What would you say have been some of the highlights of your career at IBM?
When I look back on my career. First, I started in the research labs in 1970. And when I started in research I had switched fields from physics to computer science. And it took me a while to find my bearings. But eventually I did. And when I discovered that what I really, really enjoyed was to work on technology, but to also work on how to bring technologies to the market.
So I found out that what I was really best at and was happiest doing, was having one foot in the technology camp and one foot in the marketplace.
And at the time in research, in the 70s, we were organizing more and more technology transfer activities to work closely with IBMs product division. And I was very involved in organizing a number of those activities, especially all the ones involving systems. And because of that I became very close to many of the people in IBM [who] were involved in large systems architecture. To the point where then in the mid-80s I was offered a job to go do strategy in the large systems unit. Which I did.
That was a time, as the 80s went on, of incredible technology transition. Because that was the time when the focus of the systems business was switching from mainframes and the data center to client-server and more distributed systems. And more and more of our mainframe business was at a huge risk. And one of the major projects that I was involved in was how do we transform the mainframe so we can build them out of the same microprocessors technologies used in small systems and workstations. And how we then introduced parallel architectures to mainframes so you can scale them by having multiple systems.
And that became essentially the answer, that became Parallel Sysplex and the way we upgraded the mainframe. So that became a particularly satisfying project.
And around that time in the early 90s I was asked to go and start our parallel supercomputing business. Which was based on similar principles: to use technologies coming out of our Unix workstation business; it was RISC architectures and Unix (AIX) and build a family of parallel systems—what became the SP2. That was a very nice area because when I worked as a graduate student in physics in the 1960s I had been involved with supercomputing, so this got me back into the supercomputing world.
And IBM at the time was nowhere in supercomputing. But as a result of these efforts in supercomputing, we became the leaders in supercomputing.
Later on, in the mid-90s, I was asked by Lou Gerstner to start our Internet business. So we organized the Internet division and developed what became our e-business strategy, and the e-business strategy has flourished. I did that for about four years. It was clear by late 1999 that we didnt need to separate the Internet business. This stuff was so integrated into every aspect of our business.
Then at the time, Linux was coming along and then Sam Palmisano asked me to come over to the enterprise systems unit he was leading to start our Linux initiative. Which I did. And that has also become very important to IBM in a number of ways, with Linux itself and with the culture of using open source software in more and more things that we did, and working with the open source community.
I then went on to lead our On Demand initiative in the first year when we started it in 2002. I got involved with our grid computing initiative, our autonomic computing initiative.
And then for the last couple of years Ive been involved in doing technical strategy across the company and innovation initiatives across the company. And Ive had responsibility for university relations and for the IBM Academy of Technology and other major parts of our technical strategy.
Are there any of those things you just mentioned that youd consider your most memorable achievement?
When I look back on the things I did its almost like you have various children and you say, is there any one child I love the most? And then you start thinking, Well, I love them for different reasons.
For example, the project to help transform the mainframe with microprocessor technologies and parallel architectures was particularly sweet because our backs were against the wall. So this was more of a case of we either do this or were gone. Because if we hadnt been able to reinvent IBMs mainframe business, the company would have been in dire straits. Or IBM would have become a very different company.
The parallel supercomputing, I loved the area, but I also loved working with the people with whom I was working in universities and research labs. I love the scientific community.
Sometimes people say a large company cannot do anything new; thats not true. We were nowhere in supercomputing and now we are the leader. Thats not bad. I think what happens is that if you are a $19 billion company and you hit .350 and people say: what have you done lately? But I hit .350. You cant hit a home run every time.
The Internet has been just wonderful. Ever since I started with the Internet in 1995, its almost colored everything that I do.
Linux has its own controversial challenges and its own competitiveness.
So each project has brought some fascinating angles that made them special. And theyve all been additive. Theyve all added to what Ive been able to do.
A lot of people view you as a visionary. What is it that gives you that? Where do you get the vision from?
The answer is easy: Find where the smart people are and hang out with them. Im serious. The smart people have a lot of ideas and in IBM I have been really lucky. First of all IBM has some very smart people in our research and development labs. And then Ive been really working a lot with people in universities and outside research labs. And Ive been in committees in Washington.
And what Ive been good at is talking to lots of smart people and being able to discern a pattern for when something comes up again and again in different conversations. Maybe theyre approaching it from different angles. But the way I looked at it a new idea was whether it was something we should do, and then how we should do it in IBM. Because just because its something we should do, doesnt mean we have to do it like everybody else is doing. In fact, it could even be that its something somebody should do, but not us.
So its more almost like finding out what the smart people are thinking, extract it, look for the patterns and then eventually something clicks. Then you look at how we bring it to the marketplace.
You sound like youre the early embodiment of that T-shaped skills person IBM has been talking about—the folks with both business insight and technology skills.
I think Ive been pretty good at that. And in technology in particular, Ive always loved complex systems—parallel supercomputing, mainframes, the Internet. For some reason Ive never been attracted to just a gadget. Gadgets are fine, but I love to think about things that have many moving parts.
You started in research, you said?
Right, I spent 15 years in research.
So the practice IBM has of technology transfer from research, were you instrumental in accelerating that?
Yes, I think I was very involved in helping to develop our technology transfer activities. When I first joined IBM, research was not very close to the product divisions, let alone the customers. Right now we dont even call it technology transfer anymore. We just call it problem solving. And we had a superb director of research, Ralph Gomory, and he had this thought that we should be out not just helping the world of science, but also helping IBM. And he encouraged us to get much more involved with other parts of IBM.
And we learned that the culture of helping somebody solve their problems, you first had to know what they do, then you have to know what their problems are. Then you have to become an expert in the problem yourself so you can come up with a way to do it better.
So the project started with this humility that you cannot invent anything, you cannot innovate, if youre not involved in the area you want to innovate in. So we became very close with the people in the product divisions. And I worked very closely with the people in our large systems community in Poughkeepsie and other places.
Right. I also was surprised to see that IBM Research has a great deal of interaction with customers now.
Yes, because we realized—and this had started in those days but it continues—that if youre going to solve problems, go where the problems are. And the majority of the problems are in the marketplace. There are problems back in the lab, and we need people to do that, but we think thats a small part of the problem set. The real problems are out there and customers have them. So if customers have them then you have to learn what they do. And this is part of learning what they do.