Talent, Collaboration Key to Innovation

Rick Rashid and John Markoff

“My approach to research has been exactly the same for 21 years. Hire great people and give them an environment in which they can be productive,” said Rick Rashid, Microsoft’s chief research officer.

Microsoft chief research officer says Moore’s Law is a statement of human optimism.

The scientist known as Microsoft Research employee No. 1 took to the stage in Silicon Valley and told entrepreneurs and technology enthusiasts that investing in talented people is the key to developing problem-solving innovations.

Reflecting on his 2 decades at Microsoft, Rick Rashid, the company’s chief research officer, told an audience recently at the Computer History Museum in Mountain View how he built a research division that now has 1,100 computer scientists, physicists and mathematicians, including 850 Ph.D.s. Those researchers explore basic and applied sciences that, in turn, help Microsoft products and services be globally competitive.

Rashid’s conversation with John Markoff of the New York Times was part of the museum’s “Revolutionaries” series, which is sponsored by Intel.

Rashid talked about the innovations that went into the Microsoft Tablet PCs introduced a decade ago. He said that although the timing wasn’t right for that first tablet, many lessons learned from it are now in the company’s Surface products. By contrast, Microsoft Kinect, a motion-sensing input device, was a hit from the get-go and continues to grow.

“Kinect came out of our computer machine vision work,” he said. “It changed how people could use ‘gesture vision’ to interact with computers.”

Rashid also addressed the challenge of finding and retaining talent, especially women scientists.

“Eighteen percent of our researchers are women, which is in line with national levels,” he said. “But our national numbers are dropping because we’re doing a poor job in schools and in companies describing what computer science is really like … and how it is a field where you can help solve real problems.”

After leaving the stage, Rashid sat down for an interview with Intel Free Press and discussed advances in computer learning and speech recognition technologies, the impact of sensors and Microsoft’s ability to change and innovate.

Does your research team collaborate with Intel Labs?

Microsoft Research has worked with Intel quite a bit over the years. I’ve worked with the research organization at Intel and [Intel Chief Technology Officer] Justin Rattner I’ve known for many years, [dating] back to when I was at Carnegie Mellon. Justin has had a real concern about how to get young people into the field of computer science, how to get women into the field of computer science. He and I have been board members at the Anita Borg Institute for many years together and that’s been a productive relationship.

Rick Rashid Microsoft Research

"When I first got to Microsoft, it was this small company that produced a few products primarily sold to the retail channel in boxes," said Rick Rashid, Microsoft's chief research officer. "Today we're a company of almost 100,000 people with more than $70 billion worth of sales to consumers, businesses, governments. We're really a very different company than we were back then."

We’ve also had a very productive relationship both in terms of pushing technologies like white spaces, new networking technologies, experimenting with new ways of thinking about doing multiprocessors in computers. The relationship that Microsoft and Intel have is a really positive one, and I have felt really good about that over the years.

What are the challenges of collaborating with another company’s research team?

You’re never going to have a perfect match of interests. That’s just part of the technology innovation process. People come up with new ideas or new ways of doing things, but it doesn’t always work.

Someone in Microsoft Research may think that they’ve got a great new idea that should be in some product. They go to the product team and get a look like they’ve just landed from Mars. [The product guys say,] “We don’t have that problem, our customers don’t want that, they don’t care about that … we really have this problem, can you solve it for us?”

The same thing happens between companies. Intel may perceive a particular need in the marketplace that Microsoft doesn’t because it may not be important to our customers. Alternatively, we may have a set of issues or concerns around something that a company like Intel doesn’t see or perceive. That’s one of the reasons why it’s great for us as companies to have a dialogue, where we’re constantly exchanging information and ideas … maybe not exactly what we’re doing now but about the future. What has been productive over the last few years between Microsoft and Intel is that we have this kind of dialogue, and it has been helpful on both sides.

What research findings have led to product features?

We were working on natural language processing, and that was very quickly valuable to us in doing things like building spell checkers and grammar checkers into our Office products. The work we did in program and code analysis was extremely important in helping us make the transition from 16-bit to 32-bit computing. We were able to develop a technique for dramatically reducing the working set size of 32-bit programs. This was during a period when the size of computer memories was actually not increasing the way it would normally do. Memory prices during that period from 1992-1995 actually ticked up for a while and weren’t following that normal Moore’s Law exponential curve. When we made the transition from 16-bit to 32-bit code, we needed to be able to reduce the working set size for programs to compensate for the fact that they took up twice as much space as the 16-bit programs did. That was a huge contribution to the company at a critical time in the industry.

What has been the impact of Kinect in bringing about a new era of perceptual computing?

Microsoft Research was heavily involved in the creation of Microsoft Kinect. On the one hand it is a way to interact with the game console, but more to the point it has changed the way people think about how computers can interact and perceive them. It not only does a great job of tracking your motion but it also does a great job of hearing you because of the array microphone built into it and the speech recognition technology. It has changed how people think — even in the research world — about what you can do to interact with the computer.

Do you see an end to Moore’s Law?

People always ask about Moore’s Law and what’s the future of Moore’s Law and will we run into a barrier. The way I look at Moore’s Law is it’s more of a statement about people — About us in the technology industry, and our belief that we can keep making things better, that we can keep pushing what’s possible and to some extent if we run into a problem in one area we’re going to find a solution and push it somewhere else or solve the problem a different way. This is more a statement of human optimism and our willingness to overcome things to achieve a better future than it is about technology.

Talent, Collaboration Key to Innovation
Rick Rashid and John Markoff

“My approach to research has been exactly the same for 21 years. Hire great people and give them an environment in which they can be productive,” said Rick Rashid, Microsoft’s chief research officer.

Microsoft chief research officer says Moore’s Law is a statement of human optimism.

The scientist known as Microsoft Research employee No. 1 took to the stage in Silicon Valley and told entrepreneurs and technology enthusiasts that investing in talented people is the key to developing problem-solving innovations.

Reflecting on his 2 decades at Microsoft, Rick Rashid, the company’s chief research officer, told an audience recently at the Computer History Museum in Mountain View how he built a research division that now has 1,100 computer scientists, physicists and mathematicians, including 850 Ph.D.s. Those researchers explore basic and applied sciences that, in turn, help Microsoft products and services be globally competitive.

Rashid’s conversation with John Markoff of the New York Times was part of the museum’s “Revolutionaries” series, which is sponsored by Intel.

Rashid talked about the innovations that went into the Microsoft Tablet PCs introduced a decade ago. He said that although the timing wasn’t right for that first tablet, many lessons learned from it are now in the company’s Surface products. By contrast, Microsoft Kinect, a motion-sensing input device, was a hit from the get-go and continues to grow.

“Kinect came out of our computer machine vision work,” he said. “It changed how people could use ‘gesture vision’ to interact with computers.”

Rashid also addressed the challenge of finding and retaining talent, especially women scientists.

“Eighteen percent of our researchers are women, which is in line with national levels,” he said. “But our national numbers are dropping because we’re doing a poor job in schools and in companies describing what computer science is really like … and how it is a field where you can help solve real problems.”

After leaving the stage, Rashid sat down for an interview with Intel Free Press and discussed advances in computer learning and speech recognition technologies, the impact of sensors and Microsoft’s ability to change and innovate.

Does your research team collaborate with Intel Labs?

Microsoft Research has worked with Intel quite a bit over the years. I’ve worked with the research organization at Intel and [Intel Chief Technology Officer] Justin Rattner I’ve known for many years, [dating] back to when I was at Carnegie Mellon. Justin has had a real concern about how to get young people into the field of computer science, how to get women into the field of computer science. He and I have been board members at the Anita Borg Institute for many years together and that’s been a productive relationship.

Rick Rashid Microsoft Research

"When I first got to Microsoft, it was this small company that produced a few products primarily sold to the retail channel in boxes," said Rick Rashid, Microsoft's chief research officer. "Today we're a company of almost 100,000 people with more than $70 billion worth of sales to consumers, businesses, governments. We're really a very different company than we were back then."

We’ve also had a very productive relationship both in terms of pushing technologies like white spaces, new networking technologies, experimenting with new ways of thinking about doing multiprocessors in computers. The relationship that Microsoft and Intel have is a really positive one, and I have felt really good about that over the years.

What are the challenges of collaborating with another company’s research team?

You’re never going to have a perfect match of interests. That’s just part of the technology innovation process. People come up with new ideas or new ways of doing things, but it doesn’t always work.

Someone in Microsoft Research may think that they’ve got a great new idea that should be in some product. They go to the product team and get a look like they’ve just landed from Mars. [The product guys say,] “We don’t have that problem, our customers don’t want that, they don’t care about that … we really have this problem, can you solve it for us?”

The same thing happens between companies. Intel may perceive a particular need in the marketplace that Microsoft doesn’t because it may not be important to our customers. Alternatively, we may have a set of issues or concerns around something that a company like Intel doesn’t see or perceive. That’s one of the reasons why it’s great for us as companies to have a dialogue, where we’re constantly exchanging information and ideas … maybe not exactly what we’re doing now but about the future. What has been productive over the last few years between Microsoft and Intel is that we have this kind of dialogue, and it has been helpful on both sides.

What research findings have led to product features?

We were working on natural language processing, and that was very quickly valuable to us in doing things like building spell checkers and grammar checkers into our Office products. The work we did in program and code analysis was extremely important in helping us make the transition from 16-bit to 32-bit computing. We were able to develop a technique for dramatically reducing the working set size of 32-bit programs. This was during a period when the size of computer memories was actually not increasing the way it would normally do. Memory prices during that period from 1992-1995 actually ticked up for a while and weren’t following that normal Moore’s Law exponential curve. When we made the transition from 16-bit to 32-bit code, we needed to be able to reduce the working set size for programs to compensate for the fact that they took up twice as much space as the 16-bit programs did. That was a huge contribution to the company at a critical time in the industry.

What has been the impact of Kinect in bringing about a new era of perceptual computing?

Microsoft Research was heavily involved in the creation of Microsoft Kinect. On the one hand it is a way to interact with the game console, but more to the point it has changed the way people think about how computers can interact and perceive them. It not only does a great job of tracking your motion but it also does a great job of hearing you because of the array microphone built into it and the speech recognition technology. It has changed how people think — even in the research world — about what you can do to interact with the computer.

Do you see an end to Moore’s Law?

People always ask about Moore’s Law and what’s the future of Moore’s Law and will we run into a barrier. The way I look at Moore’s Law is it’s more of a statement about people — About us in the technology industry, and our belief that we can keep making things better, that we can keep pushing what’s possible and to some extent if we run into a problem in one area we’re going to find a solution and push it somewhere else or solve the problem a different way. This is more a statement of human optimism and our willingness to overcome things to achieve a better future than it is about technology.