It’s been fashionable for years to knock Microsoft for not being hip, cutting edge, innovative. Choose your dart. The folks at Redmond have heard it all. But, of course, the picture is more complex. In its most recent fiscal year, Microsoft spent $8.2 billion on research. That sum makes it one of the largest research organizations in the world.
So when I was offered the chance last month to chat with Craig Mundie, Microsoft’s Chief Research and Strategy Officer, I jumped on it. Mundie spent a day at the University of California at Berkeley as part of a week-long visit to universities across the country.
According to his official bio: “Mundie is responsible for directing the company’s technical strategy and long-term investments. In this role, he oversees Microsoft Research and other technology and research initiatives, the company’s health and education businesses, and a number of technology incubations. Mundie also works with government and business leaders around the world on technology policy, regulation and standards.”
On a basic level, Mundie is the guy in charge of figuring out the future for Microsoft. So, what he thinks matters. He’s been at the software company since 1992. And Mundie makes a convincing case that the conventional wisdom about Microsoft and innovation is wrong.
During a presentation to students and faculty at Berkeley, Mundie began by demonstrating the World Wide Telescope. And if you haven’t downloaded this yet, you should. Especially if you have a five-year-old at home, like me. The WWT pulls together almost every astronomical image in the world to create a kind of “Google Maps” for the universe.
There were a few highlights from Mundie’s official presentation to students and faculty:
- Mundie is excited about “speculative computing.” This is where the computer doesn’t just respond to your actions, but learns to anticipate what you want next and does it for you.
- Mundie doesn’t think the future will be just “cloud computing.” That’s the trendy phrase coined by Google to illustrate the notion that all computing will move to the Internet. Mundie thinks the future will be a mix of people doing things on their computer and over the Internet. Why? “Even if you think the price of broadband would go to zero, and there was infinite speed, you can still do some things faster on the desktop.” The key, according to Mundie, is to design software that optimizes both experiences. “The cloud plus the client,” Mundie said. “We need more integrated programming.”
- He believes search is still in its infancy. “The idea that you put three words in a box and get 100,000 hits that you have to sort out on your own isn’t the future.” Instead, he thinks “domain searches,” where people search deeply within specific topics will improve search experiences.
- Mundie talked a lot about the potential to re-imagine education through new collaborative tools. “By putting these tools into the hands of young people, I think we have the opportunity to create substantial change.”
Following the presentation, I got a chance to chat with Mundie for about 30 minutes. We covered a wide range of subjects, including Microsoft’s ambitious foray into the health care field, and how Mundie tries to keep a big company innovative.
Here is an edited transcript of our interview:
Q: Where does Microsoft sees the big opportunities going forward? And what role does research play in those?
Mundie: We’re clearly well into, but have a long to distance to go yet, with the global expansion, if you will, of the company’s business. When you go into emerging countries, the model entry, the order of adoption in technologies, will actually be different than it was in the rich world. So that implies changes in the products, changes in the business models, changes in the go to market mechanisms. A lot of that is underway but that’s still evolving. It’s an area I’ve been involved with for 10 years. But I think it’s paying off reasonably well. But it’s still sort of early days there.
We also keep looking for scale businesses, where we have the technological means of disrupting them from the traditional approaches. That led to a focus on health and education, both of which are new businesses where we brought new products to market this year. With the exception of the defense budget in the United States, the top two expenditures by GDP are health care and education. I think software is going to be the solution, whether you’re a rich company or a poor country, to improving outcomes in those to areas. And it’s something we think is both societally important and an area where scale software as a special place in creating a solution. And that’s why I created those two businesses there. Hopefully that’ll come to pass over a period of time.
Q: How do you even begin to tackle a problem like health care, because it seems so massive?
A: The way we approached it was, to some extant, to say, “What is going to change, or what could be made to change, in the field of health?” That may be different than what other people might have thought. Other people might have gone in and said, “How do we go in and improve something related to the production of health care, or some other incremental solution.” But my sense, and Peter Neupert, who runs this group, supported this same idea, was that health care today is largely a remedial proposition. You get sick, you get some health care. And there’s really not been an adequate focus on wellness or prevention. At the same time, medicine itself, and even the biology that underlies it is all becoming digital. And not just in instruments or diagnostic equipment. But it’s the digits of life like gene sequences and chromosomes and the ability to understand their impact. These have the ability to completely revolutionize medicine.
In some ways, those are all at odds with the way it’s been. It’s really a combination of information technology and this revolution in medicine and biology that create the promise of that change. If you can move to that environment, you have the potential for a radical alteration of how medicine is practiced and the potential for improved outcomes at lower costs.
Then there’s the question of work flow efficiency within the medical delivery system today. Which, in many cases, is impeded because there is no platform that normalizes all the hospital systems. And therefore, you can’t get the ability for what all other sectors have done, which is to sort of cascade all the IT of the software approaches for continual refinements and continual improvements in productivity.
So when we settled on that, we figured that Microsoft could do two things. We could build a super database that could ingest all the information from a hospital environment and create this new abstraction layer. And we did that. And it’s actually been deployed at four of the top 11 hospitals in the United States. And I’m extremely proud of that.
On the other hand, we wanted to focus on the wellness and prevention part. If you want to move to wellness, as opposed to only remediation, that could only happen if you got the individual or the family involved. And so we said, since the company has a global reach and a brand, and a franchise, can we create a web service that would be a point of nucleation for bringing together an ecosystem that would focus on wellness, not just on fixing things when they are broken? And that led to Health Vault.
And so last October, we launched Health Vault, and it’s been a building phase, where we’re building an ecosystem of devices and providers. And now we’re starting to ramp up the public’s involvement with it and also link it some hospitals such that their medical records can be on deposit in this online safety deposit box for health.
Neither of things could have ever emerged out of any of the traditional providers of technology in the health care field, because it was a much bigger problem in terms of technology and scale than any of those companies had historically worked with. But it was certainly within our grasp in terms of scale and technology.
That’s why we felt we could enter, add value, and have a good business proposition. And in a positive sense, disrupt that normal pattern of the way people traditionally approached the problem.
Q: How do you try to create a culture of innovation at a large company such as Microsoft?
A: I think there are several different components for retaining that capacity in a company. One part that Bill had the foresight to do, and that I’ve always supported, is to have a pure research function, which is a place where some significant number of smart people are tasked with being subject matter experts and pursuing things that don’t have any immediate product requirement. And so that builds up and sustains the ability to disrupt or respond. And certainly provides, in many cases for us, a steady flow of things that we can use to refresh, or incrementally improve some of those legacy products that you mentioned.
So, there are three aspects. The health care one, that’s more of a disrupt. And we took in stuff from both the research world as well as other development activities to do that. It’s a new business.
Many, many features you would find, version by version in our traditional products, have their genesis in the research group. And there’s a big program of technology and transfer that we really work hard to shepherd that tries to accelerate that transfer of ideas for new technologies that represent incremental solutions to problems that exist in those products.
So the help system for Microsoft Office or Windows, you use to be able to just search by words. But now you go to the box and type a natural language question in and it parses the question and gives you different alternatives you can select from and that can get you more help. And that ability to convert that from a text look up to a natural language query came from the research guys. And that’s just a tiny example. There are literally hundreds of such things.
And then there are the responsive ones where if the market shifts in a particular direction, a lot of times our ability to be responsive in a timely way, and remain a player in evolving areas of the market comes from the fact that we have that research capability.
The second is that you have to have activities, or places in the company where people who are less risk averse can congregate, if you will. And where you have what it takes to let them start and try new things.
And so, as I said, part of my job for more than 15 years was to be a place in the company where we could do start-ups inside the company. And so while other people do that, and the product groups do a little bit of that, I probably started more new business activities in 15 years than the rest of the product groups have combined. And so, that’s another thing we can do to create a place for that.
And then the other is how management characterizes the problem for even established business groups over time. We work harder and harder to get them to deploy people against future questions in their products space that are not strictly limited or confined to working on the next release.
Taken together, these things give us that capability to innovate both big and small and over sustained periods of times.
Q: You’re on the road this week doing a tour of colleges. Why?
A: It’s something Bill and I have always done and relished. And I will continue to do even in his absence. And the goal is to just pause. And have a dialogue with the administration and the faculty and the students so that I can hear from them what the issues are on the campus these days, both sociologically and technically. And that they can hear from me about a set of problems or challenges or opportunities that I can at least make the case are going to happen but which they aren’tt going to get taught in any class in the university. And that’s sort of a value exchange. We systemically say, “Let’s take a week, and talk to our friends in academia, and figure out where there’s an interesting dialogue to have.”
Q: Have any themes emerged over the course of your travel and conversations on campuses this week?
A: A couple of things. I think of the universities I visited, we have had over a number of years some significant collaborations with Berkeley. And I think there is a genuine appreciation for what has come from that partnership between Microsoft and the university on some of these projects. At Berkeley, we have a number of things, such as modeling the water table of California, modeling the atmosphere and other things. There are a bunch of these challenging problems where Microsoft research is really proactively engaged. And I think people really appreciate that. And I think that comes across now in the interactions.
I think another thing, that might be a pattern that emerges, is the recognition that computation is at the heart of every significant research activity at all these schools. And yet, it isn’t at all clear how they’re going to continue to get access to the level of computational facilities necessary to advance their science. And that speaks to the broader policy question in the United States and elsewhere in the world where governments have largely reneged on funding basic research and some of the facilities for basic research. And in a world today where having sufficient computational facilities is the basic ante, I think it’s becoming very clear that we don’t currently have, and are not on a path to having, those facilities to the degree we probably need them.
Q: Is that a place where Microsoft or private industry can step in?
A: In a way, I don’t think private industry can be expected to substitute for government in the funding of primary research. Microsoft is one of the very few companies left that does any publishable research in the field of computer science.
Many, many other companies have eliminated it completely, or have said, philosophically they’re not going to do it. They are going to take what they can get through acquisitions. And so I think we are in an almost unique place in terms our collaborations around the computer science research programs. Intel is certainly involved with the hardware level. IBM and a few others are still involved, mostly at the hardware levels. But on the software side, we’re certainly the biggest and the best there.
Q: Is the issue basically that institutions lack enough computing horsepower to do the calculations required?
A: People are now fully awakened to the importance that computing and software are playing in their ability in different fields to be able to have significant breakthroughs in those fields. But the more they get into it, the more appetite they have for what I’ll call “high scale computing.” And despite steady progress at making computers faster and cheaper, which has been quite remarkable, we’re now looking at an expansion of the domains to which this is now applicable that go way beyond the traditional megaflop page type of computational modeling. Data mining may be as important in some of these fields, or more important, than any of the traditional types of models and visualizations.
So I think you broaden the pipe, and you have essentially an almost exponential increase in the appetite scientists have for being able to use these facilities. So I think it’s going to be a persistent issue.