Research Keeps Microsoft on the ‘Bleeding Edge’
Q&A: Craig Mundie discusses why Microsoft continues to invest in research and innovation, even during the current economic downturn.
When confronted with a tricky problem, Microsoft Research loves to tinker. The supersized think tank has 800 researchers who every day explore an estimated 55 areas of research at six labs spread across the globe. Their ingenuity and creativity is on display Feb. 24-26 at TechFest, Microsoft’s annual celebration of its advanced research work. Microsoft Chief Research and Strategy Officer Craig Mundie spoke with PressPass about innovation at Microsoft.
PressPass: Microsoft Research TechFest is happening this week. Why was it created?
Craig Mundie: In part, TechFest was created to offer yet another way for us to effect technology transfer from research activities to our product groups. We think it’s important for our engineering community to have as broad an awareness as possible of the kinds of things that Microsoft Research is doing. TechFest offers a more grassroots way for people to explore Microsoft Research’s work, then figure out a way to incorporate it into products. It’s really exciting and encouraging when you see the breadth of work that is being done and the degree of invention that underlies it. For people who live and work in a technology-driven society and company, it’s a great place to see the bleeding edge of what’s ahead.
TechFest creates an environment where serendipity can happen. People who visit might be looking for a specific idea or technology, but because they get immersed in many interesting demonstrations, there’s a pretty high probability they’ll discover something unexpected. So you get this serendipitous marriage of need and capability.
PressPass: Do you have a favorite demo or project this year that brings together multiple technologies?
Mundie: My current favorite is the virtual receptionist. We’ve got a 3D model of a receptionist that we’re going to be testing in Microsoft reception lobbies. The receptionist knows how to interact with you to arrange campus shuttle service, along with some other tasks. You talk to the model as if it were a person. It interacts with you, maintains eye gaze, and can deal with multiple people. It’s built using our Microsoft Robotics Studio technology as a new programming platform. It brings together machine learning, speech recognition, vision – in fact, a wide range of the technologies we’ve been working on in recent years. For the first time, you can see a future where our ability to interact with machines is becoming much more like our ability to interact with other people.
PressPass: In this economic climate, why continue investing in research, unproven technology and things on the cutting edge?
Mundie: We’ve always believed that being able to sustain our long-term investments alongside our current product investments is critically important to our success. First, we want to keep making improvements to all of our existing products. Clearly we need to do that because competition continues to get stronger, even in these difficult times.
Second, we can’t anticipate everything that people do. Research supplies us with a shock-absorbing capability to react to the unexpected. And sometimes economic downturns are a good time to disrupt the market ourselves. For example, we’re building a whole new business in healthcare, which has drawn a great deal from our research groups.
If you look back, the companies that excelled during previous economic downturns, including the Great Depression, were those that quickly made adjustments to their cost base, found ways to continue investing in the future, and invented new products while they were actually in the downturn. Companies like RCA, the Radio Corporation of America, went into the depression when radio was the predominant mass media. During that period of time RCA was among those companies that pioneered television, which was launched around the end of the Great Depression. Companies like RCA and CBS became household names. They were riding the new invention in the upswing after the end of the economic malaise. So if we want to excel, rather than just survive, we need to keep developing advanced new products, and bring them to market as we exit the economic downturn.
PressPass: There is a lot of discussion about technology transfer, or integrating research into Microsoft’s products. How well is the company doing there?
Mundie: We have four ways to transfer the work of our researchers into our products. The most basic is straightforward technology transfer, where we take some raw technology created by Microsoft Research, give it to the product groups, and they handle the implementation. There are hundreds of these a year – most of the time they represent an incremental improvement in a product we already have. One example from a few years back is the integration of Microsoft Research’s work in natural language processing into the Office “help” facility. The second thing we do is help design new features for products. For example, our researchers shared artificial intelligence technology to help improve the gaming experience in Xbox Live. This feature learns a player’s expertise and adjusts the game to the right level of difficulty. It also suggests pairings for online players with similar connection speeds and gaming experience.
The third way our research reaches the market is through development of new products. Often, these are the result of seeing a market opportunity or a technological way to move forward. We know where it might ultimately end up in our distribution mechanism, but it’s not a product like any we already have. One example is our robotics tool kit. When we started this several years ago the company didn’t have a robotics business, but we thought that it was going to be a big and important market driven by software in the future. So two years ago we brought to market a robotics tool kit, a visual programming language, and a 3D simulation environment. We’re now merging those things with some of the activities that are going on in the company. Eventually we’ll end up with a new product and some consolidated technology that flows through it.
The fourth area is when we actually create whole new businesses at the company. The largest new business that we’re creating is around healthcare. We’re very excited about its potential, both because of the absolute size of the market and the fact that the healthcare field is going to transition into more of a digital- and customer-driven business. We think our software can play a big role in that.
PressPass: In your unique role you’ve probably seen a lot of creative ideas come across your desk. What are some of the “out there” concepts you’ve seen? Have any of the ideas panned out?
Mundie: One of the projects I’m actually pretty excited about is our work in quantum computation. Four or five years ago, most people would have said that a practical quantum computer might be 50 years or more in the future. But one of our researchers, Michael Freedman, came to me with an idea – based on his field of mathematics, which is topology – for how to solve the problem of de-coherence in quantum computation at the physics level. We now have a group and a lab that’s one of the top quantum computation labs in the world. And I think that we may have considerably shortened the time horizon for developing a practical quantum computer.
PressPass: Are there any Microsoft Research projects research that you are particularly excited about?
Mundie: There will be hundreds of things to check out at TechFest. I think all of them are interesting in their own right. If you’re a tools developer, there are new programming languages. If you’re an operating system guy or a networking person, there’s a lot of new technology there too. Wireless, social networking, we have a very, very broad of array of things going on. Many of them are certainly focused on our historical lines of business, but the majority look far beyond that.
PressPass: Does the fact that Microsoft now has 10,000 patents show that the company is innovative?
Mundie: When you realize we’re only working in the software field, not in any of the hardware areas, we certainly have one the world’s highest-quality patent portfolios and now one of the largest. That’s just one measure of innovation. Clearly the challenge for us now and in the future will not be whether we can have the idea and get the patent, it’s whether or not we can convert that into business value in the form of a product that customers really appreciate and will pay for. That’s why all of these mechanisms, whether it’s TechFest or other forms of technology transfer, are essential to the long term health of the company, and why we must continue to focus on R&D even in economically difficult times.