Simplified a lot of things. He used it to travel. He came to visit me three or four times in the year I was there. [laughs]

Thomas:

Was this a common thing at IBM?

McGroddy:

The sabbatical was a regular thing, yeah. In fact, to back up to when I first joined IBM, August of ‘65, I told you Frank Stern recruited me, and I picked between these two groups, and picked Marshall. I had to find a place to live. As a graduate student, then post-doc we lived in a little apartment—which I think is still there, I looked on Google and whatnot—on Riggs Road in West Hyattsville. I think it was $85 a month, very expensive. A little two-bedroom garden apartment. Still looks pretty nice from the picture. And we had a little bit of furniture. We’d bought a secondhand dining room table and some chairs, some other furniture… but the way it worked out was fantastic, because Frank Stern was going on sabbatical to the IBM lab in Zurich in August or thereabouts about the time I was joining. He lived in a nice house Pleasantville, quite near here, with his wife and I think they had one kid at that point.

And he offered to rent me his house for the period when he was gone, for the better part of a year, for I think it was $160 or $180 a month. He wanted somebody to take care of it, and I think IBM probably made up some difference. Likely, they didn’t because when I went five years later, people at Research didn’t know anything about international assignments. It turned out in IBM, where they were moving people around all the time to Tokyo, and Paris, and various places—they would set you up in Tokyo as if you were living in America, get you a great, big apartment, and pay your kids’ school, just tons of money. The deal I had was simply that I shipped my stuff, myself, I packed up a box that I still have—it’s a work bench up in our New Hampshire house now.

When Stern was away my last year, we had a big capital budget, and at some point, somebody came to me and said, “You haven’t spent any of the capital budget.” And I couldn’t communicate with him easily. So, I bought a bunch of stuff, including an infrared spectrophotometer, and it came in that box. I packed that full of stuff, and I actually drove the box to the pier myself, filled out the customs forms. IBM would’ve paid to do all that sort of thing. But in Denmark, the people there, Niels Meyer’s people, who were good people, arranged for me to rent a row house there, one of a string of them on a little dead-end street with maybe eight houses on that side and eight on the other side, where behind them, you were on the garden side, and each one had a garden behind it—for some modest amount of rental.

And we lived there for a year. I ordered a Volvo from the factory. There was a deal, you could go pick it up at the factory and drive it in Europe on the plate they gave you for a year, not have to pay any European taxes. So, we had a very interesting year there. So, I did that as a test of academia, and I had a bunch of equipment shipped, most of which got heavily damaged in shipment. So, I didn’t get a lot of laboratory work done while I was there. It was not a technically productive year, but what was good about it was, it took me out of IBM and gave me thinking time. Just because you change your environment, it forces you to rethink a lot of things.

Thomas:

Did you teach at all?

McGroddy:

Yes, I taught a special course in hot electron physics in semiconductors. And I actually started to write a book on the subject. Never got finished. My partner in this effort was Peter Price, a theoretician at IBM. I was teaching in English. I worked pretty hard at that because I didn’t have anything organized to just get out the old notes. And then, the other thing I did while I was there, I went to night school with a lot of the guest workers and learned the language. I felt that was a useful thing to do. I couldn’t communicate effectively people in the machine shop in English. Back then, far fewer people spoke English.

And my kids, my daughter Kathy—who ultimately got into biophysics, worked for IBM, retired from there, actually— was going into second grade, and she just went into the Danish public school. And it was interesting to observe that in six weeks… she was a very good reader at that point. In Denmark, they were just beginning teaching kids how to read. In six weeks, she could read Danish better than any of them because she only had to learn the language, they had to learn how to read. And she got fluent, with an accent, which she was very sensitive about. I remember one time, we were driving in our station wagon with two local friends, two twin girl friends of hers, and one of them called her a foreigner, du er fremmede, and she started to cry.

Amazing how sensitive people are. Kids don’t like to be different, and that’s why they learn languages well. Adults don’t want to seem stupid, so we don’t try. So, I learned Danish pretty well, and I still use it. Later on in my career, for the last 30, 35 years, I’ve been very involved in Denmark—less so the last few years—helping them start a new university and being an adviser to the minister of research. For a long time I maintained friendships and professional connections there. So, I’d go there once or more every year. Anyway, I was there, and I did one theoretical paper that I published for a guy from the Zurich lab. It was interesting, I had never met this guy, and I don’t know how we intellectually connected, there was no email… within IBM by that time, there were the beginnings of some form of email, which was very hard to use.

But somehow, we connected, and I guess we used to write each other letters back and forth. That’s the way the world worked. And we wrote a theoretical paper that is interesting and which, when I read it now, I can barely understand it. The theory in my thesis, the same thing’s true. And that thesis, it’s all written out by hand because you had to get your thesis typed, but the equations and all, you had to just do by hand. I still have copies of that, fortunately. And by the Christmas of that year, 1970, they had a very formal faculty party, very interesting black-tie kind of thing, and my wife and I went to that. And they moved you around from table to table for the different courses.

It was a very nice environment, living there. I bought a sailboat there and sailed in the Kattegat. And soon after that, they offered me a chair. They said, “You could have your own institute.” That’s the way things worked. If you were a full professor, you had an institute, a couple of junior faculty, some postdocs, and graduate students, and a good budget. And the finance for that was very simple, it all came from the government. In Denmark, even then, even when you’re studying for your master’s degree, you get a salary that’s determined by some government body in negotiation with the engineers union. There’s no such thing as tuition at state universities.

The offer led to two things. One was, my wife wasn’t enthusiastic about that. She had not learned the language, didn’t have many friends there. And for me, I had a big discussion with them about what I would work on if I had an institute. And I said, it’s this thing called optoelectronics happening. And they were trying to do silicon, had one institute in silicon—too small. I said, “That could never make a difference. That’s a world industry. But optoelectronics, you could make a difference. But we’d have to do it in a way that we would impact the economy, not just have a research group.” And they didn’t like that at all.

Thomas:

Had you worked on that at IBM at that point at all?

McGroddy:

No, I’d done no optoelectronics. But when I was away, I read a lot about it, I read the literature. But their model of the university at the time, and for a long time after that, was, you had these islands of expertise in different things, but you don’t try to impact the economy or do anything like that. In Maryland and the physics department, certainly, I had no connection with the economy, except for NSA and whatnot. So, I told them no and came back to IBM. And when I came back to IBM, I got into optoelectronics and my work became more applied and I was moved up in management. I was still doing research, publishing papers, some actually pretty close to some of John Armstrong’s stuff, some picosecond optics kinds of things, which was considered fast then. I guess you have to do attoseconds by now. And then, at a certain point, I got into much more applied stuff around the time that the people at Bell Labs first succeeded in making gallium arsenide lasers that could run at room temperature, and that was a huge deal. And IBM wasn’t doing it, so I organized a group around that.

Thomas:

Yeah, we have an older interview with Marshall Nathan, and he said that there was a period when IBM wasn’t doing that—I guess it was semiconductor lasers—and that you were the person who kind of revived that.

McGroddy:

Yes, I started it again. I said, “This is going to be important to us.” And I did a bunch of things over a period of years after that. I actually got IBM to buy into a joint venture company with Corning, which they had set up and wasn’t going anyplace. They had a German partner, I forget who it was, Siemens maybe. I thought, “We have these huge cables under the raised floor.” If you look at an old raised-floor building, the cables that connect a disk drive to a mainframe, they’re huge things. I said, “That should be fiber. We should build technology to do that.” And Corning had this little JV that was headed in that direction, and I bought into that. It was called PCO… O for optics, I’m sure; C was probably Corning; so probably the German company’s name started with a P. I got involved for the first time in corporate governance, joining the board of PCO.

What happened with that was—and this is relevant to this issue about how research and development connect or do not—as we started to succeed in that joint venture, an IBM group in Kingston set up to compete with us. And they were the insiders, so when we eventually went to serial optical fiber, they got the business, and PCO failed, partly because of this and partly because of poor management at PCO. It was all about credit, which is much too big an issue in most human organizations. Anyway, I came back, and I was doing that kind of optoelectronic work. I was still in the Physical Science Department, but I had this project that really fit in the other hardware-related department, Applied Research.

Thomas:

If we can put a pin in the optoelectronics for just a second, I’m wondering to what degree you would have interactions across the Research organization. You mentioned Dennard’s work on DRAM earlier.

McGroddy:

Across IBM or across Research?

Thomas:

Well, I would almost say across all sorts of different levels. So, across Research, to what degree would you know about San Jose and Zurich, for example?

McGroddy:

Research generally, in many things, there were different levels and different qualities of connection in different places. So, for example, on the West Coast in what was then called the San Jose lab, which was on the grounds of the disk drive division, the connections between Research and the Product Division were pretty good there because they’re physically right there, and they contributed directly. They would do physics and materials science that related, among other things, to disk drives. The people in San Jose also had good connection with the people in Fishkill in semiconductors in that they were very strong in organic chemistry, and they would invent photoresists that you need to do the integrated circuits. And a lot of the progress in the silicon technology depended on photolithography progress, as it still does.

So, it was very varied. But between Yorktown and Fishkill, as well as Yorktown and Burlington, the connections were poor. And in fact, for a period there in the early days, there was a division that was created, which actually had a laboratory halfway between Yorktown and Fishkill, which did some of that intermediate-level work, which in a way was bad. It just put another boundary in. But what happened in the optoelectronics piece was, as I said, this Kingston group started up, and this joint venture thing then didn’t develop enough outside business, and we shut it down.

At this point my career goals at IBM were shifting. I had done enough “bench” physics to have earned respect from my peers and from management, but I was still quite obsessed by this issue of what value Research really provided to the IBM company. Of course my understanding of the company and how it worked was really primitive. But my instinct was that we could be managing Research, or at least the part which I knew, in a way that would add much more value to IBM. This really revolved around how we set goals, what subjects we focused on, how we connected both within our division and to the “product” divisions, which was where the revenue was generated. I had moved up a bit in management, leading an effort of perhaps 30 or so great people. And I was aware that as both a bench scientist and a manager, there was a built-in conflict when it came to credit. Let me go back to my college years, when some of this was nucleated.

[I’d done that partly because if I go way back to when I was in college… I remember a serious discussion with one of the Jesuit professors, Father Gannon, during which he said to me,] told me, “You’ll be good at science and technology, but you’ll be better at working with people. That’s your real strength.” He had helped me, mostly by making key connections, set up a service company of sorts, of about a dozen of my fellow-students to manage setup and monitoring of audio equipment, mikes and the like, for banquets, lunches, and the like at the Bellevue Stratford Hotel, then one of Philadelphia’s best. The schedule of their needs was very irregular, so getting it done by a team of part-timers, at, I think, a dollar per hour, was a good deal for them. [laughs] So, we did this for a couple years, and I learned a bit about business and organizing people. And this idea that my skills might be better in dealing with the human dimension than with hard-core science was somehow embedded.

So, back to IBM, optoelectronics, and the like, that idea was always working background in my head, namely, that how people and organizations are set up and how the dynamics works is so critical to success, and it’s not just the technical piece that matters. I got more experience with helping to organize a team in my first two years in I went to graduate school, where I started as a teaching assistant, which was a very typical thing in the physics department. Right off the bat, they asked me to be the manager and organizer of all the teaching assistants for the large second year physics course for all engineering and chemistry majors. They were a diverse group, so unlike my college peers. At St. Joseph’s there were no women, there were no people of color, no Asians. It was a bunch of middle-class, mostly Catholic, American guys, of whom some were veterans from the Korean War. The year I started at Maryland, ’58, the school that had the most of its graduates entering the physics department was National Taiwan University. I think there were 10 of them! Overall it was a very diverse group of people from first-class undergraduate physics places. I found this working with people, trying to get the job done well by the team, quite satisfying.

So, to get back to IBM and optoelectronics… During my year in Denmark, optoelectronics, as well as the role of institutions such as research labs and universities in creating economic impact were things that interested me… and I read a lot, more than I had hoped to have time for, because it took a long time for the experimental equipment I had shipped there to arrive. I was reading all kinds of stuff, which I’ve always tried to do. I remember when I was a graduate student, I couldn’t read a book every week anymore because the coursework just kept me so busy. [laughs] That was a problem. But I had read somewhere about Corning, Corning Glass Works, in Corning, New York. Corning was, like the Ford Motor Company, a family-founded and multi-generation led family-run company. They had set it up with two classes of stocks, so the family maintained voting control without majority equity ownership. And the CEO was usually named Houghton. It was Jamie Houghton when I got to know it.

And, at that time, half their revenue came from joint ventures. They were, still are I think, masters at collaboration, having gotten started in a lot of their business by working with Edison to make the glass part of light bulbs. When television picture tubes came out, they got into that business. And they would do joint ventures with companies that wanted to make television sets, wanted to make picture tubes, and they would supply the “bottle” and the faceplate to make the picture tubes. So, they were good at managing joint ventures. And I thought, “That’s something IBM doesn’t know how to do.” And there’d been a couple of examples where they’d done joint ventures that had failed miserably. I concluded.

So, when I came back to IBM, to Research, and got involved with optoelectronics, the idea of working with others, collaboration, was on my mind. We talked earlier about how this played out, PCO, a joint venture with Corning, aimed at providing the sources and receivers to exploit the huge advances that Corning had made with single mode low-loss fiber.

Much of what I did in IBM—and I moved this way consciously, certainly starting before 1980—was aimed at making the organization work better, both internally and via joint ventures, driven at least in part, by this long-term fear from my early days, “Someday, they’re going to realize they don’t need us.”. This more than once involved convincing researchers to switch their focus, from something which had been very impactful, but which seemed less so now. Here’s one example.

There was a problem in the early days of integrated circuits of electro-migration. You had this aluminum device interconnection wiring that you put down, and the electrons actually moved the aluminum atoms around. And pretty soon, there’s a gap, the bulk aluminum having been pushed around. It’s called the electro-migration problem. So, the circuits were failing. And the way that got solved, by a guy named Francois D’Heurle. The way he would make his samples is, he had an electron beam evaporator, where he had a copper cup, he’d put aluminum in it, and he would shoot an electron beam at the aluminum, heat it up, and vaporize it, and it would then deposit to make the wires, and be processed with lithography, and then studied.

For one run, the electron beam was misaimed and hit the copper, mixed copper in with the aluminum. And when he did the electromigration tests, it didn’t migrate. So, the alloy, the small percentage of copper in the aluminum lines, solved this problem. That was worth a huge amount to the company. Found accidentally, clearly from Research, implemented in Fishkill and Burlington. That was done and no longer a problem, but Francois was still working in that area. Physicists often tend to stick with the thing they’re doing too long, I think, driven by familiarity, by investment in apparatus, by fields becoming popular. And I remember going to Francois—and I actually have a letter in my office that he sent me many years later—I said, “Francois, why don’t you stop doing the electromigration stuff and learn everything there is to know about metallic silicides?” These materials were beginning to be used then to make Schottky diodes and whatnot on silicon integrated circuits. And he ended up doing marvelous work, and had big impacts on technology, and became famous for that.

So that’s the kind of stuff I liked to do, was to connect something or to steer something, be more of a rudder than a propeller, staying in the background, not aimed at getting credit for it. Do background things that made the organization work better. By that point, I concluded that a big problem was that we physical scientists here in Research don’t know much at all about what the company does with stuff that might relate to or be impacted by our research, and that’s not good.

I started a program called “Silicon for Lunch”. We’d have a lunch meeting, buy your own, and a topic, and somebody would be assigned each time to go someplace in the company, and learn something that might be of shared interest, and I would push them in different directions. “Go find out what do they do in East Fishkill or some segment in East Fishkill. Come tell us about that.” And we did this for I don’t know how long, probably a few months.

Thomas:

What was your position at this time?

McGroddy:

I was, like, probably what would be called a second-level manager. The management structure was sort of first-line manager, second-line manager, and then director. So, Seymour Keller was the director of physical sciences. John Armstrong eventually became the director of physical sciences. Rolf Landauer, which you’ve probably heard of, he was for a while.

Thomas:

I wanted to ask you, before we get too far into the story, about Rolf Landauer and Art Anderson because they were involved at that management…

McGroddy:

Yes, Rolf Landauer, when I joined, was (I think) the director of physical sciences. He was a genius, in many ways, a wonderful, deep-thinking theoretician. Great man. But I don’t think he was a great people manager. And the environment was, you’re not trying to connect things. I came in with this idea, “We’ve got to connect.” Art Anderson, by then, was long gone. Art left about 1970, he wanted to live in California. I remember I only met him once early on. I was getting some award for some work I’d done, some internal award which Art was to present to me. And somehow, I caught poison ivy a couple days before. My face was all swollen up [laughs] meeting Art Anderson. Survived it. But what happened around then—this is late ’70s, ’76, ’77, Ralph Gomory had then been director of Research for six or seven years, and there was a study done quietly by some of these senior people, Seymour Keller and others. There was a guy named Hirsh Cohen that always worked with Ralph, who probably played a big role in this. Hirsh was Ralph’s right-hand man for lots of things, and when Ralph left IBM, Hirsh went with him.

Thomas:

To Sloan?

McGroddy:

Hirsh Cohen?

Thomas:

Right, he went with Ralph Gomory to the Sloan Foundation?

McGroddy:

Ralph, when he went to the Sloan, yes, he took two guys with him, or hired two guys there who had been close to him at IBM. The study earlier, 1977 or earlier, led them to conclude that they should reorganize part of Research, particularly the relationship between the Physical Sciences department, where the fundamental work was concentrated, and the Applied Research department. By then, I had, in the Physical Sciences department, a decent amount of physics and materials science oriented toward silicon. Some of them had oriented around silicon as a material, totally independent of me—the work of Alan Fowler and other people who were doing two-dimensional electron gas and that sort of thing. But they used very integrated circuit-like things as samples. And then, there was, by then, a silicon integrated circuits project from what had been a germanium-based one, grossly underfunded. At that time, it was barely beyond startup mode. And the other thing that had happened then was the Josephson project had started.

[Pause in interview for lunch]

McGroddy:

A big theme in my whole career has been trying to get organizations to work better, get people to work better with each other, to try and understand the dynamics that block good things from happening and work on those. And in doing a lot of that, it’s essential that you try to make yourself as invisible as possible. One thing I never did at Research was to be its spokesperson, and that was appropriate at that time. The guy who’s leading Research now, Dario Gil, is very effective as a spokesman, totally appropriate so many years later. He’s very effective in the presentation he gives about quantum computing, for example. Although, it hasn’t helped me understand it in any intuitive sense, which is a frustration I have. One of the things that I had to do in my career there, particularly after Gerstner came, was explain to that whole new team how business works, how computers work. Some of them wanted to go fairly deep into the technology, and I actually made a little course for them, particularly for the CFO, Jerry York, who was a terrific guy. Anyway, where do you want to pick up now?

Thomas:

Yeah, so we’re in the 1970s.

McGroddy:

And I think I was telling you, there was some study done—I never saw the study—for Ralph, no doubt, by Seymour Keller, and Hirsh Cohen, and a couple other senior advisor type people, and they decided to partially reorganize the lab and have a new department, which was one level below Ralph, that would be called Semiconductor Science and Technologies, to take the limited efforts we had in silicon and the science we had—some of which was materials science related to integrated circuits, some of it two-dimensional physics using silicon structures—put all that in one department separate from the Physical Science Department and the Applied Research Department, and they made me director of that. And I started that job in ’77.

One of the key things that I had been focusing on with this Silicon for Lunch thing, and as a result of what I, and we, learned there, this became much more urgent, namely, to connect with the parts of the company that did silicon development and manufacturing: Fishkill, Burlington, to some degree the overseas plants.

Thomas:

There’s an event that I’ve seen, I think, in a couple of different places, but I don’t know where to place it in time, where apparently there was a manufacturing problem, and Gomory essentially called for all hands on deck.

McGroddy:

That was about this time, a bit later. It began about 1978 or ’79, and I was one of the ones that helped get Research involved. IBM was coming out with a new generation of mainframes, called I think the 3081’s. We had the 360s, then we had the 370s, maybe these were called 380s. But the technology in them was a big step forward because it used chips with a few hundred circuits on them, bipolar chips. And they had built a module that was called a thermal conduction module, TCM, which was about 70 layers of ceramic material. The process involved taking a thin sheet of green, unfired ceramic, and printing wiring on it, and punching via holes in it, so you could go from one layer to the next. These sheets, with the patterned wiring and metallized paste in the vias, were then fired to make a solid, wired module for about 100 chips together. The chips were then mounted on the surface of this ceramic brick using controlled collapse of solder balls. Incredible complexity!

You couldn’t imagine a more robust thing once you succeeded in making the module, but there were all kinds of technical problems in being able to make that in volume. The ceramic modules were made in Fishkill, and the team there had to learn all kinds of things about ceramics. When you make something out of greensheet ceramics and then fire it, which is a very delicate process itself, it shrinks about 20% in linear dimensions. You had to build that shrinkage into the design and the firing process, which had to be precisely controlled. You ended up with this module with pins on the back, a whole bunch of pins, hundreds, and 100 or so chips on the front, and then nine of those things, TCM’s, Thermal Conduction Modules, plugged into a circuit board that was roughly three feet by two feet—big, again, many layers. That board, called a Clark board, was made out of fiberglass-like material and, again, had holes drilled through and filled, with patterned copper plating on it to provide the wiring. And each mainframe then was built from two of these boards, called Clark boards. As production began to be ramped, problems appeared in both of these technologies. Given the criticality of the problems, delaying the shipping of the new product at the core of IBM’s business, Frank Cary, the CEO, asked Gomory, could Research help? And I’m sure Ralph then talked to me, and I was one of the key players in organizing a way to work together with them in a structured way to resolve the crisis. Like a volunteer fire department, many people put aside their personal research and looked to find ways to fight the fire. This was firefighting!

One of the problems on the table was that after you made a Clark board, sometime later in testing, an internal short circuit would develop. And it turned out, ultimately, a Research guy named Jim Brownlow figured this one out. The problem was that when they made the fiberglass to make the boards, some of the fibers turned out to be hollow, and some of the etching and plating processes would leave electrolyte behind in the hollow fibers, which would later plate out because of leakage currents, creating a plated short circuit. Could be a long time later!

And somehow, Brownlow figured out—he was a wonderful guy—what was going on and somehow convinced the people who made the raw fiberglass material to change their process in a way that there were no hollow fibers, without ever being allowed to go to their plant and see what they actually did. He was one of the first IBM fellows and had, in some sense, fallen on hard times. There were people that discounted him. He earlier had become one of the first IBM Fellows because he had figured out how to use pill-making machines to make magnetic cores to thread for memory. And when I was there, I got to know him and appreciate him. I challenged him to do some other stuff, and he eventually had a whole other career. Wonderful guy. He’s gone a long time now. That was the biggest Clark board problem. There were a whole bunch of other smaller ones.

And with the ceramics, there were issues about controlling the shrinkage and just managing the whole process. Research people got involved and helped in various ways, and I think it was a good working relationship for a while and probably built some credit, but also some resentment for later. “Why did you have to call the fire department?” Certainly at the top of the company, Frank Cary appreciated what Research had done, getting the revenue stream of billions going again!

It's also important to note that many of the people at Research who got involved in these Firefighting efforts were really turned on by the experience. I suspect that many of them suffered deep down from the same insecurity I had, that “corporate” would someday take a hard look at what they were doing, and not see its value. They jumped on the problems with great energy and creativity, bringing unconventional skills and tools to solve problems that were a huge threat to the business.

And then, there was another crisis, not too much later, that was related to the same systems—I keep thinking 3081s, whatever they were. It turned out that the chips being made in Fishkill, when you did the testing, you’d suddenly have all kinds of errors appearing made in the logic chips. Everything was going great, and the next day the chips coming out were full of logic errors. And the errors were “soft errors”, transient in nature, but hugely problematic at the system level.

And on this one I got really deeply personally involved in. Ralph never got deeply involved in it. I don’t think it ever came to Frank Cary’s attention, or he was maybe gone by then. But that there was a Research guy named Jim Ziegler who was a wonderful guy I’ve lost touch with. And he was interested in radiation physics and measuring things related to ion implantation, and range and damage that occurs when you implant boron into silicon and whatnot. And I remember one Sunday night, during this crisis, I was up in Poughkeepsie at a meeting related to the problem, and having talked to Jim, at his suggestion stopping in Fishkill on my way south, and getting a bunch of what they call “carcasses”, which were wafers where all the good chips had been picked out, and bringing them to his house.

He was having a party around the swimming pool. He eventually figured out what was happening, and it was, again, a very subtle thing that Research people or a Research person was much more likely to resolve. What he figured out… even though the people in Fishkill absolutely forbade me from coming to Fishkill about that. They said, “You can’t come on the site.” And they didn’t like him either, but I managed to keep going at it with the bench level folks. It turned out that there were some people down lower who really wanted to resolve the problem and were less worried about the credit issue. What was happening was, in one of the solutions they used to etch the wafers at one process step, they used some acid mix that came from a supplier in big plastic jugs. It turned out that the manufacturer of the plastic jugs, after he made them, would wash them and then dry them, and deionize them by shooting the drying air at it through a nozzle around which was sort of sandpaper with some polonium-containing particles, which would ionize the air going through, so you could deionize the plastic jug. Occasionally, a particle from the ionizer, containing polonium, would go into the jug, be in the jug when it was filled with acid, and get dissolved in the acid. And this polonium emitted 5 MeV alpha particles at some reasonable rate, so all those chips would have traces of polonium on them, and periodically an alpha particle would hit a gate and flip it. And the gates didn’t have extra bits for error recovery, so they’d cause errors.

And Jim Ziegler figured that out, probably saved the company billions of dollars. Never got very much credit for it—it was never really publicized—and did it with them fighting him every step of the way. And he then did another one like that a couple years later. And this’ll come up in terms of what’s the value of Research to the company. There was a time not so many years after that where there was a problem that’s called the NTF problem, “no trouble found” problem. What occurred was that these mainframes were having failures, and the service engineer would go to the customer shop, and the way they fixed things was to unplug and replace these thermal conduction modules. They’re called thermal conduction modules because they had the plug on the bottom, 70 layers of ceramic, 100 chips, and then a cooling structure on top with a piston touching each chip and water running through that to cool everything. They would change modules, “throw parts at it.”

They had some way of guessing which module should change after the system crashed. And they would then—as IBM always did, a good thing to do—send the failed parts back to Poughkeepsie where they had been originally finished and tested. And what they were finding was that in so many cases, the modules would go back to Poughkeepsie and get tested, and there was nothing wrong with them: no trouble found. But it was driving the customers crazy because the systems were crashing left and right. And Jim figured this out… like I said, he’s a particle guy. One thing he got data on was, what’s the failure rate in systems in different places? It turned out, in high altitude, so in some locations in Colorado for example, the NTF rate was much higher. So, he rightly concluded that what was happening was cosmic rays were flipping bits. And the engineers, to deal with this sort of problem, assumed it’s got to do with static electricity or power surges, and they kept adding decoupling capacitors, and increased grounding, and such stuff. Eventually, they redesigned the circuit and put error correction in, put an extra bit in, a parity bit, so you could recognize, “This one has an error in it.” And if you use that correctly, if you do it in sort of two directions, in some sense, you can find exactly which bit it is and reverse it. So, if you do parity this way and parity that way, you get the intersection, and that’s the one that’s flipped from a zero to one. And that solved the problem. Again, saved billions.

Ziegler didn’t get much credit for it, although I believe, like to believe, that I made sure he got a good financial reward. And we kept that information within IBM for years for whatever reason. It was only many years later that Boeing got involved in using it. It was a particularly serious problem for IBM because you had so many circuits there that could fail and crash the system. But it turned out, there was a study done at one time—this would’ve been in the early ’90s, where I think it was a Russian cosmonaut who took a couple of ThinkPads on a space satellite—and there were all kind of soft failures. And it turned out, there was a likelihood of, like, 1 in 10 that if you used your PC going across the country, you’d get an error.

And I don’t think much error correction has been built into early laptops. You just do a reboot. Because of the way we use them, we tolerate that. In my [home] printer that could be the problem. I don’t know exactly how it’s stored, but it could be a bit in the firmware that’s gotten flipped. I find occasionally that I have to reinstall the firmware to resurrect it!

[Before this diversion into Firefighting, were covering the period around 1978 and on, in which I had become the Director of the new Semiconductor Science and Technology department. At that point the relationship between the Research and the senior people at the chip and packaging development and manufacturing divisions, was in my view pretty poor.] And John [Armstrong], in his [AIP] IBM interview, talks a lot about that. Because he went to Fishkill on assignment to get some experience outside the Research division in about ’84, ’85, some years later, for a year or two. And he saw that from the other side. And in part, it was arrogance on some people in Research’s fault. So, the guy who ran silicon integrated circuits research at that time when I took it over, Arnie Reisman, while technically very competent thought they were not very bright up there in Fishkill and Burlington. So there was very little, almost no, technology transfer.

There was a model, which had some truth to it, that if you got hired into Research, and after a couple of years, you weren’t doing very well, we’d try to transfer you to Fishkill. So, they were a dumping ground, which was not fair. A lot of people went there because they wanted to do something closer to product.

So, one of the things I started after I now had this combination new department—which reached from integrated circuits chips all the way down to basic science but had semiconductors as its theme—is I tried to get us to work with the two laboratories, Fishkill and Burlington. The technology we were doing in Research was FET technology, field effect, not bipolar. We soon began some bipolar. In addition, the chip line in Research was quite primitive. We were processing one-and-a-quarter inch wafers, and the process was FET. And we had a different process from what Burlington had, where the gate part of the field effect transistor was made out of polycrystalline silicon, and the Burlington manufacturing plant and lab used aluminum for the gates, so you couldn’t transfer incrementally much of anything.

The focus of chip development in those years was memory, DRAM. That’s what you used to push the technology along, and it was a lot of the volume. Logic was a derivative. IBM was late in DRAM, even though it was based on IBM’s invention! The big DRAM company was Intel, which only got into microprocessors in the early ’80s. At one point, there was a major discussion—I think it was for the 64K chip. The chips used to go up by factors of, typically, four in capacity, every couple of years— “should we use Intel’s process in Burlington?” They had a silicon gate process.

And IBM was gradually coming around to using silicon gate technology. So, I managed to set up a program, convince our key silicon guy, Arnie Reisman, and I don’t remember who was the key person in Burlington, but I found somebody that I could have good conversations with, to set up a program where we would work together, on something fairly far out, to limit conflict. And I think we had Fishkill in it, too, partly because they’re nearby, and we had begun to work on bipolar too. And we set up a program. We called it TRI-One, three groups of people working together on a common goal, namely to create an FET logic process at one-and-a-quarter microns lithography, which was way ahead of current manufacturing. And we did that for a year or more and were very successful, and it led to a bunch of papers mostly that kind of startled the chip world: “Look how far ahead IBM is.” Plenty of credit to be shared! This experience made it clear to me that that’s what we had to do is figure out how to connect these… do some things together at the interface to move things ahead so you don’t have a canyon that you jump over, some valley of death between research and development.

And then, a couple other things happened about that time. It was a very active period. IBM had a multi-year patent license agreement with a number of Japanese computer companies. This involved negotiating licenses, one-by-one, every five years. This involved a team of IBM people going there to see what the Japanese had and what they were likely to develop in the next couple years. There was a trip like that in ’78 that I was part of, and it was a Fishkill guy, Bob Henle, a very good guy—that’s probably who I worked with on TRI-One from Fishkill—and Ralph Gomory led the visit. And there were a couple of lawyers and business people. At that time, Japan had a national strategy, “We’re going to win in VLSI.” And the government, MITI, Ministry of International Trade and Industry, was the driver, “This company work on this, this one on that.” And government support, probably monitoring too, followed.

We were just amazed at what we saw when we were there, that they were so far ahead and going so fast. And relative to IBM Research in particular… I took Ralph to see the silicon line that we had before that trip. We were doing one-and-a-quarter-inch wafers. Josephson, the “radical alternative” was spending gobs and gobs of money on their process line. They actually had a piece of Fishkill—Fishkill people resented this like crazy—where they were building a pilot line for certainly many tens of millions, maybe $100 million.

Thomas:

They were building what?

McGroddy:

A pilot line to make Josephson chips. And in Research, that’s where the bulk of the capital money was going. It had I don’t know how many people, 200 people or something, working on it, whereas silicon maybe had 30 or 40. So, I took Ralph to see our line as a reference point. I didn’t know in detail what we’d see in Japan, but I was really frustrated by the primitive facilities we had in terms of wafer size. People adapted to it and were able to do this one-and-a-quarter-micron technology in spite of that.

And we got there, and we saw even little companies… well, companies that we didn’t think of as competitors, show up with these fantastic pilot lines, running 24 hours a day driving a national strategy, similar to what the Chinese do now. Industrial strategy makes a lot of sense actually, where you say, “What’s the next big thing that we ought to be in to make sure the country succeeds?” Not an unreasonable approach, the way the Japanese did it, with electric vehicles and a number of other fields, pharmaceuticals for example. We also ran into an approach they had for ensuring collaboration and rapid tech transfer. Ralph years later wrote a piece about joint programs. I can give you a copy now if I can get my copier to work. I can certainly give you the reference. On the plane on the way back from Japan, I wrote a piece which said, “We need to reorganize the way we work, structurally, not based on crises and one-shots.” TRI-One, I think we agreed to do for a year, or 18 months.

And I wrote out roughly how it should work, that we should have… not on the next-generation technology, that’s the development labs because they have to be introducing it to manufacturing, but the one after that, we should do together, and the work should have some protection from Research. If it’s in the budget of the development labs, when there’s a shortage of resources, you reach back to the one that’s one more back in the pipeline and take the people away. But we should figure some hybrid structure that we would build, which we ended up calling Joint Programs. I wrote this white paper on that and gave it to Ralph. I don’t think I gave it to him on the plane. I got back and crisped it up a bit.

The net of all this was, with Ralph’s contacts and some that I had—but more his, and his direct contact with Frank Cary, who was the CEO at that point, we put together a plan that said if the company was willing to make an additional investment of $50 million a year for the next couple years, we would commit to accelerate the introduction of the technology that we were going to work on together for both bipolar and FET by a year. The work would be executed within the Research management structure, but we would hire as the head of that program somebody that came from development, so it was very hybridized that way. And there would be pieces of the work in each of those two development labs as well. And that was called a Joint Program.

I presented the proposal to the Corporate Management Committee, which is where you got decisions of that size… there were four or five guys on it: Kuehler, Cary, the CFO, and two others. And they approved it. Bob Henle, from Fishkill, joined Research as the leader of what we named the Advanced Silicon Technology Lab. Eventually we built much better facilities than we had in Yorktown in Fishkill. The Yorktown building as a building wasn’t suitable in many ways. The ceiling heights and other things were a problem. The program was fantastically successful, delivered those next technologies earlier than promised, a year and a half ahead of the earlier schedule. IBM was then among the top players in semiconductors, in both FET and bipolar.

We all felt good about that. And ultimately, one-by-one, we extended that model, first to chip packaging… soon after we’d had this recent experience with TCM’s, and Clark boards, and then to disk drives, where we already had good connections. This joint program structure ultimately became the model organizational architecture for any program whose target was the product world. You together created a Joint Program with the product people, and they would pay maybe half the cost, to expand what otherwise might get done, you’d do the planning together, and the execution was primarily driven by Research. I’ll give you this article that Ralph wrote about it I think in ’89. [ed. Ralph E. Gomory, “Moving IBM’s Technology from Research to Development,” Research-Technology Management 32, no. 6 (1989): 27–32.]

In these years I got very involved in what was going on in the company in silicon. Wandered around, and I was trying to understand the world we were living in. One thing I found was that there was a program… this is pre-microprocessor days, not very far pre-. There was a program in IBM to make a set of chips—they called them microsystems chips, three, or four, or five chips—with which collectively, you could build a small computer or other things. Kingston had part of it, and Fishkill had part of it, and Atlanta, Georgia had a piece of it. And I concluded that that’s not going to work. It’s just got too many organizational boundaries in it.

And I don’t know in detail how I came to the conclusion, but I had learned enough about Intel to know that the first microprocessors were coming. I told you about this AIM-65 system, which I have down in my wine cellar, which has a Rockwell… I think it’s an 8-bit processor in it, Rockwell 6502. You could see what’s going to come. I had bought four or five of these, as well as other micro-gadgets, and one of my team made sure our key players could see some of what was certainly coming. And I was getting into this sort of, more than awareness, but sort of having it drive me, this idea 10 times improvement every five years. So, when you think about it, don’t think about something that’s impossible ever to do. It’s going to happen quick. I concluded IBM’s microsystems program was silly, and that Intel was likely to succeed, partly because they were neutral with respect to the market. Ours would only be used in IBM, and it was a remnant, and there were many, reflecting our genes as a vertically integrated company. Why would you use somebody else’s chips? We did use a few chips from other companies for a few things, but mostly designed our own. And it was before microprocessors. You didn’t have this pervasive thing that hit like an atom bomb with the PC in ’81. So, I started going around the company and telling people—I didn’t really understand how the top of the company worked, how important some of these senior people were, I just went to see them—and I said, “This Intel company’s in deep financial trouble and at risk of going bankrupt.”

Thomas:

I just want to be clear, this is while you were still director of Semiconductor Science and Technology?

McGroddy:

Yeah, I was still that. Not for long. And I told them, “What you really ought to do is buy half this Intel company and have them be the vehicle for making this thing that you’re trying to do with these microsystems chips, and have them succeed on the outside.” So, we’d only own half of it, but that would be a better arrangement of resources. And the retail price, if you just looked at their stock at that time, was $400 million. Not a lot of money. I made this proposal to a bunch of people. I believe—he never told me himself—Jack Kuehler was very much against this idea because all those other people worked for him, and he was very protective of his people. What eventually happened… two things happened as a result of that.

One is, IBM did end up buying 27% of Intel, but they kept it at arm’s length, didn’t kill the internal program, didn’t even connect IBM Research to Intel at that time. And the second thing was, Ralph, and some folks at CHQ, saw me as a change agent of some sort, and somebody decided I should be developed. So, they offered me a development job, which was a very interesting job, to be the secretary of this Corporate Management Committee, the group of four or five at the top. You sit in the meetings, mostly listening, and you organize the meetings, but you’re not a participant in them. You take notes. Not many notes. IBM was, because of the antitrust thing, very much against keeping notes.

In fact, when I took that job over, one of the first things somebody told me is to make a 31-day file. And when you come back to that day the next month, destroy everything that’s in it before you put anything new in it, because it would be discoverable for antitrust. My boss at that time at one point had been a witness in the antitrust case, which was still very active at that time, and they had some note, which had something to do with the creation of the department that I was running, and he claimed that even he couldn’t read it in the court. Interesting. So, the idea of no notes was a big deal. But I took that job, moved out of Research, and by then, within Research, I had stopped having my name on papers, or patents, or anything like that. I would try to steer people in different directions. And there were a number of them which led to very good results. There was a guy, Jimmy Tsang, who was doing some kind of Raman scattering experiments. Very subtle physics. And I said, “With that equipment, you could probably look at a seamless chip, and when the bits switch, see a little bit of light coming out from the hot electrons. And you could use that to balance the design of the chip.” You’d have a chip with… say it had 10,000 transistors on it. In testing it along the way, you’d turn it on and see which transistors light up. “They need to have a little bit longer channel length.” And then, you can run it to a higher voltage, and therefore, a higher clock rate. And he did that, and he got famous for that. Moved over into testing completely.

So, I’d do stuff like that. Sometimes, it’d be less precise than that. I remember there was a guy that I hired, I hope I remember his name… Marc Brodsky would remember it well because I think, officially, he reported to Brodsky. I said, “We don’t really have a job for you here. We have a career for you here. So, come to work for me for a month, and I’ll send you around to look at different things, and you go find the problem.” And that’s the guy who invented simulated thermal annealing. He had done something related to that, and he went to see people who were designing, trying to optimize the layout of wiring on chips. He figured out how to use Monte Carlo techniques to do that, and that led to simulated thermal annealing generally. His family had a business someplace out in the Midwest, and his father eventually insisted that he come back and run the business. Shame on me, I forget his name. Great man.

That’s what turned me on was to operate in that mode, where I could see the organizations working better… because one of the things that drove me through those years was, IBM is good for the world. I don’t know whether you ever saw this little book called A Business and Its Beliefs, it’s a collection of lectures that T. J. Watson Jr. gave after he retired at Columbia Business School. It talks about how IBM… it has two parts. The first part says, “If you want to run a great business, have your employees be really secure: retirement program, health benefits, vacation, a way to appeal decisions they view as unjust, etc. No layoffs.” And the second part of the book is about how the company should be committed to provide benefit to not just its shareholders—the Milton Friedman idea—but to its employees, its customers, its suppliers, and the community it lives in. Corning Glass Works is doing that. I imagine they still take the view they’re responsible for the economic health of that part of New York state. So, they wouldn’t move fiber optic manufacturing to Japan in a time when that was a big way to potentially save money.

And at IBM, we took that into account when there was a time in the ’80s when they thought they were going to double the size of the company. And in Hudson Valley, we had these places, “What percentage of the workforce works for us?” That’s when they opened up Kingston, but they opened up places in other parts of the world, Tucson and whatnot, so if you failed, the community wouldn’t fail. Now, it happened eventually anyway, but it was a model in which you make tradeoffs among the benefit to those constituencies, which for me is certainly morally the right approach, and it’s coming back a little bit now. Milton Friedman made the point, and it stuck, the only thing a corporation should try to maximize is its return to its shareholders. And you see what’s happening in healthcare these days, some of the horrible abuses. I don’t know whether you follow this Steward Health Care disaster.

Thomas:

I haven’t.

McGroddy:

You should look that up on the train. A heart surgeon in Boston took over as CEO of a bunch of small Catholic-founded hospitals that were in financial trouble. Did a deal with one of the private capital firms, stripped the real estate out, sold it to them. Hospitals had to rent it back at a very high rent, so they got a great return, but the hospitals now have cash, so they’re not going to go bankrupt. They did it 14 years ago. Situation now is, it’s all bankrupt, the whole thing. And in the meantime, Cerberus or whichever private capital firm it was, borrowed a ton of money and paid themselves a dividend, took back four times what they put in. This guy now has a 190-foot yacht and an 80-foot fishing boat, and runs this company which had two airplanes, because he keeps one of his boats in the Galapagos. And that’s active day-to-day now.

Anyway, I valued IBM. It’s good for the world. I said, “There are a couple hundred thousand people working here who have families. That’s a million people. This company’s good for the world. I should help it get better.” So, when I would occasionally get a feeler to go to some other company, an Intel or someplace like that, that was one of my reasons to simply say no. I also didn’t want to move. Made it pretty easy. So, I never was seriously tempted to jump ship. So, here I am in Armonk, secretary of this management committee. I did that job for about a year, learned a lot, got to meet a lot of people.

Because people who were going to present there would often want to talk to me about this or that to help themselves get ready, and I got to know the top people well. John Opel was the CEO by then. And they actually ended up with two committees, which didn’t work well. They undid it eventually. They had this Corporate Management Committee, which was three or four of them, and another one with six or eight called the Operations Committee. So, there were a lot of meetings. And those two committees actually had underground warfare with each other, and the individual members would look at things ahead of time and make all kinds of deals. Sounds probably a little bit like Congress. But I learned a lot there.

After that, I went out and was in the product world for about six years. My next job… there were two groups in the company at that time, product groups, one that had the mainframes, and the disk drives, and the semiconductors called the T Group, and there was one that did communications and smaller systems… like, there are these systems, which you never hear about, but which have been a huge success over the years, out of Rochester, Minnesota. It was called the C Group. I worked for a guy named Mike Armstrong, who eventually, after he retired from IBM, became the CEO of AT&T. I was director of product development there. And what I did mostly was to try and get people to build realistic product strategies. They were pressed in the planning process, given that we were going to double the size of the company in a couple years, to show that their product revenue was going to go up at 15 or 20% a year for the next five years.

Thomas:

Before we get too far into this, right before lunch we were going to go into the whole Josephson experience. And you’ve mentioned it kind of in the margins, but I think it’s an important subject.

McGroddy:

I was never deeply involved in Josephson myself, but when they created this department, SST (unfortunate name given the airplane’s fate), Josephson was part of that. But I was to manage it with the constraint that I neither kill it or shrink it for 18 months, or two years, or something.

Thomas:

Do you know year-wise when Josephson started?

McGroddy:

No, I don’t. But I think it was long before… I think its total duration was a decade or more. But the spending got big as it started to look promising.

Thomas:

And do you have a sense of when it got started, was there a big hype environment surrounding it? Is this one of these things that would’ve come from the corporate leadership saying you should get into that?

McGroddy:

No, I think there was a guy—the name that I think of is Will Anacker, as being a key guy—and then, there was a guy named Juri Matisoo, who then led it for many years. And there were a lot of very smart people in it, and they did a lot of very clever things to figure out how to make essentially a three-terminal device. To do logic, you really need three-terminal devices. And Josephson is a diode. How do you control whether the tunneling is big or little? So, they did lots of clever things. In its later years, it got a lot of money from NSA. NSA has always been interested in codebreaking and having the biggest computers for that reason. They had IBM build them, over the years, a number of what you would call supercomputers, and they have for years, right up the street from you [in College Park, Maryland], funded quantum computing.

When I visited there 20 years ago, they were big into quantum computing. It’s right on Route 1 there, the NSA place, the Lab for Physical Sciences. Maybe it’s not even called that now. But when I got in, it had gone on for a long time. It had been assessed by committees that were put together involving external people because there were more than a few people that thought it should be stopped. My view was, it was just disproportionate for something that was potentially very important, but clearly still very far in the future, that we’d gone overboard in funding it, in particular in my time, at the price of not doing silicon well.

Thomas:

The overall cost was something like $200 million?

McGroddy:

It was hundreds of millions. That’s a number you hear used. When Gomory went to Opel to tell him he was stopping, he told him it was $200 million down the drain. And then those people got redeployed into silicon. Matisoo eventually ran the silicon program for me. I took this guy Reisman out because of his difficulty dealing with people. So, that was sort of in my later days of my first session in Research. The department got started the end of ’77, I think. And I left sometime early or mid-’81 to go to Armonk. Did that for a year or so, then I was in this group working for Mike Armstrong, driving strategies. The key thing I got into there, which ended up having some impact on Research but a big impact on the company was to drive IBM to develop and later manufacture LCD displays.

Since about 1980, we had be annually updating what we called the “ten-year outlook”. For years it was an internal only, Research driven thing, the backdrop for strategy development. We’d ask, “What’s going to happen, and what should we do?” We’d look ahead. It was very technology focused. Ralph created that and drove it. And one thing I became convinced of by 1980, certainly, was that flat-panel displays were going to have an enormous impact on the world. Actually, before that, three or four years before, had a project to try and make a printer with reusable paper. I had this idea you’d get your newspaper over the telephone, print it out on paper which would be reusable, and we got it to work with faint pink on white paper with electro- chemistry. So, I was big into the idea of flat-panel displays.

And the other idea that I was big on, I’m sure having a lot to do with various pieces of history that you can’t explicitly connect together, is that the way the Japanese do technology is very different from the way we do it. And it’s mostly consumer electronics. They’re very quick, and the technology comes from the bottom up. It doesn’t drip down from the top from some very complicated thing to something simpler. They start in the consumer world, and they’re doing things that could be important in our world. And their methodologies are interesting. I had read whatever I could read about that. So, I took the position at one point when I was in this…

Thomas:

When you say that, is that essentially what they mean by use-inspired research? Or is there a difference?

McGroddy:

Well, I don’t know what they call it. I haven’t heard that term. I don’t think they even think about it as research. It’s driven from the marketplace. But I remember I gave a presentation when I was in this director job at some big annual internal IBM meeting about how we should figure out where we can do a partnership with Japanese consumer electronics companies because they do stuff different. They move much quicker. We’re very slow. And on a lot of little things. And I had worked on a lot of smaller, simpler products and technologies.

At this point, 1983 or so, we were still using printed circuit boards with what is called “ pin through a hole”, where you take components like resistors and capacitors, with wires on them, and you bend the wires 90 degrees, put them through holes in circuit boards, solder them from the back, as opposed to surface mount where the devices have no wire leads. We were way behind the state of the art on a lot of stuff like that.

I had concluded that large flat panel displays were sure to be developed over the next few years, probably first in Japa, small ones, for consumer electronics. The concept of a liquid crystal display had been developed years before. It came out of RCA. George Heilmeier invented it, black and white. It got used a little bit in wristwatches, mostly Japanese wristwatches.

But I concluded that they were going to get bigger, and the picture tube is going to go away. We have a three-dimensional object display a two-dimensional image. Big and heavy. So, I had in my head that, “We’ve got to get into the flat-panel display business.” I had run a strategy exercise with all the divisions that were in our group, asking them to show me the technical underpinnings that are going to convince the world to give you a revenue that grows at 15-20% a year. And most of them didn’t produce anything convincing. But I became convinced in that process that a flat-panel display would become an aggregation point for technology. We had terminals then, which were big bulky things. The display would be the center of these things.

IBM had actually started a flat-panel display program. They were plasma displays. I forget what gas was in it, but it was color, it was red. And the glass was a quarter-inch think, and there were two pieces of it. And they actually produced those in Kingston. The University of Illinois had a whole bunch of those at one time. Terrible thing. So, I said, “We’ve got to work with a Japanese consumer electronics company.” And eventually, people around the company, in particular, Research, got interested in, “Yeah, we should have a big program in flat-panel displays.” They had earlier done some work on electroluminescent displays, and had explored some other ones in technologies that didn’t work out.

Research came to the C Group, proposing to start a Joint Program, which they wanted the group to fund, with the people who did displays, CRT based, who were in Raleigh, NC. Ellen Hancock, who was one of Jack Kuehler’s people, led that division, which also did the proprietary network products. I turned down the proposal and insisted that were going to pursue this area with a Japanese consumer electronics partner. There were some big battles over this decision, but I persisted, and Mike Armstrong supported me. There was a guy on staff in the other group, the T Group, an IBM Japan guy named Seiji Ishida, good guy, who responded positively when I approached him. He became a good friend of mine as, along the way, I did a couple other things with him. IBM Japan was very enthusiastic about going this route. that. Ishida in particular was key. Together with Research we negotiated a joint R&D deal with Toshiba. And we saw along the way how some of the things get done so fast. The goal was to develop make six- to nine-inch diagonal color LCD displays.

The deal we set up was, Toshiba would set up an initial pilot line to do that because they were already working on tiny displays. And then, IBM would build a first-class pilot line in IBM space in Japan. That arrangement violated all kinds of IBM religious principles. “An IBM person’s going to work in a Toshiba facility? Toshiba people are going to come into an IBM lab?” When they eventually came to IBM, IBM set up a separate cafeteria for them. But the thing that was remarkable was, as we started to negotiate, I think it was April, spending, Japanese style, about two days talking to each other about family, and beliefs, and all this stuff. Because we were both trying to figure out, whether we will be able to work with these people who come from such a different industrial culture.

And we eventually signed a contract in July or August. I have a long document with the history of it. As we signed, we asked, “When will the pilot line be ready?” They said, “Six weeks.” It was pretty clear that they had apparently decided at the end of the first meetings in April, “This is a go. Let’s start to build the pilot line.” That’s the way it worked. IBM Japan already knew that, but it was a revelation to many of us. And a couple of years later. IBM Japan usually worked with a design and construction company called Shimizu. And when you’re discussing about the building, and it’s designed, but you’re haggling about price, and schedule, and all this stuff, when you finally sign the contract, they say, “The first prefabricated steel will start to arrive next week.” They order it on speculation. That’s a key element of how they move fast. They don’t wait until this step is finished before they take the next step. It is based on trust.

Through much of its life the effort with Toshiba was contentious inside IBM US. There were claims that we had the wrong partner, that we had chosen the wrong technology. It was liquid crystal, backlit, color filters, all on amorphous silicon drivers. Progress was very fast, and we were soon the leader in this rapidly advancing field. The three teams worked well together, IBM research in NY, the IBM Japan team, and Toshiba, and soon had a nine inch display prototype. There’s a very good book from the early ’80s by John Zysman and Steve Cohen. It’s called Manufacturing Matters. And the point about the book is, if you give up the manufacturing, move it far away, pretty soon you’re not going to be able to do development anymore. In our case, there was reluctance on many people in IBM to go to the joint venture a couple of us were proposing to do the manufacturing. Ultimately we went this way, setting up a company called Display Systems Technology, 50% IBM, 50% Toshiba, and built a big factory on an old Toshiba plant site. Elon Musk understands this point well, and puts the developers very close to the manufacturing line. Exactly right thing to do. The next step was the creation of IBM laptops, particularly the ThinkPad. And in the early days of laptops—’91, ’92, ’93—there was a worldwide shortage of displays. It was easy to ramp up everything else, but there weren’t enough displays to go around. But we had good supply from DTI, and that was one of the keys to the huge success of the ThinkPad, which I believe became a $7B annual revenue producer for IBM. We actually had a better display on our product than Toshiba did, because, the way the deal worked, DTI would develop a process—it’s a little bit like TSMC—but each partner designs their own product to fit that process, and IBM’s was designed to be a little bigger and a little brighter. We had, I think, a 10.4-inch display and others in the industry had nine inches.

Ultimately, long after I was gone, as the displays they’re building to make them bigger and bigger, television became the main target. I’ve never understood what hollowed out Japanese technology companies, but in the case of DTI, there was a divorce between IBM and Toshiba as laptop sized displays became a sideshow to the TV business. Looking back, I would never have believed that I would someday buy a 65-inch TV with essentially zero pixel defects for around a thousand dollars! Next, IBM sold its half of DTI to a Taiwan company, and Toshiba sold theirs to a Korean one. And pretty sure, between Taiwan and Korea, that’s where they mostly came from for many years. The US essentially never participated as a display producer as this huge industry grew.

Corning still makes a good fraction of the glass used to make LCD and OLED panels. And it’s remarkable… talk about partnerships. What goes into the plants, and this is a couple years old, is pieces of glass that are, I think, 10 feet diagonal and half-a-millimeter thick going into these factories to produce these 65-inch televisions. The other US participant is 3M. They make a lot of the layers, the color filters, the polarizing layers, things like that that are in the display, in film form. But there’s never been a commercial manufacturer of a flat-panel display in the United States, even though it’s a many-hundred-billion-dollar-a-year industry. People don’t talk about how we missed that. Outrageous.

And now, if you look at the journals, the articles, all the names on them are Asian. It’s also true in Europe. Nobody ever succeeded in Europe. Phillips tried one time, different technology, failed.

Thomas:

As you shifted from your research career into this aspect of things, there’s a lot to learn about markets and manufacturing.

McGroddy:

I was learning like crazy.

Thomas:

Would you pick that up around the company? How did you pick that up?

McGroddy:

Yes, I picked it up around the company and by reading, and by going to trade shows. I never pursued an MBA, or went to MIT and took any short course, or anything like that. I learned from people. Even when I got into the silicon business, I still remember asking—I think it was Bob Dennard, it might’ve been Dale Critchlow—to come and explain to me about static RAMs and dynamic RAMs, show me the circuit diagrams and whatnot. But I was always in high-speed learning mode. I have been almost all my life. Less so now, but I still try to learn something new regularly. I did that job for a couple of years

Thomas:

Coming back to the Joint Program, ultimately you end up with more than a dozen of these advanced technology laboratories.

McGroddy:

Yes, Research did, 15 or 20, maybe more.

Thomas:

Were you still overseeing the proliferation?

McGroddy:

No, there was no oversight of the proliferation. It got to be a part of the culture that if you’re a director in a department, and you have a project that is aimed at going to somebody that’s closer to the market, you’ll figure out how to do this. I built other types of connections along the way, some of them much later, to tie things together and preserve things. The next job I had, around 1985 or 6, I was in this T Group, in the chip and packaging division and I was director of development, I think the same name job that I had in this other group, which was a staff job, so you were supervising, not managing, reviewing, critiquing, poking people. And there, I worked for this guy Jack Bertram, who came down with Guillain-Barre and died, unfortunately, very young, maybe 58.

One of the big things that we did there early on when I got there… one of the key players was actually a very broad finance guy in that group, a guy named Sandy Kane. I still have some connection with him. But we realized from the competitive stuff… that the Japanese continued to succeed on this VLSI program. And we had worked with Toshiba. We weren’t yet into manufacturing with Toshiba with the flat-panel display collaboration. Japan was very much the enemy then, and they got caught stealing product and product information—both Hitachi and Fujitsu—from IBM. People went to jail for that. They stole these modules and whatnot that we had. So, there was great suspicion when I said, “I want to work with Japan.”

And Toshiba was a bad guy because they had produced milling machines that they sold to Russia, which enabled the Russians to make quieter propellers on their submarines. That was a public issue. So, by hooking with us, they partly solved that problem in terms of their image. But the thing we recognized was, in the semiconductor industry, the US is losing out. And a couple of us got deeply involved in creating this thing that got to be called SEMATECH. This new, government supported collaboration among many US semiconductor companies got them and us working together on the front end of process development, far before manufacturing implementation, and building a shared roadmap to drive the effort. The government ended up putting in, I forget how much it was, a couple hundred million a year. It was a very substantial and successful program.

Bob Noyce ended up being CEO of Sematech, which set up a lab in Austin, Texas. Back then, there were a lot of semiconductor companies. Motorola, for example, was a big semiconductor company. TI, which still is. But there were a whole bunch. HP made semiconductors. A lot of the computer companies still made their own chips because they wanted to be ahead of where Intel was. So, Sun Microsystems had SPARC chips, HP had their own processor, etc. And there was no real foundry. TSMC was coming into existence then, but it was way behind the power curve. They only figured that out later. So, the idea was to have this company where the American companies would share the front-end knowledge and build to a common base process that a company could sort of accept as is, or bring it in and soup it up, if they had some ideas.

And I think, to a large degree, that was key to the survival and success for many years of the American semiconductor industry. The Japanese would have totally eaten our lunch—very similar to China and the electric vehicles now. So, that was a big thing that we did. But IBM was among the leaders there. And the other thing that was true was that having the best technology, particularly in the mainframe world, which was still the dominant thing in terms of money. The PC had only come out in ’81, and there were minicomputers around; DEC was around and a few other companies were in that business, Wang Labs and Prime. You go to the Computer History Museum in California now, all these wonderful companies, so many are no longer with us. People did remarkable things, but it typically lasted only one generation.

We didn’t want that to happen to IBM. So, we were a key player in SEMATECH and kept pushing our technology along, but never used the Intel connection in any good way. What happened with Intel was that… I forget what year it was, ’85 or ’86, in the third quarter, the IBM finance people saw that they were going to miss Wall Street’s expectations by 15 cents or so. They sold the Intel stock, made a huge profit on it because it was passive investment, and that 15-cent gap never showed up. But I found out later—it was written up in an issue of Fortune, in a “Wintel” issue many years ago—that Bill Gates said that he went to Armonk at some point in that period and proposed to Opel, or Kuehler, or somebody that IBM buy a third of Microsoft. He remarked, “Of course they didn’t do it, they were just selling their Intel.” So, two interesting stories one could write, I could in principle write. One is one in which IBM recognized around PC time that we could do our own chips and do our own operating system. Had we done the PC that way, you wouldn’t have Compaq jump in two weeks into the business. And the other is one in which we owned half of Intel, and used it, and bought a third of Microsoft. The way I would write the first one, eventually the antitrust people would come along and we’d have to spit out two companies. We could call one Xtel and the other called Macrosoft. But it could’ve been a very different history.

The crazy thing that Jimmy Cannavino did in the deal that he did with Microsoft at the time, giving them the operating system, which… buying it from them, they didn’t even have it. Gates had to go find it. And IBM managed that whole period badly, agreed to do OS/2 with Microsoft. They never intended to have it succeed. And we actually at one point had the right to manufacture Intel 386 processors, and the deal was, they could define a test that the processor would have to pass to make sure it was consistent with their architecture, but we didn’t simply take their masks and build it.

And they kept diddling the test. They had to take some very artificial thing like this register, which doesn’t matter what’s in it. At a certain point, it has to have all zeroes in it. And it’s just, “Make sure you always flunk the test.” Not a one ever got manufactured. IBM was very naive about partnering with others and missed huge opportunity that way. But DTI with Toshiba one big exception. That worked very well for 10 years or so. So, I was out in this technology group world with packaging, and chips, and whatnot until it got to be around ’87. That’s when John Armstrong became director of research. At that point, I had my eyes set, and I’d talked to Bertram, “I’d like to lead Research. I think I could make it much more effective.”

Thomas:

Since we’re moving in that direction, would it be an opportune time to just ask a few questions?

McGroddy:

Sure.

Thomas:

First of all, what, in general, is your view of the whole SEMATECH approach to things? Or to what degree do you think that policy is a good way of approaching national—?

McGroddy:

Well, let me back up beyond SEMATECH. I told you about the 10-year outlook, which we did and it eventually became called the Technology Outlook. And we ended up sharing it with customers, and it had a lot of customer input in it. And we refreshed it each year. It didn’t start with a blank sheet. And historically, it was very good in terms of making us realize how fast things are moving. “You ought to be doing this because it’s going to be very possible soon, somebody’s going to do it.” Which is driven by the following idea.

I’ve given talks on this many times as well. If you think about the industry as a whole, the IT industry, let’s say it’s growing at 15% a year revenue-wise. If you divide the stuff that’s in it into… call one “old stuff” and one “new stuff”. If you start out at a certain point in time, everything, you say, “That’s old stuff.” And the old stuff, which dominates the industry at the start, grows at a certain rate, it’s growing at, say, 15%, but it flattens out. And over here’s a little piece, new stuff that doesn’t contribute to industry revenue in the early years, yet. But when you look back from five years out, and you say, “What is it out here now?” a good fraction of the revenue five years out, with the aggregate is growing 15%, is the new stuff. In 1980, PC’s didn’t exist, and by 1985 were important. Laptops didn’t exist, and then they were there. Same in the software world. Lots of things.

It’s almost always the case… and Clayton Christensen, I think, made this point in his book. I sent you a paper, I think, “Raising Mice in the Elephants’ Cage.” It’s very hard to get people to manage the new stuff into a business in an environment where your company is a big player in the old stuff. And that’s the IBM problem. And many other companies have had that problem. And IBM has, as far as I can tell, never really solved that problem. I started to write a book on that. I mentioned John Zysman and Steve Cohen. I was going to write a book with Steve Cohen. There’s a group called the Berkeley Roundtable on the International Economy—I don’t know if it exists anymore—that they were both part of part of, and I used to work with them, and they were somewhat involved in SEMATECH ideas. Eventually, Steve went off into corporate finance, and the book never got finished.

But the idea is, you need a mechanism to spit those things out. I used the word excubate. Don’t incubate them too long. Put them outside, and fund them, and maybe provide staff support so they don’t have to hire their own personnel department, or lawyers, or things like that. But take a chance that they’re going to fail. And DTI worked somewhat in that mode. For many years, it was a development project. Once it got to a certain point, it wasn’t old stuff anymore. And that company got to be, I think, $3 billion a year in revenue or so. So, that’s, I think, an unsolved problem for many companies. And you can obviously make mistakes in the other direction, of which Josephson is probably a good example. That was never excubated, but it just got too big for no evident value to anybody as a technology.

Thomas:

I was going to ask you about the Japanese experience with the fifth-generation computer. Is that what that was?

McGroddy:

I don’t resonate with the term fifth-generation. What year are you talking about?

Thomas:

This would’ve been in the 1980s.

McGroddy:

The Japanese government, MITI, told Hitachi and Fujitsu to go after the mainframe business and Toshiba to go after minicomputers. They laid out roles for them in the industry, and for Hitachi and Fujitsu it was obviously, “Compete with IBM.” And they made clones, and those clones were around for a long time and were very successful. They were plug-compatible, and then, of course, Gene Amdahl did the same thing on his own. And then, in the disk drive world, lots of people left IBM and made all kinds of disk drives. It was part of the antitrust settlement that IBM had to expose the specifications of the connectivity. If you made a disk drive, what do you have to do to have it plug into an IBM mainframe? So, that model was out there for a very long time.

But getting back to this 10-year outlook thing, I expect that MITI did something like that, and I think the Chinese government… I think a lot of places do. And we don’t do it in this country. In those cases the national government develops a view of what might happen in the world in the next 5 to 10 years, and where might we, at a national level… not that we, the government, are going to execute on it, but we identify what holes might we let get created that are going to leave us falling behind in some critical areas. Where the Chinese presumably said… in 1990, they had no car industry. They eventually had two companies come in, Jeep went into Beijing and, with old tooling, installed it, made the Beijing Jeep. And Volkswagen went to Shanghai, did an equivalent thing. Buick eventually did it, too. But they presumably looked ahead and said, “Electric cars are going to happen. What should we do to be getting ready for that? What should we make sure some companies in our country are doing?” And in China, of course, it’s very simple because they were at that time mostly state-owned enterprises, so you can tell them to do it. MITI had not that same level of control, but not terribly much diminished from that.

In this country, we have very little of that other than in defense technology. There was a time in this country, more than one time, where big defense development and research people used to worry about, “We might not be able to get the components to make the hardware we need. We must make sure we can get it made in the United States.” So, there are a couple semiconductor factories still around. I think the former IBM Burlington plant maybe one of them. In New York the state, and the feds, have invested huge sums of money in R&D and I think in Global Foundries aimed at ensuring secure manufacture of radiation-hard CMOS.

Thomas:

Is this Trusted Foundry?

McGroddy:

Trusted Foundries. And the trust is the key thing. Morris Chang makes this point all the time, that the biggest asset he has is the trust of his customers. They trust him with their designs because he doesn’t produce any products that compete with them. He started at least one other company that does, Vanguard Semiconductor, but it doesn’t share TSMC customer information. They’re trusted.

Back to the idea of government involvement, creating roadmaps, and not missing out in key technologies. In the flat panel display case, at the display module level, the US, Europe too, missed out almost completely. And if you take the idea that for every $200,000 or so of volume, that’s a job. A lot of people have jobs working in a flat-panel display factories in Asia. And we missed it. Doing such planning and promotion at a national level would in my mind be a very good thing to do, but it would likely be highly resisted because of the way the American system works. And the flip side of that is, we’re very good at starting many new things, but a lot of them never make it to manufacturing.

Thomas:

There are a few people at IBM… I don’t know to what degree you had interactions with them or not. First of all, there’s Erich Bloch, who went to NSF.

McGroddy:

Yes, he led NSF. He was an engineer, not a scientist.

Thomas:

Yeah. I’m kind of interested in the degree to which he brought insights from IBM to NSF. He started off the Engineering Research Centers.

McGroddy:

The biggest thing that I saw—and of course, I do not have an incredibly broad scope here—was the founding of the NSFNET, 1987, when a couple of the supercomputing labs, Cornell and others, had an ARPANET-like network at—I forget if it was 16 or 64 kilobits a second—connecting them, and they wanted to have more bandwidth connecting them because the files they create are huge. Somebody convinced Erich, I’m not sure who, but I think the fact that he had the background that he had enabled him to accept it, to fund the NSFNET, which was a two-year project, not for the government to own, but for a company to provide, communications service linking a dozen or so labs. The initial NSFNET showed its telephone roots in its speed being T1, one-and-a-half megabits a second, and NSF put out a request for proposals.

So, I give him tremendous credit for doing that. I had known him at IBM, stayed in touch with him over the years. He was a very feet-on-the-ground kind of guy. He’s not a high-energy physicist. There’s a lot of stuff NSF funds that he probably didn’t understand very well at all. But I think he made the agency effective in some ways that a different person would not have. And then, of course, the history of NSFNET was that IBM ended up playing a key role in implementing NSFNET. The key IBM guy was Allan Weis, who was then the director of Computing Systems at Research. The Computing Systems often were fighting with the Computer Science people. Computing systems was in principle charged with providing the computing infrastructure for the Research division. But they had bright, ambitious people. They always hired the best. So they did a lot of research. And they bid on NSFNET together with MCI and Merit, and that team won the bid and delivered NSFNET in less than a year.

Concerning NSF and Erich Bloch, I don’t know, I’m sure people have studied what happened to the distribution, the mix of what NSF funded on his watch as compared to before. But the NSFNET one really sticks out. Al Weis and his team put it together using a PC as the base to build a router. The IBM PC had a very fast back channel called Micro Channel. They figured out how to make the PC’s act as routers where they routed over the Micro Channel. Then the NSFNET traffic just exploded, 20% a month, for a long time. Because everybody started to do email on it, and all the universities connected their local networks onto that. So, it ran out of capacity, even though it was an increase from, say, 64 kilobits to 1,500 kilobits—more than 20X. NSF then put out an RFP to raise the bandwidth to T3, which was 45 Mbps, another factor of 30. And the same group bid on it, won the upgrade bid, and delivered on that. There were no routers for that, so they used the same PCs, but the Micro Channel wasn’t fast enough to ship the bits from one card to another. It had all these flat cables around the back of it to do that.

And Weis personally tried to convince Ellen Hancock, who ran the Communications Division, which had managed the IBM proprietary network technology—it was called SNA, Systems Network Architecture, which was a couple-billion-dollar-a-year business then, big, heavy hardware. It was very much a store-and-forward, step-by-step, the whole file. Get it to here, check it all out, take another step. Whereas the NSF just ships a package, piece by piece, by whatever route they can get there. You reassemble them at the end, and if it doesn’t work, you do it over. So, it’s much more immune to local failures, which is why ARPA was originally interested in it, in packet switching. So, he tried to convince her to go into the router business. And I covered this in that paper I sent you, “Raising Mice …”. They were trying to raise a mouse in the elephants’ cage. And the elephant wants to step on the mice because they irritate it, or it accidentally poops on them. But in either case, it kills them. As in many other companies, this Internet thing was seen as a threat to huge existing businesses.

And the remarkable thing about the internet, I think, is that it’s creation was driven by people who needed it, these supercomputer centers initially, and the telecom companies, who presumably viewed that as a tiny, niche thing, didn’t pay any attention to it. Those folks, ATT, IBM, DEC, DT, etc. were all fighting in the standards bodies because they each had their own proprietary thing. IBM had SNA, DEC had DECnet, AT&T had their own kind of packet switching. And they were optimized in different ways. Like, for voice, where you want really tiny packets, and for data files, you want big packets. And standards people were just haggling with each other year after year. And NSFNET and its successors just zipped right past them, and they didn’t even notice it. IBM didn’t notice it even though at one point, not too far in, companies started to approach the entity that Weis’s team had set up, Advanced Networks and Services, ANS, saying, “Hey, we’d like to use that sort of thing. We have a bunch of branches of our enterprise here and there,” many IBM customers who were using very expensive IBM stuff. And ANS then set up was a for-profit subsidiary of the not-for-profit that could do for-profit business, a very awkward structure, because among other things, in that structure, there’s no easy way to raise capital. So, the customers that wanted to get on the network would have to put up something like two years of their expected fees to provide the cash to buy the equipment to set up to points of presence. And there was huge demand, which threatened ANS with bankruptcy! By that time, I had joined the ANS board, and what we did eventually—and Al Weis deserves a lot of credit for what he did—was to put the for-profit piece up for sale. It was sold it to America Online for $40 million. Unfortunately, we had banking advisers—JP Morgan, I think they were—who said, “You’ve got to take that money that’s now in the not-for-profit and, a) you’ve got to find a mission for the not-for-profit,” which we figured out pretty quick, “and b) you have to diversify the portfolio.” If we’d kept the AOL stock, it went through the roof. But we had this $40 million and got ourselves into the business of trying to get the internet effectively used in K–12 education and did that for 15 or so years, had a contest with kids doing things toward this goal all over the world, prizes given, million dollars a year in prizes, and eventually handed that off to the Oracle Foundation. As it turns out, it’s almost impossible to shut down not-for-profits. But then, I was the chairman of ANS, which ended up being just me and Al. Eventually, we had lawyers wanting to help, and we didn’t have the money or didn’t want to spend the money on doing that. We decided, after filing reams of paperwork, to simply declare it dead.” We did that, and it’s long enough now that we’re protected probably by the statute of limitations. And by now, Al’s passed on, unfortunately.

But this issue about when you do something in Research that is new stuff, how do you get that to be valuable to the company? IBM certainly never solved that problem. And the way I used to deal with it—I would never say I solved the problem—to some degree, we did things, particularly in my later years in development… I went back to Research in ’89, but in, say, ’85, ’86, ’87, that time period, I started to push for us to be doing a better job at that. I’d done the flat-panel display thing, which at that point was still not an obvious success. The first ThinkPad only came out in ’92, so that’s when it probably started to show. And even then, people would bitch about the price, but I think it’s clear that IBM wouldn’t even have had laptops. And it’s still the ThinkPad, which they unfortunately sold to Lenovo, a premier business brand.

I was working on that problem on other fronts. One thing I did in that period in the year or two before I went back to Research, IBM, at that point, and for many years, had never forced people who used our patents to pay any money for licenses. So, companies like Intel, Motorola, who made DRAMs—IBM didn’t get a nickel out of them. And that approach was, I think, off the back of the antitrust suit, which was settled in about ’82, where there was huge fear in the company of another antitrust suit. That’s why they kept Intel at arm’s length. I had long considered that the whole way we did patents was crazy. We had a system that had a formal disclosure form that you filled out: “I’ve invented this thing”, and you had notebooks that you got signed every once in a while by a witness. And then, there was a process to decide whether to patent the idea or to publish it, which would prevent other people from patenting it. And the way they published it, they actually had a journal, called the IBM Technical Disclosure Bulletin—big, thick books they’d publish, with all these disclosures in it, these ideas in it. It seemed ready for change to me.

Many people figured out how to exploit the system, which rewarded engineers, and scientists, and inventors. They got one point if it was decided to publish your disclosure, which demanded very little work on your part, and three points if they filed for a patent. That was a lot of work. You had to work with the patent lawyer to write a good patent application. And there were people who concluded, “Well, let’s just write disclosures and agree to publish them.” There was one guy who actually had his son do this in summers with him. They wrote a patent disclosure every day, and the attorney they worked with was so overwhelmed, he would just publish it. So, they got a point every day. This guy had a huge number of points. And you got money for your points at certain levels. You’d get a couple thousand dollars for perhaps a dozen disclosures.

I started to try and change that in the late 1980’s and when I went back to Research in ’89, I pretty much unilaterally changed the system for the Research Division. My goal was to require to increase the number of filings and issued patents, and not reward patent publications alone. Publications, we should just publish them in some random place and keep track of where we published them. Publish it in the West Jutland Technical Journal in Danish, and that way, you don’t give away curated collections of ideas, and you can always, if it comes up in a patent issue, say, “Here, it was published years ago.” And we should get more patents, and we should also insist on licensing. I put that in place in Research and convinced the corporation to start licensing things. And there was a thing that was created around that time, the IBM Academy of Technology, a senior group somewhat similar to the US national Academies, advisory in nature. They did a study of what I had done in patenting, because people in Fishkill and other places were not happy that Research was different. And that group concluded we should do the same across the whole company, and we did, about 1989.

And having done that turned out to be very important when Lou Gerstner came in 1993, because this approach had gotten the company to be reasonably aggressive in filing patent applications and insisting on people who were using our intellectual property to pay us the typical licensing rates. And by the time Gerstner came, that was producing about a billion dollars a year at the bottom line. And in my first one-on-one with Lou—we can go over it orally, too— I was trying to explain to Gerstner that, “Research is a wonderful investment for IBM. Specifically, IBM as a whole gets a return on it every year that’s a multiple of what is invested. How could it be better than that? And I think I can prove this to you quantitatively.” So, for the patents, I got the people in corporate to give me an estimate of what percentage of the patent licensing money comes from Research-related patents, and they said around 40%. We’re getting a billion a year. If there’s $400 million a year, our budget’s 500. 80% of it’s paid for right there which is not the core purpose or the biggest piece of value that we deliver.”

So, I was a change agent all the time in IBM. Anyway, in ’85 or thereabouts, ’86, I wanted to be the next head of Research. We knew when Gomory was going to leave because IBM had a rule, sometimes called the Learson Rule, that in the top positions, when you’re 60, you leave. And you had to sign papers all the time when you took a job like that, confirming that you’ll give it up by that age. You didn’t have to leave the company, you had to give up that job. But then, of course, your salary would go way down, and your retirement pay depended on your salary, so nobody would do that. So, we knew when Gomory was going to leave.

Thomas:

He’d spent an extraordinary amount of time in the position as well, since 1970.

McGroddy:

He was in there for more than 15 years, I think. And then, what happened was, at that time, they also had a job called Chief Scientist, which was Lew Branscomb for quite a few years, and he ran a staff operation in Armonk that didn’t have much direct effect on anything. He used to bring in this outside committee, the Scientific Advisory Committee, and they would show off to each other around a table with some inside people and outside people. I used to tell them, I thought what we really ought to do was start the meeting, put a ruler on the table, let everybody lay their organ on it, we’ll decide whose is biggest, then we can have a real meeting. Lew had to leave when he turned 60, and Ralph took Lew’s job.

So, that’s what took Ralph out of Research. He wasn’t happy about that, I don’t think. So, he left Research before he was 60. When he went into that job, they needed a new director of research. That must’ve been about ’86 or ’87. And I had certainly let it be known that I would love to get that job. I still remember a senior VP, Dean Phypers who was always going [impersonating guttural noises]: “we’ve decided, it’s going to be John Armstrong.” And they tried to put me in a new position, vice president for education. They had a collection of stuff, training salespeople—other kinds of investment in training, costing in all about $2 billion a year, and I told him that I don’t want to do that job. I said, “I want to just stay where I am or leave the company.” And I stuck, I stayed where I was for about another two years, about, during which John was the Director of Research.

And then, Ralph turned 60, so he left, went to Sloan, and John got moved into that Armonk position. That would’ve been ’89. I was made head of Research then. And I was raring to go to do that, and I knew a lot about what was going on there, and I had a lot of things I didn’t agree with, mostly related to their ability to impact the company. I started to work very hard on understanding in a more crisp way, “Why is Research valuable to IBM?” I gave a lot of talks about it, both inside and some outside. I also met with some recent Research Directors from such places as GE. My net was that a simple way to capture what I thought should be the case was that IBM invests in research to win in the marketplace. And I ended up capturing this idea, and its underlying elements in the packet I sent you. There’s a chart in there that I made.

Thomas:

Is that the arrow chart?

McGroddy:

Right. Ralph had laid in, in the early ’80s, a very good fundamental goal. “We must be a research division which is famous for its science and technology and vital to IBM.” And when he talked about it, he made the point that it was an inclusive “and”, not separately, together. And I tried to flesh that out, so that’s at the root of that arrow. And the top of the arrow is a list of seven or so things of major and visible value to IBM that you can quantify. So, talk about leadership in our products and technology. By then, particularly with the Joint Programs, we had a very good record in silicon, with the stuff we’d done in the disk drive world, particularly in the magnetoresistive heads, which put IBM way out in front of everybody else. And then, many years of software. I said, “We don’t pay any licensing money for any of that revenue.” And I think patents was second. But there were a bunch of others. So, we’d bring great people into the company who wouldn’t join IBM otherwise, and then many of them go other places and do great things.

Thomas:

Yes, I have your document here.

McGroddy:

Does it have firefighting on it or not?

Thomas:

Firefighting is not on this particular one, but you mentioned it.

McGroddy:

I sometimes put it on because can be irritative, and is more than a bit proprietary. But mostly when I talk present this, I would include firefighting as part of it.

Thomas:

Should I just read them quickly? “Technical leadership, patents, marketing support, vision, people, luster.”

McGroddy:

People like to do business with a company that has Nobel laureates in it. The honor reflects on how we are in it for the long run. And the company has luster in other ways, too. But it’s recognized as a positive thing.

In terms of value and impact one of the big things I did after I went back to Research, shortly after, is to get Research much more connected to customers. Let me tell you the story how this came about, how it triggered it. I used to like to visit customers. I was doing a visit, along with Abe Peled, the director of our computer science efforts. We were at Citibank headquarters on Park Avenue just north of Grand Central. And their CTO, the guy we were visiting, was Colin Crook. Great name for a banker. He’d actually been a semiconductor guy at one point, so we hit it off. At one point in the discussion he made the point, sketching on a whiteboard, that, “All this stuff, if you look at the value chain starting at the base with science, and raw silicon, and integrated circuits, and chips, and all that—for us, that’s a commodity. What we want is, right up here at the top, you have to help me compete with him, Chase, across Park Avenue. If you don’t do that, I won’t have any preference for IBM products. That’s where you distinguish yourself.”

And the marketing teams, ideally, and Research can help them a lot, can do that. You have a private relationship with this client and that client. You’re always going to get into their drawers in a pretty deep way. And off the back of that, I thought about this goal, and vital to IBM, I said, “How much of what we do fits up here near the top that Colin Crook was pointing to?” And I had a rough assessment done by the staff, and it was a couple of percent. There were a few things in the math department that they were doing to help American Airlines. Operations research, the kind of stuff that Ralph did. But not much else ….

Thomas:

Right, right. Linear programming.

McGroddy:

Yeah. There’s a schedule. Particularly after an outage, how do you get the planes back in the space in the optimum way to reinstitute the schedule? So, we had some of that. And I had been pushing some of that stuff. The math guys particularly were valuable. There were a couple of cases where what they did was really valuable to customers and changed the whole relationship.

But it was only a couple percent of what we did. And I concluded then and there, this was probably ’90, that was ridiculous. If we’re vital to IBM, you can’t be at 3% or whatever it was. It’s got to be at least 20%. And we’ve got to do it fast. So, I cooked up a plan with my staff. The goal was set to have a new sector in Research. It’s going to be called Services, Applications, and Solutions. Customer-focused. And within two years, that’s got to be 20% of our effort.” And what I told people was that there was a lot of what we were doing that had fit well in terms of this arrow chart, and the nature of the value in our industry when it was younger but didn’t differentiate very well anymore. So, a lot of the work that we’re doing, in III-V compounds, for example, we were leaders early on, getting started. But five years in it, everybody’s doing it, and it is increasingly unlikely to impact our mainframes by displacing silicon.

And people would come back and say, “But I had the invited paper to the conference last year.” “I agree with that. It’s really good work. But it’s not that valuable to IBM.” And I actually set up a process in the division. We did an annual plan, and we started a thing called Project Descriptions. Every group of whatever it was, 5, 10, 15 people, we had on the organization chart, which had a piece of budget with the finances to manage that would create a PD. The PD would include a description of the value you deliver. Tell me what it is. And tell me what you will deliver in the next year, so we can talk to you about it in a year from now.”

And I went around telling people, “We’re going to stop doing some of stuff that we do. It’s very good work. If you’re an academic, it’d be fine. But we haven’t pruned in, I guess, ever, and we’re going to shift resources.” There were two reactions to that. One was, there were a lot of people who said, “That’s wonderful.” Because they were like me. They were concerned about, they’re doing some thing, “Why is this valuable to IBM?” And they were happy to jump into something quite different. Not only software people. Many hardware people did that too, some switching to software. For some others who did not want to change, my message was “If you want to continue what you’re doing, we’ll make an arrangement where, if you want to go to a university that wants to support it, you can retire,, and continue your current path.” A lot of the people were retirement-eligible by then.

The division was pretty mature by this point. The rapid growth was in the ’60s, and here we’re talking 25 years later. So, people have their quarter-century in and are retirement-eligible. Some people never forgave me for that. Some people were enormously successful in doing new things. And then, I would push specific groups to do specific things. I clearly saw clouds on the horizon, more than clouds, big trouble coming. I knew from my time in Armonk that the way that IBM managed through the mid and late ’80s was not sustainable. They kept the profit up by selling off the rental base, so it all dropped to the bottom line real quick. But you run out of the ability to do that after four or five years.

To get ready for harder times, I asked my financial team to create a specific accounting system for Research. We were treated as a cost center in IBM, and I hated that idea. I always pushed, “We’re an investment center. We produce value.” Not like the legal department’s a cost center. So, “I need an accounting system where I can manage how I operate on sort of a sources and uses of funds basis.” The corporation only looked at us as cost-net. So, anything I brought in from the outside, I could spend, and they never saw it. They very quickly set up the accounting system, very macro. Then, I brought in a guy named Dan McCurdy, who’s still around. He’s living in Connecticut now. Very good guy. He had been working with the IBM’s legal and licensing and government relations kind of business in Washington. But Dan was a very good, adventurous, young guy. I said, “What I want you to do is work with me to develop a stream of outside income.” And the target I gave him, which could have been considered outrageous, I guess, was, “In two years, I want that to be $50 million a year.” So, 10% of our budget. “Because I see trouble coming.”

One really key piece helped get us close to that goal. At that time, the semiconductor laser was alive and well, and running at room temperature. IBM was not doing much in that area, buying some for fiber channel. This group in Kingston was making what is called serial channel, which is an optical channel to connect boxes together with fiber instead of big gray cables. I tried to get two groups. One was a very strong gallium arsenide group in Yorktown under a guy named Jerry Woodal to focus on this area, particularly relative to high power reliability. Jerry is the guy who first developed aluminum gallium arsenide heterostructures, which was very important later in lasers, and visible LEDs, and things like that. And it never became important to IBM, but you could make incredibly efficient solar cells with it, although at very high cost.

So, I challenged him, as well as a tiny group in Zurich working on lasers to figure out why it was that when you pumped them hard, tried to get a lot of power out of them, the facets blew off. That was a huge problem because, by then, the fiber optics had become very real, and the erbium-doped fiber amplifier had been invented. The systems were using about 1.5 volts photon energy to carry the data. And to pump the erbium-doped amplifier, which is just a piece of fiber with this erbium dopant in it, which, you have to pump it with a higher energy laser to amplify the signal going through it. For long distance systems. Telecom, using these amplifiers was , it was so much simpler than converting the light to electrical, fixing the error bits, and turning it back into light. But nobody could make the pump lasers last any length of time. I challenged these teams to figure out what’s going on there, solve the problem.

Woodall had little interest in this, being a guy that always wanted to decide what he would work on. The group in Zurich got cracking on it, solved the problem very quickly, figured out that when you made the little laser bars, after you’d made them, you cleave them to make the facets, do the cleaving in ultra-high vacuum, and then deposit a little bit of silicon on the facets and let that oxidize, the facets didn’t blow off anymore. So, we started a laser business in the Zurich lab with about 20 people, and we were soon selling these things, the size of a grain of salt, for something like, $1,000 each. Fantastic business. And we soon had a revenue stream from it which rose to about $20 million a year. Being treated as a cost center by corporate accounting, and having our own more businesslike system, the corporate people never saw this revenue, and we were able to use it to buffer some of the later cuts in our budget.

The crisis started to come in ’91 or so, as the easy money from rentals faded away. John Akers was the CEO. Nice man, maybe too nice, but not tough enough. At one point, the accounting people came to me and said, “You’ve got to close this Zurich lab. It’s the most expensive lab in the world per head count.” It had, like, 200 people and had a budget of, I don’t know, $40 million or so, $200,000 a head. And I showed them my books. “Yeah, it costs $40 million, and I have $20 million of revenue from it. It’s actually the cheapest lab in the world.” [laughs] So, they went away.

I started licensing stuff, and taking pieces of work that we were doing, and spinning it out, selling it. So, we had a project, for example, on 3D printing. I think we even called it that then. And that was never going to go anyplace in IBM, so we sold it to Stratasys, which is still one of the big players in the field. And we got a good amount of money for it. We’d been doing some robotics work in the division and had a project with a couple of medical people out in California to make a robot to do the cavity cutting in the femur for total hip replacement, ROBODOC. We spun it out, even did an IPO on the NASDAQ! We did a bunch of things like that and earned a whole bunch of income from that. We also focused on teams chasing contract money for things where they would only minimally have to modify what they were doing, but they could get significant money to support the work.

All this business effort turned out to be enormously helpful when the company, in ’92 or ’93, first did layoffs. It was unheard of at IBM: we don’t lay people off. Fire people, but don’t lay them off. And they were going around the company trying to get people to lay people off and cutting budgets to force it. I was able to avoid doing layoffs. I urged a lot of individual people to leave, to retire, but we never had a general layoff of any sort in Research.

There were clearly some of the research staff who had become less effective, many then beyond retirement eligibility. I remember one particular guy, his name I forget, when I talked to him, he said, “I’d love to retire. The money would be fine. I don’t need more money. But I don’t want to be trapped at home”. I went over to the men’s room and peed, and by the time I came back, I had in my head an Emeritus program nearly fully out. You retire and become emeritus. You get a shared office. You can’t have your own laboratory, but you can work with colleagues (just as you likely do now) in their laboratories. You can be a source of wisdom. You get a business card, you’re on the computer system. You’re identified as a research staff member, emeritus. You come to work every day. You preserve your honor!

Lots of people took that. They were more than ready to retire. You don’t have to come to work every day, and you don’t have to come to work on time. So, that reduced head count a significant amount. Others, such as Woodall, who took all his molecular-beam epitaxy equipment and went to, I think, Purdue initially. Maybe someplace else first. He moved a couple times. But a number of people did that, left and did other things. The effectiveness for IBM, the value to IBM of what this group does, or what this individual does, or what this division does was the thing that drove my thinking through a good fraction of my career. And it’s captured in the arrow chart. So, ’92, ’93, company in deep yogurt…

Thomas:

Can I ask a quick question about the services and applications aspect? You gave the mathematics example, which was illuminating. What other things would come out of Research?

McGroddy:

We put a project up with a marketing team to build a hospital information system for the Kaiser Hospital region in Colorado. I pushed that. I’ve been a proponent for as long as I can remember of using IT more effectively in healthcare. So, I pushed that one. And it was initially successful, but the sales people preferred to sell piece parts, hardware, databases, into a broken system. They didn’t want to market an HIS, a much more complicated proposition.

One time, I went to talk to Cerner, proposed to buy half of them to get IBM into the business. Only a couple years ago they were bought by Oracle. They’re the number two in the country now. Number one is this private company, Epic, which is well ahead of Oracle, run since its founding by Judy Faulkner. But healthcare still doesn’t use IT at all well. If we used IT well in healthcare, the cancer people that dealt with me would’ve been able to find out that there was no benefit to me of taking Keytruda. They may have done it anyway because this is a huge financial benefit to the hospital. I likely wouldn’t have lost the ability to walk. The IT systems in healthcare delivery are mostly focused on reimbursement, on finance.

One of the things I did after I retired, which was end of ’96, early ’97 when I turned 60, was to invest in a startup electronic medical records system, which was clinically focused, not reimbursement-focused, and I ended up as the major shareholder and the chairman, worked there for 20 years. We were only in the end-stage renal disease space. I had the idea we could show how good it was, and then do a partnership and acquisition with Cerner or somebody like that. For lots of reasons, we never succeeded at that. But healthcare is 20 or 30 years behind. My healthcare records at the local place, Phelps Hospital right down the road here, are in at least 10 different EMR systems, which don’t talk to each other. So, I had a telemedicine, as they call it now, appointment with my urologist a week or so ago about the bladder cancer. It’s a six-month follow-up. And I had received a PET scan, which detects any significant cancer anyplace in your body. And I had blood work from my primary care doc. The urologist didn’t have access to either of those. I had to call these places. The PET scan, I got a copy of the CD ROM, and I had to call people up and ask, “Would you fax the record over on paper to him so he knows something about me?”

Thomas:

So, when you envision a research entity… you have a lot of people who have worked in devices, and in materials, and that sort of thing, and you ask them to shift in this direction to developing a system, are you basically asking them to retool their expertise, their career?

McGroddy:

In many cases, yes. And most of them loved it. And they were very good at it. They’re very smart people. They’re people who got into other things that were useful to customers, and worked together with the marketing people, and we developed a process which we called “first of a kind”, FOAK. And the idea was in that space, this SAS space, as we called it—Services, Applications, Solutions—you’d find somebody in the marketplace who had a customer who wanted to do something advanced that took some technology, often it was a lot of software, but it could be otherwise. And we would do a project with them, cost-shared, to meet their needs, but make sure we generalize the product so that it didn’t have any weirdness or odd constraints in it associated with their particular requirements. So, we had to find, “What’s the market it’s going to go to later?” But you develop closely connected to one customer.

So, you do a hospital information system for Kaiser Colorado, which is a good one to do it for because the patients there are fully capitated. Kaiser is responsible for the all aspects of their populations clinical healthcare. But you’re not going to just sell it into Kaiser, although that would be a pretty big market by itself. But you try to sell it, and they’re now called electronic medical records, EMR’s, not hospital information systems, which is a better name. But the ones that exist are absolutely dominated by reimbursement. And the expert systems write the notes, and they write the notes to meet the insurer’s requirement of what you must have done. “I spent 28 minutes with the patient, and we discussed this, that, and that.” Because if you don’t have that in, the insure will deny the claim. The clinical stuff is still very much background. And it would be so powerful… one of the things I learned in this company that I was deeply involved with for 20-odd years was that if you could study the information, the clinical information in the systematic way, you could learn all kinds of things. So, we had a piece of that product. This is well past the IBM years. There was a tool in there called the Outcomes Wizard. It was a very simple thing. And the whole thing was built on relational database technology. But you could ask it, for example, for me, “Go in the database and find all the patients that had bladder cancer, and then had surgery, and then had BMG immunotherapy, and it worked. And then, show me their health trajectory beyond that for those, a) who took Keytruda, and b) those who did not.” So, it’s looking at anonymized historical data. You would be learning from others’ recent experience.

But the thing that drove me to that, just the general idea that we should be doing something like that, is that there are cases, maybe the most famous of which involves the same company that makes Keytruda, Merck. 25 years ago. Merck had a product called Vioxx. Aimed it to people who had arthritis for pain reduction. It was very successful sales wise. After a while, somebody happened to notice, some medical practice, “Hey, it seems like a lot of our Vioxx patients are having heart attacks and strokes.” And somebody did a systematic study, and the results were so bad, they pulled the product off the market. And then, a study was done to ask, “How much damage was done with Vioxx?” And the result was, “We caused about 200,000 excess heart attacks and strokes, and 50,000 of those people died, over a couple of years.” And we never noticed it. And when you think about it, other people use IT which rapidly detects anomalies, Walmart for example. I learned a lot about that over the years from other companies.

One thing that began in IBM late in my career was the Partnership Executive program. For big and important customers, an IBM executive would be assigned to be the partnership executive. You were to work with the customer and help them work with IBM, you weren’t there to sell anything. And I had some big customers, Walmart, for example, which is a pretty big company. And I knew that in Walmart, they get information from checkout every 15 minutes. They get a dump. And if something’s going on in a store that’s different and interesting, they notice it, and they do something about it. I heard about one case where—I was told, probably—they had some PCs that they wanted to sell. Excess supply, I guess. And they told people, “Put them at the end of the aisle, put a sign up,” and all that. Managers have a lot of power in the stores there, even the sub-managers in the store. And they found some stores, they’re selling like hotcakes, and the others, they’re not moving at all. And the difference was, when they dug back on it, did it within hours of the sale starting, the ones that were selling like hotcakes, had an application running on it and a keyboard. And the others, they just had this box there. So they told the people, “Put the application on it and sell the PCs.”

Medicine should have that sort of thing. And what you would find if you did that is that a huge fraction of the expensive drugs in fact don’t work. And a huge fraction don’t work any better than some generic thing. There was a guy… there’s a kidney place down in New York at New York Hospital…

Thomas:

Sorry, Jim, can I interrupt? I have to start thinking about my train back to DC. And so, we haven’t really hit the crisis at IBM yet, so I’m wondering if we should.

McGroddy:

Okay. Yeah, and we can do some form of follow-up, if you want. Easier after we’ve met, can do it by FaceTime. You want to talk about the crisis when the company got in deep yogurt. You’re talking ’92, ’93. What happened was, the finances, IBM showed a huge loss at one point, one quarter there. It became clear that the company was in deep trouble, and in fact, the finance guy, a guy named Frank Metz, was not a financially trained guy. I think he was an engineer originally. Didn’t have good handle on things. And they were at risk of running out of cash, which is what triggers bankruptcy. You don’t go bankrupt because you’re losing too much money; you’re bankrupt if you can’t pay the salaries, or the loan payments, or the rent, or whatever it is.

One thing I did, I used to do every year, over Christmas, I used to think about my job and the company and write a piece. This for myself, something I never shared with anybody. But looking back, each year, I tried to capture what I had learned. Early in 1992, when it was clear there was big trouble, I got a call from Akers, the CEO, or was messaged somehow. He wanted to audit Research. And the way he wanted to do it was to have 10 senior IBM officers from outside the division come in and go over everything in some form for a couple of days. And we scheduled it for August 1992. So, I started to get ready for that fairly early in the year and used the arrow chart as the basis of it. And I made the same point that I told you a couple times. I tried to make the point that IBM Research is not part of the problem, it’s part of the solution, and it is a great investment…

Thomas:

IBM Research.

McGroddy:

IBM Research. It’s part of the solution, and it’s a great investment, and that you get a return, not all precisely quantifiable, that’s a multiple of what you put in every year, not 10 years out. So, I used the arrow chart as the theme of the whole thing and built a story around that with numerous examples. I had some of my key people do some of the presentations. CHQ picked a particular VP to lead the review, Pat Toole, who was a technology VP. I knew him very well because I’d been in his world, and he was one of the key people in solving this Clark board problem, and we had done many other things together. I knew a lot of the other reviewers, and they respected me because they knew much of the stuff I’d done in the company. I had fortunately had some influence on who was on this visiting committee, and when CHQ insisted that it include a country VP of marketing I was able to arrange that that would be the guy for Spain, who was a technical guy by background, and I’d gotten to know him fairly well over the years.

The net is, we did the review over two full days, two days, in Yorktown, with these 10 guys. (I don’t think there were any women involved!) And it went very well. We had presentations, and we showed them things, took them to different things where there was something interesting to see. And they went off to prepare their report. I always assumed that Akers’ goal was to take something in the range of 30% out of our budget. They came back with a report to Akers that said, “It is really working very well. You shouldn’t cut their budget at all.” And they had a few small things that we could do to improve, and they were good ideas. In preparation for that session, We created a book, which had text and pictures in it, I think it was maybe 30 pages, to give to them well ahead so they could read it before the intense two days. And it made our case reasonably well.

So, they went back to him, and he was probably very disappointed, and they didn’t cut the budget. Over Christmas of that year, the key business thing that I focused on was that book, because by then I was convinced that either of two things would happen. One was, Akers would fire me, because the top people were, at that time, trying to sell off divisions of the company. The Storage division, then called Adstar, and they acted like it was a separate company, was on the block, and it would include Research pieces with it. To me this didn’t make any sense. And I was very visibly opposed to it. And it was the very top people, some of whom… Paul Rizzo had come back from retirement to be part of that. In fact, Toole was one of them. I think he was probably somewhat sympathetic to what I wanted to do, which was keep the company together and fix it.

The position I took, and I’d written a couple white papers on it, was that the problem is, we’re acting like a supermarket, and many of the customers want to eat at a restaurant. They don’t want to buy the piece parts. This related to the meeting with Colin Crook. They want us to help them with their integrated business, not buy some eggs, and some bread, and some beef, and go home and cook the meal. They want a meal on the table, and we should offer them that. And we had, historically, when the computer was the solution early on… you had an insurance company, and you needed to do the billing. The software’s pretty simple. But over time, obviously, it had gotten much more deeply involved in the business. But I said, “We should keep it together. There’s value in that.” And I wrote that up. And actually, later on, when Akers had been fired, and the search committee started, I got that paper into the process, and it got to the committee, and Gerstner had read it before I first met him. So, over Christmas, I took this book from the summer review, and I basically made a much more refined version of that, shorter—I think it was 16 pages when I got done.

The idea was, either Akers is going to fire me, and I should leave something behind like this for my successor to try and preserve the legacy of what so many have done for so many years, or Akers is going to get fired, and I’m going to have a new boss, and I’m going to have to explain this place to him. This small book would be good for that, too. So, I wrote the piece, put it aside. But I think it was by January, they had announced Akers was leaving, and they had a search committee led by Jim Burke, who had run Johnson & Johnson, whom I had met before. I made sure he got a copy of this small book, and another paper I had written about what I thought was wrong with IBM… “What diseases does IBM have?”, a diagnostic paper.

The search went very quickly, and Gerstner showed up for a start date on April 1, ’93. He actually showed up a couple of days before and had a meeting in the headquarters building, which is no longer the headquarters, he built a new one, in the boardroom of the surviving senior executives. They had let go about half the senior executives at that point to save money, including John Armstrong. And Lou told us in that meeting… he’s not a very big man, came in looking very serious, and he came in and said, “I don’t know anything about this business, and I’m going to be the CEO here. You have to teach me about it. I want each of you to write me a white paper, a dozen or so pages, where you describe the nature of your part of business. What are your products and services? Who are your competitors? What’s your economic model? What are your strengths and weaknesses?” etc. “And send that to me in about two weeks, and then I will come and spend some time with each of you to discuss it. That’s how I’ll learn about the company.” Very sensible.

IBM, in those days, in almost all parts of the company—you’ve probably heard this before—did not write things in sentences. These used flip charts, a standard-sized paper chart. It had phrases on it, but no sentences. In Research, we had always written our strategy and our plan in narrative form, sentences, pictures. And I had this thing that I had refined over Christmas, which came damn close to being what he wanted, so I went back and worked with my staff. I turned it in in two days, well ahead of the two-week deadline. The result of that was that when he joined on April 1, the first place he went outside headquarters was to come see Research, because I’d turned my book in.

Everybody was trying to figure out how to write out, in narrative form, something that fit the outline that he’d given, which was a very sensible outline. I used essentially the same theme. So, on April 15, two weeks into the job, Lou came to Research at Yorktown. I didn’t have a desk in my office, I had a table, cleaned it off—usually had a lot of junk on it. All it had on it was a ThinkPad, which was less than a year old as a product then. After we did the pleasantries, shook hands, and he sat down, I said, “With this ThinkPad, I think I can show you some of what we do here that adds value to this company.”

So I talked to him a bit about some of the stuff that we had done in Research, like the track point thing. Which was done actually in the math department. And I had an agenda form. I said, “I’ll take you around and show you four or five different things in the lab and have you meet the people, then we’ll meet collectively at the end of the day.” And what he did very early on, was to ask me, “How do you figure out what to do?” And I said, “Well, we have a very good process for that here at Research. We have this technology outlook, that’s the backdrop, and then we do a strategy that deals with some of the issues that are revealed in the backdrop and meet the long term goal that we have. And then, every year, we make a plan, and we update the strategy.” “When can I see that?” “Right now, if you want.” I had a paper copy of it there in my drawer, not for any particular purpose, and I pulled it out, gave it to him.

He took it, and we spent the next hour-and-a-half going through it page by page. And he’s scribbling on it. He’s left-handed. I wish I had that copy of that copy! Then, I took him around and showed him some things that made more real some of those things, like the 3D printing that we were doing was still there. I told him, “this doesn’t have a home in IBM, and we are selling it to a company which does.” We looked at some very fundamental stuff in ion-implantation-related physics, a big accelerator, and some other big machines. I took him to meet the people and see some of the stuff the math people were doing. I don’t remember which of the projects. But I explained to him how we were organized and how we worked. And then, we had a meeting at the end of the day in the conference room next to my office, about eight of us. I had a group called the Research Management Committee, which was the people that reported directly to me, including the Zurich lab director and the Almaden lab director.

I’ll never forget what he told me at the end of that meeting, when we went back in my office to say goodbye. He said, “You know, until two weeks ago, I was on the board of AT&T”—this is almost verbatim—“and one of the biggest frustrations there among the board and the senior management is that they know that in their research area, they have some of the smartest people in the world, and yet they make very little difference to the products, the business, or the bottom line. And I want to thank you for not putting that problem on my plate here.” And I didn’t realize until later, in some sense, what he was telling me… at my retirement dinner, he explained it more fully... I think it was privately that when he got asked to lead IBM, he talked to some IBM executives. He concluded, based on his Bell Labs experience—and in fact, he said that at least one person he talked to in IBM reinforced it—that he should break IBM Research up into its pieces and send the pieces to the group they were supposed to partner with. And what he was telling me was, he wasn’t going to do that.

Lou instantly became the biggest fan Research could have. In fact, he didn’t have a clear idea where the boundary of Research was and product began. I was reporting directly to him in the new organization. I don’t know whether I knew that even then, probably it hadn’t even been set. But he asked me to go all over the country and all over the world to be part of meetings that he was going to have with CEOs of customers to explain how IBM… as I said, how it was going to become a restaurant again. “We’re going to help you with your business, particularly this internet thing which is coming along.” They said, “What?” And we did that, all over the world. He’d have a bunch of CEO level people come, typically multiple industries to avoid competitors being in the same room.

But we’d go to England, or Germany, or someplace, and he’d have a two-day meeting with a grand dinner in between with 8 or 10 CEOs to explain what IBM’s going to be like going forward. And it fit this model that I had been pushing, which he really bought. So, we became the best of buddies and were for the remainder of the time I was there. That was, as I said, April 15, ‘93, and I left Research around the end of ’95. A year or so later I left the company, I handed Research over to Paul Horn and then moved my office to CHQ in Armonk so I’d be out of Paul’s way, but I’d be available to give advice.

During the time from April ’93 to when I left Research I worked very closely with Lou and all the top management to help them get settled in and deal with a lot of issues that came up along the way. And the one thing about that that amazes me—I think I put it in one of my notes to you—Lou eventually wrote this book, Who Says Elephants Can’t Dance? I spent several hours with the writer, giving input. Eventually, after I had been gone a couple of years, I got a copy of the book with a scribble from Lou in it that says something, I forget what. And first thing you do, you turn to the index. “What’d he say about me?” And it was very simple: my name’s not in the index. He’s a guy that is particularly focused on Lou Gerstner. So, he rewrote a good bit of the history. And that’s okay with me, I never needed much credit.

I stayed at IBM after stepping down as Director of Research for a year to try and get some of the things that I’d been trying for a long time to make happen, for there to be a process to spin things off, for example, to get them into the medical field in a different way. Couple other things, keep the flat-panel thing on track.

After I left, I had lots of opportunities. Hewlett-Packard, for example, asked me to come out and be the CTO and be on the short list to be CEO, before they had Carly [Fiorina] on the list. And I said no to them. Before I actually retired… since you knew you were going to retire, that’s a very good thing—60 is going to come. And I had thought through, “What am I going to do then?” I had gotten involved in a bunch of things locally, and I had some board involvements, particularly the local hospital board, small stuff. I picked that up in these areas later. But I made three rules for myself: I won’t have a boss, I won’t have employees, and I won’t move. My youngest daughter Aileen at that point was 8 years old. We had a Chinese nanny at the time—we’d had two of them, terrific, from when she was born—who kept the house going. My wife was an entrepreneur at that point and was working crazy hours. And I thought, “This kid, that’s what matters, Aileen, so, I have to be the available parent. That’s my job”. Teach her stuff, and take her places, learn ice skating, and horse riding, and all the stuff that kids do that a Chinese nanny couldn’t do. They couldn’t drive a car, they didn’t speak English. Chinese was half the language in the house.

So, I never took another job, or moved, or had any employees. The only time I had a job was at this electronic medical records company. Occasionally, we’d get in trouble, I’d have to fire the CEO, and I’d have to be acting CEO for a couple months to get a new CEO in place. So, I did a whole bunch of things, and before retirement, I’d joined the board of the local hospital, figured I maybe can use to for leverage to both improve the hospital and learn more about how to get technology in. It’s a pity, it’s not a very good hospital now. It’s been acquired and consolidated. Sad. I avoided it being acquires years earlier. The CEO was going to give it to Mount Sinai Hospital in the city, and he was going to get a big job, and all the “interesting” work would go down there. And a bunch of the docs could see that, and they asked me to intervene, and I did. And a couple of the veterans from IBM set up a network of four peer hospitals here in Westchester, high-quality hospitals without a big New York City hospital in it. But those days are long gone. The whole industry has changed radically and continues to consolidate at a rapid pace, largely to the detriment of clinical quality and availability.

I also did a bunch of stuff with the National Academies, and off the back of that, with the FBI and the CIA, and a couple of corporate boards I was on for years. I built a vacation house on a lake in New Hampshire to gather the family, which is a wonderful thing to have. They’ve organized it now so that what was a bedroom on the first floor that only had a half bath now has a full bath. My third daughter’s an architect, and about a year and a half ago, figured out how to put a shower in there. And that’s now the room I can use so I don’t have to go up and down the stairs. It’s a beautiful place on nine acres right on the lake, with about 800 feet of lakefront. So, I continue to spend time up there. It’s not easy for me to get there now. And when I’m there, I’m very limited in what I can do.

What other questions do you have?

Thomas:

I don’t know. They would be essentially random questions.

McGroddy:

That’s fine. [Discussion of Thomas’s travel plans.]

Thomas:

Well, I think we may as well just call it here. And I think we can settle anything else via email or a follow-up. This has been five hours. [laughs] Which is very generous for your time.

McGroddy:

Well, I have the time available, and I value what you’re doing tremendously because there’s no place else… IBM’s never interviewed me for any history thing.

Thomas:

Well, there are a lot of people who don’t do interviews. I’m hard-pressed to think of organizations that do it.

McGroddy:

IBM does not do a very good job on exploiting, benefitting from, its retirees. Many would love to be more involved, not for the money, but because they are proud of, and value, the institution. This is a thing that I think we did well in Research with the research staff member emeritus program, we got a huge value out of that.

Thomas:

I’m going to pause this.

[End]