Last Friday Matt Welsh, my former advisor at Harvard, announced that he was leaving permanently for Google, some five months after being tenured. He discussed his reasoning on his blog on Monday.
Matt's departure brings up all kinds of issues appropriate for the first post on my new blog — how much hacking should computer science professors do, the real difference between academic and industry research, reflections on the current funding climate. But as the student that worked with Matt from (almost) his first day at Harvard to (pretty much) his last, I'm drawn to begin with a personal note.
During my early years working with Matt some of his Berkeley buddies use to refer to me as "mini-Matt", citing physical (and perhaps temperamental) similarities. People even confused talks that the two of us had given, at times attributing my (fantastic, mind-blowing, life-changing, etc.) presentations to him.
And while I chafed at these comparisons (as, I imagine, he would have), Matt was playing a formative role in my life. He recruited me as a fresh-faced RA my first summer out of college, the summer he arrived at Harvard. He employed me during the next year while I decided what to do with myself. He talked me into applying to graduate school, and made sure that, even with my Physics-degree and lack of CS background, I was accepted somewhere. He and I worked together closely on projects, most notably MoteLab and our volcano monitoring work. He pushed me to be more than "just a hacker", the derogatory label he would throw at me when I was being sloppy or unscientific.
He wasn't easy to work with. Neither was I. There were loud arguments, office ejections, firing threats. But there were also accomplishments and adventures. And so despite the bits of unpleasantness, I loved it.
Eventually all of that friction started to wear me down, as it was bound to, and Matt and I started to go our separate ways. Our interactions were as fraught as ever, but missing were the calmer, friendlier moments that had provided balance at the beginning. By the end of my Ph.D., we had signed a mutual non-irritation pact: I was free to pursue my own ideas, regardless of how misguided he found them; and he was free to point out how dumb they were, expecting not to deter me. I think, in the best case the trajectory of the apprenticeship we call graduate school takes the adviser from mentor to colleague, the advisee from dependence to independence. In our case, I achieved a measure of independence, but we never became colleagues. Since I've graduated our relationship has become calmer. Recently he's been dispensing good career advice, and I hope that that continues.
Looking back on my experience in the light of Matt's recent decision, some things make a bit more sense. It's hard to work joyfully on work you don't enjoy, and Matt's frustration's with academia probably played a major role in the negative environment that I experienced. Some of his younger students hadn't been there in the early years when he was more open, more fun, easier to interact with. They always assumed that either he was just difficult — which I knew wasn't true — or that the tenure push was affecting him — which I found plausible but not sufficient to explain the change in his behavior.
So why did we continue on? Obviously there was something in it for both of us: a degree for me, tenure for Matt. And these are prizes of sufficient worth that people sometimes seem to elevate their pursuit over their own happiness. This might work temporarily, but succeeding in each has very different temporal implications. Graduation is a point of departure, providing the freedom to go on and do something else. Tenure wants you to commit to continuing what you are doing, requiring that you hide your doubts behind a veil of certitude. Matt is airing those doubts now, but a year ago he was recruiting students and post-docs while allowing colleagues to drape their arms around his shoulders and say to Harvard: "This is our guy, deserving of a lifetime appointment!" Here it seems that late, while always better than never, might be less than ideal.
Ultimately, the degree might help cure what ails you. Tenure won't.
After graduating last spring my wife and I spent the summer in Sydney, in a beautiful house on the top of a 500-ft hill facing the ocean. We fed lorikeets in the morning and fell asleep to the surf at night. I hadn't planned on doing much with myself, which was good, since I was surprised at how many complex feelings I had to process. It took me most of the first month we were there before I could relax fully. (And then it was absolutely beautiful.) All to say that I know how both freeing and difficult it can be to escape an unpleasant situation, and I wonder if this isn't how Matt felt when he started at Google, or how he feels now.
When I returned in August I walked over to MD one day before starting at MIT to pick up some mail. I went up the side stairway and navigated down the hallway back to my old lab. Inside: all of Matt's students, the air thick with tension. And then I realized: it was Friday. Friday! Friday was the day that Matt held one-on-one meeting with his graduate students, and our weeks tended to pivot around that point. 30-minute meetings could, and did, go badly. If you had had a week where you just didn't work that hard, or you did work hard and struggled past some obstacles, or you worked really hard and just didn't have a clear result to show for it, you were in trouble. We probably found you at your desk with a huge coffee wearing the clothes you had on the day before. And even if you worked hard and successfully and had a stack full of graphs (always a good sign), you were still anxious. Friday was a huge and repetitive source of stress, and before leaving for Australia, I never would have forgotten a Friday.
So I hope that Matt is as happy at Google as I was in Australia, as I am now at MIT, and as I hope to be next year at UB. I hope he's forgotten his own Fridays, whatever they were. Because when Matt is in good spirits I remember him being a lot of fun to work with. (That's true for most people.) And I do consider myself amused at how life turns out: Matt, who disparaged me for being a hacker, now returns to Google to hack. And as Matt's doubts about academia were mounting, mine were receding. Now he's headed out, while I'm headed in.
In the wake of Matt's departure I think it's more important than ever for all of us to reflect on what we're doing and why. There are so many opportunities for smart computer scientists, each providing a different mix of tasks and responsibilities. It just doesn't make sense to do anything that isn't a good fit, especially given that while you may do it well, you will almost certainly do it miserably, and you might make life miserable for others around you. And so my next post will discuss my own reasons for pursuing academic research.
