By Mark Leibovich
Washington Post Staff Writer
Monday, April 5, 1999; Page A01
REDMOND, Wash.—The company well known for its
aggressive domination of the software world has set itself a
new target: the best minds of academia. With cash, stock
options and the promise of vast resources, Microsoft Corp. is
luring faculty elites to its research center at a pace so fast that
some campus departments say they're being picked clean.
Last month Microsoft hired Yale University mathematician
Lazlo Lovasz, recent winner of his field's prestigious Wolf
Prize. He will start here in June and will join, among others,
Fields Medal-winning mathematician Michael Freedman, a
recent arrival from the University of California-San Diego;
and MacArthur fellow Jim Blinn, a computer graphics expert
from the California Institute of Technology.
Microsoft Research, known as MSR, is aiming for 600
"faculty" by the end of next year. It already is among the
biggest computer science laboratories in the world, with 350
researchers. While other private research labs, including
Lucent Technologies Inc.'s Bell Labs, International Business
Machines Corp.'s T.J. Watson Research Center and Xerox
Corp.'s Palo Alto Research Park, recruited faculty stars long
before Microsoft, no company has raided universities so
brazenly, school officials said. And none has dangled stock
options that have mass-produced so many millionaires.
Microsoft Research is seeking big names in computer science
foremost, but also leading thinkers in graphic arts, linguistics,
biology and mathematics. While they may never write a piece
of software, they could hatch ideas that the company's
programmers one day might turn into big-selling products.
Here on a lush-green campus crisscrossed by shuttle vans
painted with the Microsoft slogan, "Where Do You Want to
Go Today?," the software giant is practicing its "dinner party"
philosophy of basic research: By assembling the right mix of
brilliance, eliminating the usual concerns (teaching, tenure and
grant-proposal writing) and leaving people free to develop
ideas, the next wave of great computing could flow.
But some school officials said Microsoft's elite guest list is
depleting their own.
"Microsoft Research has become a parasite on the academic
establishment," said Jim Morris, chairman of the computer
science department at Carnegie Mellon University. "They are
eating our seed corn. If you take away great people from
schools and put them in a place where they're not teaching
anyone, who will train the next generation?"
Microsoft officials deny they've spurred a university brain
drain. Nathan Myhrvold, the company's chief technology
officer, said MSR probably hires six full professors a year.
The rest are junior-level professors and PhDs. "We haven't
taken all that many top people in all," he said.
Some universities scoff at this, saying that Myhrvold is
counting only fully tenured professors, a shallow measure of a
department's teaching cache. And a few departments have
been hit undeniably hard: Carnegie Mellon, for example,
beginning in 1991, when computer operating system expert
Rick Rashid left to become MSR's first director. Since then,
"between 15 and 20 top people" followed, said Raj Reddy,
dean of Carnegie Mellon's computer science school. "You
never can completely recover from this," he added.
Still, Reddy is a member of Microsoft's high-tech advisory
board, a group of deans and department heads -- many of
whom, like Reddy, have been hurt by the company's talent
hunt. While seemingly at odds, Reddy's joint roles underscore
the mixed feelings of many in academia toward Microsoft. On
one hand, the company is depleting universities' staff. But on
the other, "Microsoft is not a clear-cut force for bad," said
David Dopkin, Princeton University's computer science
chairman. The company, he said, should be praised for
contributing to basic research at a time when other
corporations are cutting back.
Microsoft spends $3 billion a year on research and
development. Carnegie Mellon's total endowment, by
comparison, is $600 million. In recent years, the company has
opened basic research labs in Cambridge, England; Beijing;
and Silicon Valley. The main facility in Redmond, Building
31, has a giddy Nerd-in-Wonderland quality. Barefoot PhDs
mingle with computing legends such as Gordon Bell, the
computing pioneer of the 1960s, who last month was seen
fumbling with an office printer.
Company officials insist MSR is not at odds with the broader
goals of academia. Researchers often retain affiliations to their
former schools and the papers they publish benefit scholarship
in general. Likewise, the company has courted university
favor with $80 million a year in cash and software "gifts."
This is not pure altruism. Microsoft wants university scientists
to invent things that will support its software. (Until the
practice was stopped last year, Microsoft paid faculty
members a $200 "stipend" to lecture at conferences at which
their speeches were deemed sufficiently pro-Microsoft).
While universities might bemoan Microsoft's recruiting drive,
no one is forcing the professors to leave. In fact, they are
inundating Microsoft Research with job feelers , often through
former colleagues. Staff recruiter Steve Clyne said MSR
receives 50 to 200 such inquiries a month, and 80 percent of
those offered positions accept.
Base salaries -- well into six figures for senior researchers --
are on par with those at most major universities. But stock
options are the X-factor, and it doesn't take a Wolf Prize
winner to calculate the profits reaped by Microsoft
options-holders this decade.
"Stanford has stolen people from us before, but at least that
was a level playing field," said Princeton's Dopkin, who has
lost two top professors to Microsoft. "[Microsoft] is implicitly
saying, 'If you come here, in a few years, you will probably
be a millionaire.' "
Researchers insist they've not been lured merely by wealth. In
a dozen interviews last month, they characterized Building 31
as research nirvana, where ideas can transcend the numb
abstraction of academia and shape a mass market. Several
MSR innovations have found their way into software. For
example, Microsoft Office's "grammar checker" feature was
created with the help of the lab's natural language group.
"To me, this corporation is my power tool," said Steven
Shafer, a Microsoft researcher who came from Carnegie
Mellon in 1995. "It's the tool I wield to allow my ideas to
shape the world."
The abundant resources make for a collegial environment.
"Fights in academia could get pretty vicious because the pie
was so small," MSR cryptographer Josh Benaloh said.
Yet there still is ample pressure. Researchers periodically
undergo reviews from group leaders and managers. They are
rated on the quality of their ideas and papers, how active they
are in their disciplines (what conferences did they chair?) and
how well their work has contributed -- or will contribute -- to
the company's pursuits.
They also must endure a less formal but perhaps more
profound evaluation. It comes from a one-man faculty review
board named Bill Gates.
Two or three times a year, the World's Richest Man holes up
in a room with piles of accumulated reading. Called "think
week," Gates uses the time to devour trade journals, memos
and research papers -- many of them produced in Building 31.
He then volleys e-mail with his researchers. Gates recently
argued with one group on a physics phenomenon called
"incipient infinite cluster," according to MSR mathematician
Jennifer Tour Chayes, who was at the meeting. By the end of
the discussion, several white boards in Gates's boardroom
were covered with math equations.
After think week, MSR director Rashid receives a flurry of
e-mail from Gates. "He'll tell me what he's excited about and
what he's not excited about," Rashid said.
If one's work falls in the latter category, it can be
demoralizing. But it's healthy, too, Rashid said. "Bill's an
intelligent guy, and if he has an idea, you might want to listen
to him . . . If people don't tell you you're stupid periodically,
you lose your incentive to try hard."
Building 31 also is not immune from the siege mentality that
generally can mark Microsoft. "We're very much a whipping
boy in the academic community," said senior researcher Dan
Weise, who is exploring new software de-bugging tools.
Microsoft's row with the Justice Department hasn't helped.
Weise said he tries to stay focused on his research, not the
antitrust trial. "I can't help it if our lawyers are clueless,"
Weise said.
Researchers said the principal difference between Microsoft
Research and universities is the lack of students. Some
describe this as a big relief. Former Carnegie Mellon professor
Shafer, for example, said teaching steals from research time.
But others lament the loss and miss the implicit altruism of
teaching in an academy.
"Sometimes I feel like I'm abandoning my duty to students,"
Benaloh said.
"My annoyance with Microsoft is that there aren't 300 extra
great PhDs waiting to replace the people they steal," Dopkin
said.
UC-San Diego math chairman Jeffrey Rennel said the
departure of a faculty star, such as recent MSR hire Michael
Freedman, can multiply. "When you lose a Freedman, you
might have 13 or 15 top people under him whom he'll never
teach again," Rennel said. Often, he said, they leave, too.
Carnegie Mellon's Reddy said that when he complains to his
friends at Microsoft about faculty raids, they seem
incredulous. "They always talk about it being an open
market," he said. He jokes that perhaps Microsoft should just
offer $1 billion for Carnegie Mellon. Microsoft's Myhrvold
rules this out, saying, "I would never move to Pittsburgh."
But a department must learn to reinvent itself, Reddy said,
because Microsoft Research isn't going away. After it took
some of Carnegie Mellon's key robotics and
speech-recognition experts, Reddy looked to new fields. He
calls it the "natural order of things."
Now, Carnegie Mellon boasts a widely hailed electronic
commerce institute, where graduate students learn the fine
technical points of business on the World Wide Web. It is a
blossoming field and Reddy realizes that Microsoft is
especially interested. They could be coming around again
soon.
Copyright 1999 The Washington Post Company