Being a theoretician is like being an artist

This winter break I attended two computer science conferences: Workshop in Network Economics (WINE) (http://www.dis.uniroma1.it/~wine09/) and Innovations in Computer Science (ICS) (http://itcs.tsinghua.edu.cn/ICS2010/). Beside meeting interesting people, learning assorted facts about game theory and mechanism design, and gleaning possible research directions, I took away this important observation: when I see the most eminent theoreticians talk about research, they seem like blissful children playing in a sandbox; they seem to enjoy asking and thinking about mathematical questions in and of itself, treating the process as a deserving end instead of a means. In many ways, good theoreticians are like artists: they obtain tremendous enjoyment and fulfillment from the act of creating their work, regardless of whether they reap other rewards. I think that this is the only "right" way to go about research, because the efforts and commitments needed for success are enormous, while the pay, the possible fame, the job prospects, and the chances of having one's work materialize into practical applications are slim.

Reflecting on why I do research, I realized that the primary reason I engaged in research was not that the research itself was so captivating, but that I wanted to churn out papers to boost my grad-school resume. The simple litmus test is that after my papers got accepted, I no longer cared about the research topic, except whether I can milk another paper out of it. Although I could appreciate the beauty in the underlying order I was uncovering, I was not drawn to the question by itself in the way that a painter is captivated by his/her painting or a musician is obsessed with his/her song. Hence, I don't think theoretical research is the right career for me. Instead of enjoying simply being in the sandbox and satisfying my imagination through it, I don't like sand that much; I build sand castles just to win competitions so that I can have money to buy other toys. I should figure out what toys I actually want.