Impressions from my year at Microsoft Research

I am not a fan of blogs, but for my own interest I recorded some impressions from my year at MSR Redmond.

For a reader who doesn't know me, it is important to realize that in many ways I am the complete opposite of a typical MSR visitor. I am 57 instead of 27; I own neither cell phone nor laptop; I like to work on topics no-one else in the world is working on; and I'm a Mac dude. So, for young people in particular, your experience at MSR would likely be quite different from mine.

Not an evil empire ....

Though I liked to make "working for the Evil Empire" jokes, it ain't really so. In the entire year, the only whiff of sulfur was an internal email encouraging employees to expose interns to major MS technologies -- which is fine and sensible -- but continuing to say "and actively discourage competitors' tools such as Java, Eclipse, Matlab, Latex and Google". Harumph. But after all I could probably find some internal Berkeley emails about discouraging grad students from going to Stanford .......

... but in fact a great employer ....

MS is in many ways a wonderful and creative employer. Here is one minor instance that makes a good story. In casual conversation with a stranger, if asked my occupation, I had two possible answers: Statistics Professor, or working for MS. Each has a stereotypical response, either "I took a college course in statistics and couldn't understand it", or a complaint about some piece of MS software. Now statisticians don't do anything about this situation but MS does. They give employees 3 cards, to keep in their wallet, intended for such occasions; one gives a card to the stranger, entitling them to one free phone call to get technical support on any MS product. Great PR!

... with great employees.

Amazingly, I came across no individual (here I mean from all MS employees, not just researchers) who was less than competent at their job. This contrasts with my experience in academia, where 5% of employees should have been fired years ago. Somehow MS does a great job of selecting who to hire in the first place.

Of course, as in any large organization the systems may be less than perfect. The story I tell, to illustrate a complete opposite from academia, involves moving office (to 3 offices down the hall). In academia there would be no alternative to physically moving stuff oneself; at MSR we were explicitly forbidden to move stuff ourselves, so instead of taking 5 minutes to move my few books/folders myself, it occupied a couple of hours of filling in online forms, having empty boxes delivered, packing and labeling two boxes, waiting for them to be moved, and then unpacking.

Non-technical talks kicking off a book tour

As well as many technical talks from different MSR groups, there are frequent general talks (some addressing the Microsoft campus as a whole, others aimed at MSR specifically), often by an author of a newly-published book who is kicking off a book tour. I found these fell into two opposite categories.

1. Would-be best sellers on technology and business/management, of the genre one finds in airport bookstores.
The most polite descriptive phrase I can devise for these is "overwhelming hype around some trite catch phrase", as illustrated by the following example.

ABSTRACT: The world has changed profoundly, and the old tools that led to success in the world of "push" won't work anymore. "Pull" helps us to understand the shift we are experiencing and provides us with a new understanding of the implications of how our digital world really works. What can we do to thrive in the environment dominated by the forces of pull. With pioneering research we can show how to access people and resources when you need them, attract resources you didn't even know existed and achieve potential with less time and more impact. Few of us are systematic in how we use the tools available to us, and no institutions are effectively dealing with the startling changes wrought by new technologies and the attitudes they encourage: pull will change everything.

OK, so this talk was, like Wagner's music, "not as bad as it sounds" - cute stories, but without any coherent intellectual theme beyond phrases like "instititional innovation" and "scalable capability building".

2. Fortunately, there were also genuinely interesting and less commercially-oriented talks, for instance by

Paul Davies, the Chair of the SETI Post-Detection Taskgroup, on The Eerie Silence: Renewing Our Search for Alien Intelligence

Shane Harris on The Watchers: The Rise of America's Surveillance State

and Greg Bear, one of my favorite SF authors.

Theory ≠ product

MSR Redmond is made up of about 40 groups each with about 10 permanent employees, each group nominally working on some particular topic. Of course this is the only feasible style for an outfit like MSR -- how else could one structure it? But to me it suggests a danger of individuals and groups getting absorbed into topics that fall between "theory" and "product". More precisely, it's fine if such topics represent a moving wave propelled by outside developments but not if they stagnate into some inward-looking topic.

Elsewhere I have written a rant theory is not a product which is aimed as those of us interested in theory -- do not treat it as product development! Now there's an obverse meaning: "if you're interested in product development, don't treat it as theory". And indeed MSR management has been saying things like ....[let's make] an effort to capitalize on the intellect and abilities we have brought together in MSR to take on problems that are not just "difficult" but which some people might call "impossible" having in mind some actual product, not mere theory.

Did you meet ......... ?

No I didn't meet Bill Gates or Steve Ballmer, and wouldn't have had anything substantial to say if I had met them.

The general ethos is that the multi-millionaires of early MS seek to do something substantial, a prime example being Nathan Myhrvold's founding of Intellectual Ventures. For reasons unconnected with my MSR stay I took a tour of their Intellectual ventures lab which is an amazing place -- where Edison would be working if he were active today. At an opposite end of some spectrum, for some reason I don't recall we got a briefing from Nick Hanauer and Eric Liu on their (progressive) True Patriot project.

Research areas

MSR's own site outlines the different research fields they cover. I am hardly revealing secrets by saying that there is lot of interest in cloud computing and in NUI (natural user interface), the latter exemplified by Project Natal, just launched publically as Xbox Kinect. Amusingly, the only products in the pipeline that are actually secret within MS are the games. I guess they're not worried about employees shooting their mouths off in a bar about cool features of Windows 8, but as for Halo 4 ...........