Surviving modern Ph.D. program

Many master students I know are applying for doctoral programs now, which makes me want to write down some of my thoughts. I won’t talk about comonly discussed topics. Instead, I want to talk about training in an industry-relevant field.

I sometimes joke that I have a cursed taste for advisors that are more likely to leave academia. For example, the PI of my undergraduate research left to join VC, and my first graduate research PI left to be a CEO. I used to think that I was just unlucky and this is a rare situation, but my perspectives have changed. First of all, advisor leaving is actually fairly common. Maybe they didn’t get tenure, maybe they relocated to a different university for personal reasons, or maybe they passed away. This issue has probably gone worse recently with the popularity on AI-related topics. More advisors are proactively choosing to spend a fraction of their time or even go all in for startups, rather than leave because they have to or leave for better research environment.

Even before this whole AI era, senior students had been telling me to make sure there are at least 3 PIs I really want to work with when deciding which school to do doctoral study. Although I think back then this advice was more about mitigating the risk of choosing an advisor that doesn’t really suit you. This tip might sounds cliche, but I think the key point to emphasize is “PIs you really want to work with”. The real tough decision is when you only have a few so-so offers that don’t meet this criteria, are you going to turn a blind eye and hope for the best? I am not necessarily suggesting that you should give up on grad school in that situation. I followed the tip, made my choice on the school, was unfortunate in a situation that called for this tip, but I took new possibilities instead of the 2 PIs I originally had in mind, and eventually turned out okay – at least on paper. Back then, senior students have also told me that 5+ years of Ph.D. is a very very very long time, that something unexpected, something unpleasant, something bad, is statistically almost bound to occur at least once. You can gather all possible tips from friends and mentors and online posts, and they might be still not enough as a recipe for a smooth Ph.D. experience. Are you ready (not only in terms of backup planning, but also mental maturity) for this level of uncertainty?

Even if an advisor doesn’t leave academia, that doesn’t mean you will get the kind of training and mentoring you wanted. I am noticing that students are increasingly expected to come in experienced, even skilled, and ready to start churning out results. I don’t know if this is a matter of some PIs can be very busy and hands off, hence students have to be independent to survive grad school. Or is this a matter of increasing competitions, both at the applicant level and at the PI level? If PI is facing more competition, they might be more risk averse. So then, the question is, when and how should students get trained? Should I accept that, the situation has become filtering instead of training and I am either a winner or a loser? Well, that’s not a helpful mindset, and a loser still gotta live. One should recognize that we each have our own growth trajectories. Perhaps some are more gifted, some are more lucky, some are more mature, so at a given time point, some appear more ahead and made the cut. It sounds very out of touch to say that the important thing is not about getting into a prestigious program or landing an amazing job offer. But in the long run, I do agree that the thing to optimize for is, where can I get the real learning experience? Is it in the form of a post-bacc, master program, post-doc? Is it in the form of a stable job that allows me to do self-learning in spare time? Is it through industry internship? Is it through mentors you meet at external occasions like conferences or discussion on open-source projects?

In the end, I don’t have an answer to anything. There are many systematic issues, and I don’t know to what extent an individual’s action helps. I was raised academically by meals from a hundred household (吃百家饭长大的). Perhaps for many other people, this is the modern way of getting training.