Academia Deromanticized

I decided fairly recently to not pursue my dreams of a PhD. I felt surprisingly relieved (maybe a bit guilty) as I turned down a PhD offer in ML & comp neuro from CMU for an overpriced MS degree. It was a tough decision, but here's my reflection.

Up until a few months ago, I thought I was certain of my career trajectory. As I romanticized the role of becoming a research scientist at DeepMind, the pressure to go through the obvious route of a PhD intensified. During my undergrad, research became my life and occupied 90% of my mind. I was in three research labs simultaneously and took on two additional research internships in the span of three years. All five were in different fields: ML, robotics, neuroscience, psychology, and statistics.

I should mention it was incredibly interesting to observe the differences in traits and the way people thought about things within different departments. For instance, in the robotics lab, which was part of the CS/engineering school, the researchers were more logical and sharp, but tended to be less abstract. The cognitive science researchers in the psych department were more creative and inventive, but less rigorous in their thought processes. Regardless of these trait differences, I realized people from both groups still shared a commonality: an intense commitment to their respective fields.

After this realization, I decided that I would probably be a mediocre academic/research scientist given my dilettante instincts. I think the most successful academics have a vertically oriented mindset, which helps them dedicate their cognitive resources into deep and focused, but more narrow slot. I, on the other hand, was inherently a dabbler. I liked learning a bit of everything and never really felt specialized in anything. Part of it was also just that I got to see a snippet of the dark side of academia. Toxic advisors, publish-or-perish pressures, nocturnal working hours, shitty reviewers, and depressed PhDs.

What I personally don't love about academia is ending up spending way too much time on niche problems that (unfortunately in hindsight), were neither interesting to my original curiosities nor productive to the bigger picture. Results of research are oftentimes disentangled from the real world, solving simulated toy problems in labs constrained by a precise set of controlled parameters. And only a small percentage of groundbreaking works truly make an impact. More often than not, people go in asking interesting questions and end up being squeezed into working on slightly less interesting problems. The neuroscientist who wants to figure out how the brain works ends up spending 6 years studying the impact of UV rays on retinal ganglion cells. Same with the chemist who wants to understand the nature of matter ends up focusing on the geometric properties of pressurized gallium arsenide. The topics of most of your work becomes so niche that there are maybe 2-10 other people that can truly understand what you do (and this may or may not include your own PI lol).

I mean, I guess this is expected because obviously science is hard. Researchers must hone in on a microcosm of their field, otherwise they end up accomplishing – or rather publishing, nothing. For one, research is an incredibly slow and lengthy process that requires much patience and grit.

Don’t get me wrong, I absolutely still have love for research. I love the struggle of learning something difficult. I love the feeling of finally understanding something difficult. I love grueling through dense, boring papers only to appreciate the beauty of it after the 10th read. I love being able to mentally simplify an abstract topic that can, after some time, jibe with my intuition. I love the rush that comes with a new idea. I love critically thinking through some reasons why my initial idea failed. I love the brainstorming sessions, the whiteboard discussions, the random wake-up-in-the-middle-of-the-night-to-jot-down-a-bad-idea moments, and finally, it's absolutely fun to engage in jargony, technical discussions with a peer, that to an outsider, sounds like a whole new language of nerd gibberish. But the best part is the people you get to surround yourself with. These are some of the smartest, most curious, and determined people you can find on campus.

So here was the dilemma. Research felt too slow and unproductive to real world problems for me. Yet, I loved the challenge, the people, and the environment. I had to figure out how I could integrate the stuff I liked while marginalizing out the variables I didn’t like. That left me with the option of doing an MS degree. I would still be able to learn new things, work on difficult problems, brainstorm and discuss ideas, befriend academics, and reap all the promises of research, but without the perils of rabbit-holing, publishing stress, and a 6 year commitment. I figured the MS and the school would also help out with my long term eventual goals of a startup. Of course there’s that part where I go into student debt, but for now I’m choosing to remain fairly confident (maybe naive) that I’ll be able to pay it off soon after.