Skip to content

Advice for the budding theorist

November 3, 2015

As a Ph.D student, I searched for advice on succeeding in grad school, and often thought that eventually I would repay the favor when I became a professor. I never got around to doing that, but now decided to give it a shot. So, here goes:

Tip #1: Don’t get advice from the Internet.

Grad school, and more generally research life, can differ so widely based on the field, the university, and the personalities involved, that what is a sound strategy for one person can be a terrible idea for another. It’s always better to get advice from a wise and experienced person who knows and cares about you.

Given this advice, I wouldn’t blame you if you stopped reading this blog post at this point. In fact, you probably should. But I will offer some more tips just in case, though keep in mind that they may not be applicable to your situation.

Tip #2: Remember you have other options.

Much of the advice you see online (and maybe even get in person) will include a long list of do’s and don’ts and hoops to jump through on the way to the final goal of becoming a tenured professor. Now, from my personal experience, being a tenured professor is pretty nice, but if you are even considering a career in theoretical computer science, chances are you have the skills and talent for many alternative career paths that have a number of significant advantages over the academic life. The main benefit of the latter is that you get to set your own goals and not jump through other people’s hoops. Don’t lose track of that.

Tip #3: Research is hard. 

Research is very hard, or at least it’s very hard for me. Doing research is often not so much about solving a fixed problem given to us, but about setting out goals and revising them as we learn more. As we do so, the pendulum often swings between the states where the current problem we’re trying to solve is trivial, impossible, already known, or not well formed. Most of the time though is spent staring at a brick wall, trying to think of some way to bypass it.  (Metaphorically speaking, of course; what you’ll actually be staring at is an empty notepad and a cup of coffee.) An added difficulty is that all this hard work is often invisible, so you get mostly to observe other people’s successes and feel that you are the only one that is having such a hard time.

One of the common fallacies of beginning students is thinking that it’s all about innate talent. When you see people solve mathematical problems in seemingly no time, or hear stories about those geniuses that solved the main question of someone’s Ph.D thesis in a day, you may feel that success in research is out of your control. I have met many highly successful researchers over the years, and while some are insanely quick, others can take time to do even the simplest calculations. Success in research comes in a great many forms and people can have very different styles of work, personalities, types of questions they are interested in, and more. The only common denominator is that, as a rule, successful researchers are passionate about what they do and even those for which success seems to come “easily” actually work very hard at it.

Tip #4: You are your own boss.

As an undergraduate, you are used to getting feedback on how well you’re doing in the form of grades. Many corporations have periodic reviews and evaluations for employees. In the academic world, feedback is quite rare, and can come in varying forms. Part of this is because professors don’t like to have awkward conversations. But fundamentally it is because you are truly the measure of your own success. As a theoretical computer scientist, I don’t need my students to run experiments for me, write code, or even prove theorems. I view my role as truly an advisor to the student as they find out what they are passionate about and great at. The best ones eventually find their own “research compass” and set and pursue their own goals in a unique way.

Tip #5: Be a good boss.

Since you are managing yourself, you should try to do a good job at it. Here I cannot really give any general advice. Some students should be tougher on themselves, and push harder. Others are too hard on themselves, and anxiety and/or depression are quite common in graduate students (and beyond that). As I said, being a researcher, at any career stage, is a hard job. In theoretical studies, maintaining motivation is especially challenging, since one can spend days, months or more with no visible progress – there are no lines of codes being written, no data being collected. I would also caution against measuring progress by publications. My philosophy is that any day in which you learned something is a good day. Such learning can take many forms, including thinking about a problem, learning some new or not so new cool ideas by attending a seminar or talking to a colleague, reinventing old ideas, reading a paper, figuring out why some approach doesn’t work, and more. I’ve heard a talk by Sanjay Gupta in which he gave the following advice — “try to do every day one thing that scares you” – I think this applies to research as well.

Just try to spend a good amount of your time on the things that truly matter to you, and less time worrying about the job market, who got published where, or reading advice blogs from some random professors on the Internet.

19 Comments leave one →
  1. Ankur Moitra permalink
    November 3, 2015 2:31 pm

    Boaz, this is really terrific advice! I couldn’t agree more with what you wrote here

    • November 3, 2015 5:55 pm

      Thanks Ankur! I assume you agreed most with the part where I said people shouldn’t follow my advice 🙂

  2. November 3, 2015 4:29 pm

    all excellent advice especially the contradictory parts :p
    now after criticizing the internet, how about a followup that describes how useful it can be? actually “on the other hand” it fits in neatly to nearly all the goals you describe. but agreed, it is also a highly evolved way of wasting time, & have noticed that still, many profs tend to avoid it (arxiv notwithstanding), at least “social networking” aspects of it. so then my question is, how do we change that? 🙂

  3. PhD student permalink
    November 3, 2015 5:48 pm

    Were the successful researchers with varying speed you met also in TCS?

    • November 3, 2015 5:56 pm

      Of course. Research in TCS is a long term enterprise, requiring vision, taste, creativity, and tenacity. Speed in doing math on the fly is a useful skill to have, but neither necessary nor sufficient for success.

      • PhD student permalink
        November 3, 2015 6:20 pm

        This doesn’t make me happy or sad. But I think the problem is how to bootstrap this serious/great advice without using the Internet.

  4. Hoeteck Wee permalink
    November 3, 2015 6:14 pm

    Thanks, Boaz!

  5. November 3, 2015 7:16 pm

    Excellent advice!

  6. Alon Rosen permalink
    November 3, 2015 7:45 pm


  7. November 3, 2015 8:29 pm

    Good to hear!

  8. November 4, 2015 1:49 am

    Thanks to all!! Hope people found this useful!

  9. November 4, 2015 4:14 am

    Very nice, and were I to make a comparison with terry tao’s advice, I’d say that yours is easier to digest, believe, and truer to my experience.

    However, first-paragraph irony is a dead giveaway for a theorist.

  10. November 4, 2015 4:39 am

    This is really good advice Boaz. It is very different from the grad-student advice that I have come across till now.

  11. November 4, 2015 7:31 am

    Thanks a lot ! Very inspiring 🙂

  12. November 4, 2015 4:59 pm

    Very nice advice Boaz!

  13. Bob permalink
    November 6, 2015 10:34 pm

    The main thing I took from this is

    “You have to do things in your own way”

    Don’t spend too much time worrying about what others are doing, or how they’re doing it (or how many papers they’re getting published, big jobs they’re getting, etc etc) because ultimately that doesn’t help you specifically too much. Your best possible results are most likely to be achieved when you’re doing things in a way that makes a lot of sense to you – even if it’s not how any one else would do it.

    But still: You’re going to have to work hard. By this I simply mean consciously pursuing a somewhat disciplined and systematic approach. But also really taking time to figure yourself out, what are your personal obstacles and how you can do better. Any program pursued halfheartedly or haphazardly probably will not give satisfactory results. For many people this can be the harder part. It’s all very well being smart and creative and hopeful, but consistently executing and following up and improving yourself over years and years is a whole different story. It’s definitely something that I personally struggle with.

    And lastly there is no guarantee of “success”. You can do all this and still be very middling, despite considering yourself as talented as the “stars”. But it’s still better than desperately trying to imitate someone else and definitely a lot less misery.

    Anyway, in the worst case if you don’t get tenure you can always get a job at Google for the harsh price of making more money, working shorter hours and dealing with less politics (and yes, I believe this is reasonably true but obviously YMMV).

    So this is a loquacious and roundabout way of saying that, based on my experience, this post is totally correct.

  14. November 10, 2015 8:56 am

    Could you elaborate on the part about “The only common denominator is that, as a rule, successful researchers are passionate about what they do.”

    • November 12, 2015 1:30 pm

      e.g. could an unimpassioned but still interested curiosity coupled with a moderately professional work ethic make a successful researcher.
      Or, what if you think a certain field is interesting but you’re not passionate in a sort of exclusive-claim-to-most-interesting-thing-out-there sense. Would you nip such a theorist in the bud?

      • November 12, 2015 1:44 pm

        I think such a person could still make very positive contributions to theoretical computer science even if they are unlikely to be great breakthroughs. So I would definitely not “nip at the bud” such a theorist from my viewpoint.

        However, from the viewpoint of the person themselves, they might want to ask if there is an alternative career path they are more passionate about. There are going to be several hurdles along the way, both during the Ph.D and afterwards (especially if you go on the academic vs the industry route) and it may be harder to go through them if you are not passionate about TCS, especially when you have the skills to get a good high paying job without those hurdles.

        I am not at all saying that only the people who live 24/7 for TCS should pursue it, but at least from my experience it is a challenging (but also rewarding) career path and you should make sure it is the right one for you.

        Generally, as long as you are enjoying what you are doing, you don’t need to think too much about these issues and definitely not worry if you are passionate enough. Only if you stop enjoying it then you might want to take stock and see if it’s the right fit for you.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: