Peter Lee’s mission at Microsoft is to think big, think far and think unconventionally — and to get others to do the same.
As head of Microsoft Research (MSR), Lee is shaking things up at one of the largest computer-science research organizations in the world. Its researchers’ work has been incorporated into just about every company product, from Azure to Xbox, Bing to Windows.
But the division has also taken criticism for not helping Microsoft innovate enough, allowing companies with smaller research and development budgets, such as Apple and Google, to leap ahead in crucial areas.
Shareholders have wondered whether Microsoft gets enough in return for the $10 billion a year the company spends on R&D — a relatively small portion of which goes to Microsoft Research.
- Nurse dies from injuries in attack near CenturyLink Field
- Woman knocked unconscious by falling drone during Seattle's Pride parade
- Residents return to ‘war zone’ in wake of Wenatchee wildfire
- ‘Historic’ tuition cut sets state apart from rest of U.S.
- Tukwila group to submit expansion application to NHL
Most Read Stories
The task before Lee, who became head of the organization last year, is to balance the pure research for which his division is best known with more immediately applicable work, all the while focusing resources on what could best lead to the next big thing.
Some of the changes recently put into place, including embedding at least one of Microsoft’s advanced researchers in each of the company’s product planning teams, have come from above, as a continuation of a larger companywide restructuring that started last year.
That reorganization, begun by former CEO Steve Ballmer and continuing under new CEO Satya Nadella, is geared toward fostering more innovation and collaboration.
Others come from Lee himself, such as making sure his managers have strategies for, and can show achievements in, each of four quadrants: Research that solves immediate problems; improves existing technologies; is disruptive and game-changing; or is purely curiosity-driven and exploratory.
The intent is to have a portfolio of research activities that together form an innovation pipeline, where blue-sky ideas feed into industry-changing inventions, which then help with the shorter-term activities.
Some of those changes have caused discomfort. But that’s not necessarily a bad thing, Lee believes.
“It’s delicate because, on the one hand, you need creative discomfort to spark creative thought, to think out of the box,” Lee says. “On the other hand, researchers really need stability. It’s always a struggle to find the right balance.”
Creating a safe place for creative discomfort, thinking unconventionally and seeing the practical potential in abstract ideas are all hallmark Lee.
Those traits were evident from his days heading Carnegie Mellon University’s computer science department, to his stint at the U.S. Defense Advanced Research Projects Agency (DARPA), and now at his job leading more than 1,100 advanced researchers working in 55 areas of study in a dozen labs worldwide.
This is a man who says he likes to “encourage wackiness.” While at DARPA, he launched a nationwide search for 10 giant weather balloons to convince four-star generals of the power of social media.
Since 2010, when Lee arrived at Microsoft to head the Redmond lab of Microsoft Research, he has looked for surprising technological advances not just from fellow techies, but also artists.
“If you look around at the world and where technological surprises are happening, one place is in the art world — artists who are using tech to create new experiences,” Lee said.
A tall, salt-and-pepper-haired man with a playful sense of humor and an affable demeanor, Lee, 53, speaks engagingly on a wide range of topics.
He easily segues from his expertise in software security and reliability, to developments in machine learning that excite him, to the joy of rounding a corner just right in a fast car (he used to race cars as a hobby).
“Peter has the ability to get other people to see past where they normally look,” said Seth Copen Goldstein, an associate professor of computer science at Carnegie Mellon. “He is aware of, and interested in, many, many things. And he is able to put them together in ways that are interesting and that are useful.”
Young “coding geek”
Lee’s interest in science crystallized when he was 9 years old, watching the moon landing on TV at his home in Houghton, Mich.
“I got so sucked in,” he recalls. “I was drinking Tang and eating Space Food Sticks.”
The son of a chemistry- professor mom and a physics-professor dad, Lee figured he was “probably destined to grow up a research scientist.”
But it wasn’t until he watched the moon landing that science became real to him — something that could get people to the moon or could create ancillary products.
“I didn’t have the thought that I wanted to be an astronaut,” Lee said. “I did have the thought that I wanted to be part of the organization that makes these things.”
Another turning point came in high school when his father enrolled him in a summer computer-science camp.
“I became a coding geek,” Lee said.
He earned bachelor’s degrees in mathematics and computer sciences, and a Ph.D. in computer and communication sciences, all from the University of Michigan at Ann Arbor.
After earning his doctorate, he joined the faculty of Carnegie Mellon’s computer- science department, eventually becoming department chairman and the university’s vice provost for research.
Carlos Guestrin, a former Carnegie Mellon colleague of Lee’s who is now CEO of Seattle-based GraphLab, recalls that Lee was always interested in “big projects that can have big impact. He always has his eye on what the next big thing is going to be.”
Saluting social media
In 2009, Lee took up an offer to establish a new office at DARPA focusing on technology with potential for disruptive transformation and surprise.
The Iranian “Twitter Revolution” was happening at the time and Lee thought social media could be of use to the military.
Military brass, however, was not so sure.
Lee and a group of DARPA interns he was guiding decided to send up 10 red weather balloons moored at fixed locations across the country and offer a $40,000 prize to whoever could locate all the balloons first.
Intelligence agencies had told DARPA that the balloons were too small, and the search area too big, for spy satellites to be effective, said Daniel Kaufman, director of the Information Innovation Office at DARPA, a colleague of Lee’s at the agency.
But in less than nine hours, a team from the Massachusetts Institute of Technology, which used social media as part of its winning strategy, located all 10 balloons.
The project was “awesomely fun,” Kaufman said. Moreover, “it opened up the eyes of the military because they started to realize that potentially some aspects of social media could be used as a new source of intelligence, a new way to see things.”
Lee tries to bring that same spirit to Microsoft Research.
When he first arrived, some colleagues told him about their work seeking to improve teleconferencing.
“I asked why they were doing this. They said: ‘To make business meetings more efficient,’ ” Lee said. “Which is fine. But ultimately, it isn’t very surprising. The chances for technology to surprise aren’t driven by the business-meeting scenario.”
Lee, who is married to artist Susan Lee and has a 17-year-old son, asked whether the technology would enable musicians or sculptors from around the world to collaborate in real time, or children in different locations to play together.
“Those ideas are uncomfortable and wacky. But that way of thinking led to the idea of low-latency teleconferencing,” Lee said. “When you teleconference, even with Skype, there’s always a bit of a lag. If you want to play a duet with someone, that latency makes things hard.”
Thinking through such different scenarios can help researchers develop new ideas on reducing latency — something that could lead to better Skype videoconferencing or more responsive wireless connections between Xbox game controllers and consoles, he said.
Next big things
These days, Lee sees machine learning and ambient computing as two next big things.
Machine learning is the ability for computers to automatically learn from data to continually improve their future performance. Ambient computing refers to having computing devices, embedded all around in the environment, that can understand people’s intent and operate on their behalf.
Taken together, he said, “we see an era … where we move beyond the personal computer to a world where computing all around us becomes more personal.”
In addition, Lee has established a Special Projects group within Microsoft Research to focus specifically on projects with the potential to expand what’s possible with technology.
Beyond science, what gets Lee excited is when his research arm — representing less than 1 percent of all Microsoft’s employees — gets recognition from the rest of the company.
“I think people in MSR are very aware of how connected they’ve been to the company,” Lee said. “But they’re also aware that there’s an external perception — I think it’s a misperception — that they haven’t been.”
Making that connection more visible has become easier since Microsoft researchers have become embedded in product planning teams.
“Satya has said that a big part of his job will be to accelerate the company’s ability to bring innovative products to Microsoft’s customers,” said Harry Shum, Lee’s boss, who started the effort to embed researchers in product planning teams after he became Microsoft’s executive vice president of technology and research late last year.
“Both Satya and I believe that there are opportunities for the company to do an even better job of unleashing the value in Microsoft Research than it has in the past,” Shum said.
Some of the matchmaking between researchers and product groups arise naturally, based on mutual interests or already established relationships. Other situations may require “wackier” pairings, as Lee puts it, such as matching a sociologist with insights into American teenagers with a mobile-apps group.
“It’s now assumed that if Microsoft Research wants to participate as first-class citizens in product development, the door is open,” Lee said.
Janet I. Tu: 206-464-2272 or firstname.lastname@example.org. On Twitter @janettu.