Hacker News Reader

As a professor (well associate professor actually), I can second pretty much everything you have said.

One additional thing I want to mention is, that the job of the supervisor is not only to guide the student to the PhD (i.e. the mentor role), but also to ensure that the student actually has done sufficient work to earn the PhD (the role of a gatekeeper). Which can lead to the situation, that the student thinks they should be graduating now, but the supervisor is still asking for more work. Now there are some supervisors that are abusing the gatekeeper role. Furthermore even for supervisors that are not consciously abusing the role there is also a conflict of interest, because it is in the supervisors interest to keep successful students as long as possible. So this is difficult to manage.

Agreed.

I think of my PhD students as my primary "product" I am creating as a professor. Not to say I don't think that teaching is valuable, but I'm by far proudest of my PhD students who graduate and go on to have fantastic careers. That said, I'm proud, but also sad, and worried because once they become good they leave. I let them depart once they have met our agreed upon goals, which are pretty much the bare minimum for being competitive for research-oriented jobs in industry (not "just a coder," but designing and doing independent research).

>>> The first thing I did when I became a professor was write up an expectations document, and when students want to join my lab I walk them through it, and even then we do a trial run if they still want to.

A relative of mine got his PhD with a prof who was kind of famous for such a document, and all of his students continued that tradition in their own labs. Of course you know the drill. Part of it is to influence your students, but the other part is to influence yourself.

> Almost always by year 4+ most PhD students are fatigued and starting to get very weary

Currently experiencing this weariness and my output isn't great. Covid restrictions + WFH are making it worse I feel but ultimately it's my responsibility to do well. Any tips for not looking like a hopeless case to your supervisor at this stage?

The struggle is real... my most productive deep work has always happened while standing next to my advisor in his office, staring at his chalkboard and working through research problems together. Can’t do that over Zoom...

Same actually. I haven't had those type of conversations, and moments of inspiration/motivation in a year now, with everything.

Be structured and have clear goals.

After we went to remote work, my students have been making weekly presentations to me, which serves to update me on their work and helps them become better communicators of their research.

They recap what I suggested, they give a status update on what I asked them to do and whatever else they did, they compare the results of their experiments to past experiments (baselines), and they propose next steps and I make suggestions with regard to what I think makes sense.

In my field, machine learning, we often stitch multiple papers together, so each project is aimed at being a paper in the dissertation. For my lab, we come up with the idealized main contributions for a paper, and then work toward those goals. We may not meet them, but it provides structure and their informal presentations to me facilitate organized discussion and analysis of results.

> I eventually decided I felt that way because he didn't make his expectations for me finishing and tailoring those expectations to my career goals clear up front, and very few advisors do so.

I come at this from a different field (the biomedical sciences), wherein it is rare to find clearly-articulated career goals from the PhD students, and even harder to find supervisors who care. This problem is made worse when supervisors often rely on their students' naivete to keep their labs staffed.

The students can be forgiven for this naivete: they often start their degrees with the assumption that, if they work hard, a tenure-track or similar position will be available to them at the other end. Of course, reality is that they may fail to achieve that goal no matter how hard they work, simply because they have the wrong project/wrong mentorship/poor funding/etc. The supervisors, on the other hand, are often quite happy for their students to remain naive, as the incentives of students and supervisors are often misaligned.

The solution is probably twofold:

First, it should be essential for prospective students to come in with a very clear idea of their career goals (e.g. "I want a tenure-track faculty position in a leading research university" or "I want a dedicated teaching position at the community college level").

Second, said prospective students should be given clear, objective advice, along the lines of your expectations document, on what it takes to achieve those goals. This advice should help guide students in choosing a lab that will help them achieve these career goals, and ideally should come from a neutral third-party. One can't expect supervisors to be unbiased in this regard.

Is 1-2 hours per week really that much? I suppose they also spend additional time reading what you've written, along with revisions, etc?

They often advise several other students, and still have their faculty and research duties.

Also they are your advisors, not teachers. At PhD level, you are supposed to make a meaningful contribution to the discipline on your own.

it's good if my advisor give me 30m every 6 months :)

(sigh...)

I did my PhD in the UK and the communication structure with my supervisors was:

1st year: we had a 1 hour meeting every week

2nd year we had a 1 hour meeting every 2 weeks

3rd year (1t hslf) we had a 1 hour meeting every 2 weeks

3d year 2nd half onwards we had a 1 hour meeting if I emailed them and asked them for it

If I needed more time I could always knock on their door. And oftentimes we went to have lunch for informal meetings.

Now, i always felt my supervisors were absolutely amazing. But i quickly understood that to complete my PhD, i needed to scratch with my own hands and that my sups were there only to guide me.

All in all my experience was great. Nevertheless after finishing and doing 4 years as a postdoctoral i decided the "publish or perish" mentality was not for me, and moved to industry.

That's certainly true, but it isn't every week a student has something for me to read.

What professors at research universities do:

1. Mentor PhD students (and often also MS and BS students)

2. Write papers and edit student papers

3. Write grant proposals

4. Deal with a lot of administrative issues

5. University service committees - some are on multiple some are on none (admissions, COVID response, undergraduate program, graduate program, etc.)

6. Review papers from others

7. Teach one or more courses

8. Develop new courses (takes 1-2 days to make a good lecture)

9. Update lecture materials (I spend probably 2 hours per lecture for courses I've already taught)

10. Meetings

11. Conference/Workshop organization

12. MANY miscellaneous duties

13. Budget management

14. Prepare talks

15. Serving on MS/PhD proposal/defense committees for students not in your lab

16. Staying on top of the scientific literature

17. Making homework assignments (some can delegate to TAs)

18. Grading (some can delegate)

19. Write letters of recommendation (I write around 10-15 per year)

Of course you can get this load down by learning to say "No" in some cases. But a professor during their first couple years when they start the job can easily expect to put in 60-80 hours per week. It gets a lot easier after that.

Then of course most of your list (besides 5) is also what PhD students do, at least in the UK.

That’s more than I ever got. One day I was waiting outside my supervisors office to ask a question and he had left the country for a month or so. I didn’t even know.

An engineering manager having 5-6 direct reports would spend the same amount of time in a company. I guess PhD students would be mentored by postdocs, the same way juniors are mentored by senior developers in the industry.

Those direct reports aren't (nominally) trainees though. Part of the "deal" is that you're being underpaid because you're being trained.

Mentoring by more senior folks in the lab certainly helps everyone (https://www.pnas.org/content/116/42/20910), but it's barely measured or rewarded for PIs and not at all for postdocs. Changing this would be an easy fix though--ask for a reference from a junior colleague when hiring.

More generally, there's a huge disconnect and sudden between the skills needed to get a faculty job (i.e., publishing a high profile papers as an individual contributor) and the skills needed to do well at it (e.g., management). Everyone is all antsy to shorten postdocs, but I'd actually be in favor of making the career path more gradual with so people get some mentored management experience.

When I started my PhD my supervisor said to me "Welcome to the faculty, see you in 3 years when you submit". They kept that promise.

Working in industry, I found this pretty surprising. I'd expect to have around that much facetime with my manager, but also colleagues who know other, completely different things very well. That, and in industry, depending on the company and manager, the manager's role might be enabling your success and not being an expert on what you're doing.

Any chance you'd be willing to share some version of said expectations document?

It is basically the requirements for getting a job as a scientist in my field (machine learning) as far as graduating goes:

1) 1 journal paper as the first author, so that they learn how that process (often is a review or perspective paper on their field, which goes in their dissertation)

2) 2 first author papers in top-tier machine learning venues (goes in dissertation)

3) 1 first author paper in a top-tier or second tier venue (goes in dissertation)

4) 1 collaborative paper with another PhD student (have to learn how to collaborate)

By the end, they have at a minimum around 4 first author papers and 1 additional paper. They can then turn these into their dissertation and these are signals to employers that they are competent scientists. If a student tells me they want to be a faculty member, we increase the numbers a bit (need at least 10 first-author papers to be competitive).

I also require that they be organized when we have our one-on-ones in terms of their experimental output.

That's pretty much it. It may sound like a lot, but I try to put training wheels on for the first couple projects and make them as tractable as possible because I fleshed out the project and found some low-hanging fruit, before I start trying to make them be more independent and drive the process.

That's an insane workload to put out during a PhD. 5 Papers including 2 top-tier ones in 3-4 years? Is ML really full of such low-hanging fruit where that is possible?

In my stints around Denmark and Japan (Computational Chemistry) you basically need ~1 top-tier or ~3 mediocre papers during your 3-year PhD.

Well, five smallest publishable units... optimizing for the wrong incentives that be.

Norms vary per academic discipline and niche. My requirements are pretty much the standard for my field if you want a job doing research at FAIR, DeepMind, Adobe Research, Google Research, OpenAI, etc. Without meeting those goals students will not be able to get research oriented jobs at top places.

In the US, PhD students in Computer Science and similar fields usually require 5-7 years to finish. They typically do not have MS degrees, unlike in Europe, where the expected graduation time is 3-4 years and they must have MS degrees.

You would be surprised how often there is a lot of low-hanging fruit. I'm good at asking "easy" questions that nobody knows the answer to, so a student just has to put in the work to get the answer. Many advisors don't give as much initial hand holding as I do.

So far my PhD students have taken 3-6 years to finish. Those that have graduated so far have finished with 3 papers (3 years), 9 papers (3 years), and 12 papers (5.5 years). The senior ones still in my lab (year 4 or 5) are on track to have around 3 first author papers in top-venues and about 9 papers total each. Their first author papers are in CVPR, ICCV, ACL, ECCV, BMVC, NAACL, ICLR, AAAI, etc.

Thanks! That makes sense. My requirements for my students aren't quite at the same level. (And I arguably didn't meet those myself for my own PhD.) 10 first author papers is seems a little crazy to me. As a relatively new faculty member, I haven't met that bar myself yet. But of course as you say in another comment below, it differs by field.

Thank you for this perspective. My supervisor is in many ways a caring and kind person, and I appreciate his job (and yours!) is difficult and has lots of pressures, many of which are invisible to trainees. But at the same time, I can't imagine being in his role and treating trainees as he's treated me or others. I hope to keep that feeling with me as I progress in my career so that I don't subject people over whom I have power to the same treatment that I've experienced through grad school.

I can relate to the "fatigue" issue. I am an ABD from a big world class university, and plan on never going back after suspending my studies. During the course of my studies I became ill and even though some people on my committee expressed support, my chair did not really know how to mentor and encourage me while I was going through chemotherapy, trying to work full time (to keep medical coverage) and do the work necessary to complete the dissertation. My experiences working in other universities are even worse. Faculty who exploit students or punish them for perceived slights or lack of dedication are one of the key reasons for the mess that we are in. It's dreadful overall and would not suggest doing any doctoral studies to anyone. Your viewpoint about your students as apprentices is very encouraging but this is an attitude that is not conducive to most university settings that I have been involved with. Its about the money plain and simple.