Advice for thinking about whether to pursue a PhD

Question

I've included lots of context because I thought it may be useful. If it's too much to read, I've marked the important parts of the question.

I graduated from college last May with a degree in Computer Science and Math and just recently turned 23. I'm in San Francisco now working in tech. The money is nice but I've found it's not quite the life I want. Over the past 7 months, I've developed a near insatiable taste for curiosity that I had never seriously indulged in the past (I was naively focused on getting a lucrative job that might be a stepping stone to maximizing my wealth and "success", not learning for curiosity's sake). I've learned some Statistical Mechanics, Bayesian Statistics, Topology, Logic, Information Theory, Social Psychology, and the list goes on.

I think this part of me came out partly because I feel disillusioned with tech and partly because I feel free from the college->job market rat race and can pursue what I am genuinely interested in.

Important 1 Start

I'm considering switching to a PhD track for 2 reasons:

1) So I can better shape my life by what I'm passionate about. It's frustrating to work knowing that I would rather be learning something else.

2) In learning more I've discovered how little I really know. This smallness not only humbles me but also motivates me to learn more and use what I learn to reshape my worldview. I feel particularly lonely with this feeling and know very few people in my social circle who relate. While I'm aware there may be other social circles that I could join with more people who feel this way, I expect students pursuing PhDs and professors share this feeling as well as they've surely learned more than me.

I am considering a PhD track because it seems like it may satisfy both of the issues I described above.

Important 1 End

I'm setting my potential start age of a PhD to 25 because I want to take the time I have (especially with the current pandemic) to think extremely carefully about this decision before I devote at least 4-7 years of my life. I also want to be mindful of other options that may satisfy both (1) and (2) above. I'm finding over the past 7 months I've begun to understand my emotions and personality much better than I ever thought possible. I see this new understanding as overwhelmingly positive since it helps me make the best decisions I can for myself. I'm willing to take the time now to think about this decision if it will lead to a more fulfilling life 20 years down the line.

The final question I have is how someone decides where it is they want to focus their research. I find that I get easily distracted by new information as I learn. I may have an interesting idea that leads me to explore one path down some material. As I learn more, the material sparks a new question or idea that leads me down a new path. While this feels like a great strategy to self-learn with a constructive and natural progression, it doesn't seem like an effective way to research. How does one decide they're ready to study a very specific question that may close doors to other material and questions in related fields they find interesting?

In my case, I spent a lot of my undergrad time learning about machine learning and working with a startup that explored some research questions related to its business. I find machine learning interesting because of how natural the algorithms can feel to think about. The ideas feel intuitive while my math education helps me wade through the more nuanced derivations and implications. On top of that, both the advancement of computers and my familiarity with them helps me solve seemingly difficult problems. But at the same time, I feel more drawn to questions on causality, truth, logic, observable effects at large scales, and efficient problem-solving. I have found physics fascinating since I was a kid but never started learning anything beyond basic relativity and quantum mechanics until recently. It seems like that opportunity is gone; to pursue physics long-term I would have needed to study it in undergrad. I enjoy math as well, but more so as it comes up as an abstraction to a concrete problem I'm thinking about. I would not like to spend my "passion-time" solving unsolved problems or proving various conjectures simply out of my enjoyment for math. I enjoy having a tether back to our world and the means to produce and use any result I find.

Important 2 Start

In short, I'm not looking for an answer to my life question (I must answer that myself), but rather answers to these 3 gaps in my reasoning:

1) Is it more common at a PhD program to find others who share a humility and sense of wonder at the immense amount of knowledge both discovered and undiscovered?

2) How do PhD candidates decide what question(s) they are interested in pursuing, potentially closing doors on other related fields and questions they find interesting?

3) Am I expected to be set on a question to pursue before I join a program? How flexible is this if I learn something that makes my question seem unsolvable?

Important 2 End

That being said, if you feel that you have life advice to give, I will gladly welcome it. More information will help me make a well-informed decision.

Answer 1

Is it more common at a PhD program to find others who share a humility and sense of wonder at the immense amount of knowledge both discovered and undiscovered?

Yes, but it's also more common to find people completely jaded and disinterested in their work due to poor research progression, troubles with supervisors or colleagues and deep financial worries. Research itself can be the wonderful, carefree pursuit of new knowledge that you describe, but getting a PhD is often the opposite. Go in with your eyes open: the reality is, it's very hard work and takes a long time, and there will be many days/weeks/months where your feel like you're banging your head against a brick wall and like you're not learning anything at all.

How do PhD candidates decide what question(s) they are interested in pursuing, potentially closing doors on other related fields and questions they find interesting?

I did a lot of reading. There are (and always will be) a ton of things in physics I find interesting, but in terms of finding a topic for my PhD it was pretty much a case of "what's the coolest thing out of all these interesting things". And in fact, I've grown to love my research area more the more I've worked on it and the more depth of knowledge and understanding of the field I've gained. Plus, if you stay in research after your PhD, there's plenty of time to work on all the ideas you've had on the back burner.

Am I expected to be set on a question to pursue before I join a program? How flexible is this if I learn something that makes my question seem unsolvable?

No, although having an understanding of the field and the big open questions helps. A null result is still a result. You can change direction at any time.

Good luck.

Answer 2

I would keep very much in mind that a PhD is a professional degree. If you are not certain you want a research career, then it's likely not the right correction to go in. Luckily in computer science a research career doesn't strictly mean an academic career, so there are some options. That said, going into a PhD program driven wholly by a sense of wonder at the landscape of human knowledge, with insufficient attention to concrete career goals, is a common failure mode. You need to work into your plans the possibility that all this stuff might seem a lot less wondrous once you've spent five (or six, or seven...) years trying to compete with a lot of other really smart people to prove that you've got the best insights on it all. Even if not, you still have to get a job!

The smartest PhD plan I ever saw was an old colleague who worked as a coder for a long time, almost ten years, spending lots of nights at home getting familiar with the research landscape. After a while he knew exactly who he wanted to study with and went and did it, with great success. While as others have said, you're not expected to know what you want to work on on entering a(n American) PhD, it bears keeping in mind that the vast majority of PhD students do not end up in research careers at all. Those who do are often over-prepared. It's not necessarily a waste to do a PhD and then go into a pure teaching job or back to coding, but far more people end up on such a track after their PhD than intended to on the way in. Make sure you know what you're aiming for, and what it takes to get there.

Answer 3

Too few people around the world have the option to actually choose what they will do as their life's work. If you have the chance to do so, pick something that you feel passionate about and forget the rest.

Are you driven to do research in some field, say CS or math? If not, don't bother. It is often said that "You don't choose mathematics. Mathematics chooses you." That was my experience. I also really wanted to teach and be a productive faculty member.

But the goal is a ways off and there are many hurdles. Do you have the drive and the patience to push through them? Including disappointments?

From the basic sound of your question, I'd think that you would be a good candidate, but it is up to you. Do it if you really can't envision spending your life any other way.

You only get one shot at life. No do overs. If you aren't forced into something unpleasant by circumstances. Grab it by the ...

Answer 4

The other comments are great, but based on my experience at an R1 school in a CS department with a large international student population and strong connection to industry sponsored grad students:

1) Is it more common at a PhD program to find others who share a humility and sense of wonder at the immense amount of knowledge both discovered and undiscovered?

Not necessarily. Although PhD students on average may relate to this, I've never thought to describe grad students as such. People's motivation will differ. I know some international graduate students who want to stay in the United States and being in a PhD program helps ensure they can stay. Others may be doing it because where they work needs trained scientists/engineers and the company sponsors them.

2) How do PhD candidates decide what question(s) they are interested in pursuing, potentially closing doors on other related fields and questions they find interesting?

Unless you are reading papers for work or connected to people who are familiar with the literature, it is going to be difficult. One strategy is to look up the proceedings of a conference in a field you are interested in, for example CVPR or NuerIPS, and browse the papers that catch your eye. More realistically, and which is my case, my current research direction is fairly different from what I put on my grad school application and has largely been dictated by my advisor's interests, collaborators, and data available. In fact, most programs have classes or structures in place for PhD students to find a suitable advisor and research direction in their first 2-3 years in the program.

3) Am I expected to be set on a question to pursue before I join a program? How flexible is this if I learn something that makes my question seem unsolvable?

Mostly answered above. You aren't expected to be set on a question, though having an area of interest, such as machine learning, would be helpful for the admissions committee to pair you up with an initial advisor. If a problem appears to be intractable, get feedback from others or try to solve a constrained version of this problem. Your advisor should be able to help you navigate this.