I said this opinion exactly once, and it had a lifetime well beyond what it probably deserved. In any case, I feel the need to defend myself.

In popular culture, researchers lead spiky lives. With every breakthrough discovery comes a mathematician or physicist with a bite-sized anecdote illustrating how odd they are. Wiles proved Fermat’s last theorem in his attic in total secrecy, Djikstra wrote acrimonious essays by fountain pen, Feynman was a lovable misogynist. I don’t dispute that scientists are deeply strange people—the famous ones as well as the ones currently sitting in their MIT offices. But reading all of these stories, you could come away believing that an unlivable lifestyle is excusable for the sake of producing great research. Or that serious personality flaws and brilliancy in your career go hand in hand.

I don’t know if the empirical claims are true. Maybe Nobel prizes go to the worst people. But I disagree with the stronger, prescriptive claim that you should overlook your personal flaws. It’s always worth working on your personal flaws, because personal flaws leak into every facet of your life, including work. This is so obvious a point I’m not sure it even merits this many words. Some examples, drawn from experience with the people in my field (sorry!):

• If you have trouble communicating, you might also have trouble communicating your ideas in your papers and disseminating your work through talks.
• If you have trouble understanding people, you might end up producing work that nobody cares about, since it does not meaningfully contribute to the literature.
• If you have a big ego, you might develop an unsavory reputation as a bad or difficult collaborator, limiting your research abilities.
• If you have low self-esteem, you might do research which is “too safe”, or otherwise not trust the instincts which would lead you to making progress.
• If you are depressed, you might have your research tastes dulled, leading you to stay on dead-end projects and ignore promising opportunities.

I know great researchers with each one of these issues. I certainly have some of my own. And, speaking from personal experience, dealing with my personal issues makes my research, on the whole, better. That’s why this idea has stuck with me: as someone who has trouble not working, it’s a reason to do other things. We live in such a specialized world, where everywhere you look is a rabbit hole full of knowledge, mysteries, and things to do. It’s more important than ever to be well-rounded, to have the best perspective we can have on our little corner of the world. I wouldn’t want to have a whole career in one little subfield and only realize after that I was better-served doing something else, whether that be writing fanfiction or working on a topic whose the world expert is in the same building I work in.

By the way, I do recognize the obvious objection. The best way to get better at research is not to fix your personal flaws: it’s obviously to do more research. This is definitely true for junior researchers, where the main barrier to their success is not some consequence of their personality but just their inexperience and lack of technical skill. Even for me, if I spent more time doing research, I bet I would produce more, better research. However, (and now we are getting into the well-worn grooves of grad student self-help), if the goal is to sustainably do good research over the course of a career, then the question is not “how can I do great research?” but “how can I be the kind of person who produces great research?” The way to do this is not to be a tortured obsessive (if you are reading this, you are not one of those anyways), but to be someone who finds meaning in their work and also in their life.

I also recognize that this argument hinges on a subtle implication of my personal research taste. There are famous and brilliant “tortured obsessive” types, and there are famous and brilliant “normal people” types. What I’m saying is that you want to model your life after the latter. The best reason I can give for this is: I know many successful researchers. Some of them have good lives outside of work, some of them don’t. I can’t really distinguish a difference in quality between the two groups’ output. Perhaps one or two will become extremely successful, enough to achieve some degree of fame. Trying to hit this lottery is not worth doing; you might as well consider it random. Then, why not, on the whole, choose to be the kind of successful researcher who has a life? Do you really, in your heart of hearts, want to live the life that Erdős lived? It just feels like a no-brainer to me.