Vivek Haldar

Advice to (prospective) grad students

In my final year of college, when I was applying to graduate schools, I spent a lot of time reading “advice to graduate students” pages. Some of them helped shape my thinking about the decision of whether to do a PhD, and I feel indebted to them. (And by “grad students”, I really mean “PhD students.”) Now that I’ve been through that experience, here is my humble contribution to the genre.

The category representative of the genre is Ronald Azuma’s timeless classic “So Long, and Thanks for the PhD.” Much of what I wanted to say has already been said there, and although the piece is more than a decade old, it holds up remarkably well. What I want to do here is provide some personal perspective.

My credentials are that I’ve actually done a PhD in computer science, and since then spent as much time out in the real world as in grad school, so I’ve seen both sides of the fence.

Do you really want to get a PhD?

Before climbing the ladder, make sure it’s up against the right wall. Should you get a PhD?

The answer to that depends on two questions? Do you need one? Do you want one?

You need a PhD if you want to go into academia. If you find the process of getting a PhD unenjoyable, that is a strong sign that you are not cut out to be in academia. If a PhD makes you hunger for more, then you are.

The numbers are stacked against you though. PhD graduates in computer science outnumbered tenure-track professor positions in CS by a factor of 10. The next best option for those who want to be in academia is to become a postdoc,which is why postdoc numbers surged, but let’s face it, that is a crappy second option.

And let’s not even talk about the financial oppurtunity cost. It is nearly 100% certain that someone who started in industry on the day you joined grad school will overtake you in salary by the time you graduate and join the “real world.”

There used to be certain positions in industry, especially in R&D, that absolutely needed people with PhDs. That has been steadily eroding over the years, and while a PhD is certainly a plus, its necessity has been watered down significantly. Even if you do not formally have a PhD but can make an impact and prove yourself in an industrial research setting, you can perform the exact same job as someone with a PhD.

Why am I being so discouraging here? Because I have seen my share of people get into grad school for the wrong reasons and go on to waste their time at best or get stuck in an unhappy path for life at worst. If you realize this is not for you, cut out after your Masters.

What it boils down to is that this is one of the most intense questions of self-knowledge you will ever face. The answer is simple: you should do a PhD if you really want to. Look into yourself to figure out if you really want to. Everything I said above was to list out things that should not factor into your decision.

This is what I like to call a clamped decision. If you are not fully convinced that the answer is “yes”, then the answer is “no.”

So why should you do one? I’ll get back to that after considering the more practical question of choosing where to go.

Choosing where to go

Assuming you are brilliant enough to have scored multiple admission offers, choosing where to go is a big decision.

What should your criteria be? If you are like most fourth-year college students, you will only look at one thing–the rankings of the schools. This is a big mistake. (With one important exception. See below.) The choice is much more nuanced than that. There are actually three big factors at play here, and you have to decide which ones you value more, and what balance to strike among them. They are:

  • Your happiness and quality of life during the 5 to 6 years you will be in graduate school
  • The quality of the learning and research at the school
  • Your future prospects after graduation

Your prospects after graduation have have nothing to do with your school if you are going into industry. Industry (at least, the tech/software industry) does not care about your alma mater. They care about whether you can program. They do not care about your credentials. They care about whether you do well at their interview.

There is an important exception. If you want to go into academia after your PhD, then school ranking matters, because whether they admit it or not, in practice a school will only hire new professors who have graduated from schools close to their own rank. This works the other way round too. Candidates will only want to join schools with ranks close to the one where they graduated from.

Which leaves us with the first two criteria.

The one choice that affects both of these is who your PhD advisor is. That should make up 80% of how you choose where to go. Repeat after me: PhD advisor. PhD advisor.

Are they someone that will nurture you, and give you room to grow and explore, while not breathing down your neck, and having the wisdom to gently nudge you back on track when you will inevitably be lost in the dark jungle of the trough of your PhD? Are they someone with a track record of successfully funding their research, so that you and your project are not suddenly left high and dry without money in the middle of your PhD?

How is one to find out these things? You can look at their webpage, but take everything there with a grain of salt, just as you would someone’s resume. A better way is to email their current grad students. You would be amazed at how helpful they might be. A large group of grad students is also a good sign of being funded.

I have seen first-hand what a stressed out and crappy life some of my fellow grad students had due to their advisors insisting that they put in a certain number of hours at the lab everyday, or not letting them graduate even though they had done more than enough work and published more than enough papers to justify getting a PhD already.

Graduate school is a long haul, and almost your entire happiness and quality of life depends on your PhD advisor. I cannot stress this enough. This is the only choice you should be stressing out about. All other variables are in the noise.

Coming to secondary factors, the next most important one is location. I wish someone had impressed this on me. I chose California, but that was luck, not wisdom.

You will have a much richer and more stimulating life if you are in a densely populated area as opposed to a small town with a ninety minute drive to the nearest big city. If you are in CS and want to go into industry, try to be on the coasts, preferably the West one. If you come from a place that does not have long snowy winters, beware of places that do. The “oh look the snow is so beautiful!” feeling goes away very quickly after a few months of shoveling snow.

Why do a PhD?

Let’s come back to the question of why you would want to do a PhD.

Like I said above, this is a question of self knowledge. The circular answer is that you should do it if you are the type of person who would enjoy it.

Let’s take another excursion and ask: what is the value of a PhD? Which begs the question: value to whom?

The value to employers is that it indicates deep specialty in a narrow field. Or so one might think. Does that mean that when you dissertation topic is no longer fashionable and becomes obsolete you have no value? No. So it must be more than that.

Another value to employers is the indirect signal a PhD sends. It means you took an a substantial intellectual task and stuck with it for a long time to achieve some sort of closure. That is valuable. But even more valuable than that is the fact that you faced an ambiguous ill-specified problem and charted it through an uncertain path. That is the central value of a PhD: to be able to comfortably work with ambiguity and uncertainty. Note that this implies the serial-learner “learning how to learn” capability usually ascribed to PhDs.

I’m certainly not claiming that doing a PhD is the only way to get the capability of comfortably working with ambiguity and uncertainty. There are certainly other activities and experiences in life that can equip you for that. But a PhD is structured to do that.

But I’ve only spoken of value to someone else. What about the value to you, as an individual? That would include all I’ve said above, but what else?

Before I come to that, I want to take a slight detour to talk about boredom, solitude and entertainment.

William Deresiewicz on solitude and boredom:

Boredom is not a necessary consequence of having nothing to do, it is only the negative experience of that state. Television, by obviating the need to learn how to make use of one’s lack of occupation, precludes one from ever discovering how to enjoy it. In fact, it renders that condition fearsome, its prospect intolerable. You are terrified of being bored — so you turn on the television.

I speak from experience. I grew up in the 60s and 70s, the age of television. I was trained to be bored; boredom was cultivated within me like a precious crop. (It has been said that consumer society wants to condition us to feel bored, since boredom creates a market for stimulation.) It took me years to discover — and my nervous system will never fully adjust to this idea; I still have to fight against boredom, am permanently damaged in this respect — that having nothing to do doesn’t have to be a bad thing. The alternative to boredom is what Whitman called idleness: a passive receptivity to the world.

So it is with the current generation’s experience of being alone. That is precisely the recognition implicit in the idea of solitude, which is to loneliness what idleness is to boredom. Loneliness is not the absence of company, it is grief over that absence. The lost sheep is lonely; the shepherd is not lonely. But the Internet is as powerful a machine for the production of loneliness as television is for the manufacture of boredom. If six hours of television a day creates the aptitude for boredom, the inability to sit still, a hundred text messages a day creates the aptitude for loneliness, the inability to be by yourself.

Hal Abelson, co-author of SICP, said of entertainment and education (and this is entirely from my memory of a talk by him I attended six years ago, and is paraphrased, so is almost certainly warped, which is why I’m not even going to put it in quotes): We see all around us that the marginal cost of goods and services eventually approaches zero, but one thing we all want and pay for is entertainment. And that’s the value of education. Ultimately, the more educated and creative you are, the easier it will be for you to entertain yourself and never be bored.

And that finally brings me full circle. The value of a PhD to you is that you will never be bored, and that you will be comfortable with solitude and cherish it. This is partly why I keep referring to universities (especially grad school) as the closest things to modern-day secular monasteries.