more on Talent leak drains AT&T think tank

[Dave Farber <dave () farber net>]

Delivered-To: dfarber+ () ux13 sp cs cmu edu
Date: Sun, 21 Mar 2004 16:07:40 -0400
From: Claudio Gutierrez <cgutierrez () improvement cl>
Subject: Re: [IP] Talent leak drains AT&T think tank
To: dave () farber net

Dave
	reading through the story we can learn: "Others have gone to Israel's 
Weizmann Institute of Science and to Microsoft Research -- maybe the only 
shop rich enough to support basic research as Bell Labs did in the Cold War."

Regarding Microsoft Research, Financial Times published an article 
about  them just a couple of weeks ago. I think it is worth reading

Good old-fashioned innovation (subs req)
http://search.ft.com/search/article.html?id=040312001006&query=microsoft+research&vsc_appId=totalSearch&offset=10&resultsToShow=10&vsc_subjectConcept=&vsc_companyConcept=&state=More&vsc_publicationGroups=FTFT&searchCat=0

Innovation used to be easy, in principle if not in practice: build a 
laboratory, hire brilliant scientists, treat them well and wait for ideas 
to flow. The technology giants of the last century - DuPont, General 
Electric, AT&T, Xerox, International Business Machines, Glaxo - relied on 
this recipe to find breakthroughs and fuel growth.

But times have changed. The generation of companies that flourished from 
the 1980s onwards - Cisco, Oracle, Intel, Nokia - has relied instead on 
acquisitions, corporate venture capital, and academic partnerships. While 
each of these companies still does in-house research, today's technology 
titans have all eschewed the temptation to build big labs.

All, that is, with one mighty exception: Microsoft.

The Seattle-based software group has spent the past decade assembling a 
team of scientists to rival the great industrial research labs of 
yesteryear. Microsoft Research now has 700 researchers, including many of 
the biggest names in computer science, and an annual budget estimated at 
between $500m and $1bn (£280m-£560m).

"Microsoft really is an anomaly," says Hank Chesbrough, executive director 
of the Centre for Technology Strategy and Management at Haas School of 
Business in California. "It is about the only company in the world that is 
still trying to round up the best and the brightest and put them on the 
payroll."

The drawbacks of this approach are well documented. Big research labs can 
become inward-looking, obsessing over their own ideas while discounting the 
inventions of others. For companies trying to stay at the cutting edge, 
this can be disastrous. For example, in the 1970s and 1980s IBM failed to 
grasp the importance of the shift from mainframe computers to client-server 
architecture - even though IBM Research virtually invented the field of 
computer science.

Moreover, in the modern era of start-ups and venture capital, can any 
company hope to corral the best researchers and the brightest ideas within 
its own laboratory?

The breakthrough technologies of the digital age did not arise from orderly 
industrial research. They came instead from Silicon Valley's chaotic mix of 
accident and experimentation.

Notable examples of this include Apple (developer of the first mass-market 
personal computer), Netscape (progenitor of the first web browser) and, 
most recently, Google (undisputed champion of internet search).

"Technology firms must break away from internal innovation and redefine 
their 'innovation perimeter' to include ideas that come from outside," says 
Georges Haour, professor of technology management at IMD in Lausanne.

In his new book, Prof Haour calls on companies to use a range of techniques 
- including corporate venture capital, spin-outs, acquisitions and 
licensing - to blur the line between "inside" and "outside". Against this 
background Microsoft looks distinctively retro, a throwback to a golden age 
of orchestrated research.

If Bill Gates is Microsoft's chief software architect, the architect of 
Microsoft Research is Rick Rashid, a computer scientist hired from the 
faculty of Carnegie Mellon University in 1991. The lab's growth came in two 
distinct spurts. Between 1991 and 1995, Mr Rashid built a team of about 150 
scientists, almost all of them based at corporate headquarters in Redmond, 
outside Seattle.

Microsoft was determined, he says, to avoid the mistakes of Xerox's Palo 
Alto Research Centre, the California lab that in the 1970s laid the 
foundations for personal computing but failed to persuade its east coast 
parent to turn its inventions into products.

The lesson? Proximity matters. By locating its researchers cheek by jowl 
with business managers, Microsoft hoped to encourage more effective 
transfer of new technologies from its labs to its range of products.

The second growth spurt ran from 1997 to 2000, when Microsoft Research's 
headcount rose to 650. New labs were opened in Silicon Valley, Beijing and 
Cambridge.

The management philosophy, says Mr Rashid, has been consistent throughout: 
select a subject for research (there are currently more than 60 on his 
list, from artificial intelligence and audio to wireless and web services), 
hire a top name in the field, let them set their own research agenda and 
provide the right environment for them to shine.

"I see myself as a ringmaster," says Andrew Herbert, director of 
Microsoft's Cambridge Lab. "Here are a collection of great performers. My 
job is to make sure they can put on the best show they possibly can."

If this sounds too good to be true, consider that one of Microsoft's most 
eminent researchers, Jim Gray, has spent the past few years tying together 
the world's astronomy databases into what he describes as "the worldwide 
telescope".

Do not expect to find this feature embedded in the next generation of 
Windows. The project, says Mr Gray, a veteran of AT&T's Bell Labs, Xerox 
Parc, IBM Research and numerous universities, sprang from his interest in 
large, distributed computer systems.

This intellectual freedom - combined with freedom from the grind of 
applying for research grants - is what makes this individualist happy to be 
a part of Rick Rashid's international army of brainiacs.

Microsoft's researchers are brilliant, well-funded and free to advance what 
Mr Rashid calls "the state of the art" in software and computer science. 
Their work is published in peer-reviewed scientific journals, presented at 
conferences and discussed with the hundreds of PhD students who do 
internships at Microsoft each year.

Prof Chesbrough remains sceptical: "When you bring in so many really 
talented people there is a natural tendency for them to justify themselves 
by their own ideas and inventions, not by their ability to spot really 
important ideas. There is a marginal tendency to talk down what is 
available on the outside."

Gary Hamel, the academic and consultant, points to another concern: "How 
many companies [such as Microsoft] with more than 50 per cent market share 
and more than 50 per cent margins have had a mean-ingful second act? None."

The reason, he says, is that most important new businesses are built 
through a process of experimentation and rapid learning. Big companies tend 
to be bad at both. They do not experiment enough in new markets and they do 
not learn fast enough from the results.

On this diagnosis, no matter how good its research capability Microsoft 
will tend to see opportunities late, by which time it will face stiff 
competition.

Sure enough, much of Microsoft's history has been spent playing catch-up 
with market leaders. Whether in computer operating systems (Windows versus 
Apple's MacIntosh), desktop applications (Word versus WordPerfect), 
commercial databases (SQL versus Oracle), web browsers (Explorer versus 
Netscape), internet portals (MSN versus AOL), or hand- held organisers 
(PocketPC versus Palm OS), Microsoft has played the role of "fast follower".

But so far this strategy has been hugely successful. The brainpower 
corralled within Microsoft Research - combined with the company's dominance 
of the PC operating system market - has allowed it to arrive late at the 
party and still go home with the belle of the ball, almost every time.

Back in Redmond, Mr Rashid is reluctant to endorse the idea that Microsoft 
Research exists mainly to enable a strategy of catch-up. However, he does 
concede: "We are a reservoir of technology and people for when the world 
changes. When you suddenly decide that you need to do something new, if you 
don't have people who understand it, you can't do it."

Last year, for example, Mr Gates declared that it was a strategic priority 
for Microsoft to take on Google, which dominates the emerging market for 
internet search and, unlike so many internet ventures, has built a 
profitable business on the back of its market position.

The call went out to Microsoft Research. Starting with a "summer school" in 
Redmond attended by 30-40 researchers, brainpower has been focused on 
out-thinking and out-engineering the latest upstart.

The danger, says Mr Hamel, is that as Microsoft tries to grow from its 
already very large base, it will face much more formidable competitors. 
Beating AOL, Palm or Google is one thing; taking on the likes of Sony (in 
games consoles) and Nokia (in mobile phones) presents tougher challenges 
altogether.

Microsoft will have to latch on to new ideas and act on them more quickly 
than ever before. Will Rick Rashid's big lab then prove to be a help or a 
hindrance? Think of it as a big experiment in the management of innovation.

-------------------------------------
You are subscribed as interesting-people () lists elistx com
To manage your subscription, go to
 http://v2.listbox.com/member/?listname=ip

Archives at: http://www.interesting-people.org/archives/interesting-people/