Should you do a PhD?
Note about my background: I did my PhD in Physics, in Grenoble, France. My impression is that the PhD system is quite similar for the rest of Europe and Canada. The US is significantly different, and I don’t know about the rest of the world.
Thanks to: Nandi Shoots, Gavin Leech, Daniel Filan and Callum Hinchcliffe for helpful comments and proofreading.
Sometimes someone asks me something along the lines of “Is it worth trading X years of my life to get a PhD diploma?”
My reaction to this is:
If you put it like that, then obviously NO!
You have no idea what a PhD is, do you?
A PhD is an entry level research job – nothing more, nothing less. Having a PhD is a requirement for some jobs, but so is X years of software development experience, or basically any other skilled job. There is nothing magical about a PhD, and like any other job (or anything really), if you’re just in it for the credentials, then there is almost certainly something better you can do.
You should not do a PhD just so you can do something else later. Only do a PhD if this is something you would like to do, in itself.
If you find yourself confused about the question “Should I do a PhD?”, just replace it with “Should I take an entry level research job?”
Now, if you are still considering doing a PhD, then there are some things you should know about that are particular to this career path. This is because there are ways in which academia is not quite like the rest of the world.
I’ve met many physicists who look back at their PhD as the time where they were most free to actually do research, before their time was taken over by teaching duties and bureaucracy. If you are only using it as a stepping stone to an academic career, you might miss out on the best part of this career path. I recommend that you try to do the research you really want right from the start – although this might be different in different fields.
You probably only get to do a PhD once. There is no rule about getting more than one PhD but some universities do not accept applicants who already have one, even in a different subject. I also think it is hard to get a second chance if you are a PhD dropout, but I have a lot of uncertainty around this. Because of this you should take extra care to choose your PhD position before you start.
As a general rule, you can take as much time as you want before starting your PhD without many consequences to your career (although I’ve been told that this is not true in maths, and there may be other exceptions too). But once you’ve stepped on the academic career path (getting a PhD is the first step), it is very hard to come back if you ever leave. It is also hard to change subject.
Think about how much you care about having an academic career after your PhD. If you do want an academic career you have to at least somewhat optimise for prestige. This means getting into the right institution, and publishing as much as you can in the right journals. Also find out how competitive it is to have an academic career in your particular field, and have some backup plans in case you can’t get a post-doc.
Probably the most important thing to keep in mind: make sure to find a supervisor that you get along with!
Professors become professors because of their research skills (or sometimes random luck, being in the right place at the right time, etc.), not for their people skills. Some PhD supervisors are great and supportive – I got along great with my supervisor – but I’ve heard stories of other PhD students who were not so lucky.
From what I have heard, these are some problems you might have:
Your supervisor never has time for you. This seems super common.
Your supervisor wants to be the first author on all of your papers even though you did most of the work, or your supervisor wants to be co-author even if they did nothing to help.
Your supervisor decides what projects you should do. This may not be a problem if you fully agree, or just don’t feel ready to choose your own research. But it is worth being aware that the amount of freedom you get can be very different depending both on your supervisor and on your funding situation.
You will probably talk to your potential supervisor when interviewing for a PhD position. Don’t be afraid of bringing up things that are important to you, like the amount of freedom you want, or making sure that you will get the support you need. If the professor is not supportive of you raising concerns, it is better to find that out before you take the job.
Multiple comentors on the draft of this post told me that it is an even better idea to talk to your potential supervisor’s current students. This seems like excellent advice.
At some PhD programs, you will not have a supervisor when you start out, but will pick one later after some initial studies. This gives you a bit more of an opportunity to get to know the potential supervisors before you commit to working with one of them. On the other hand, your pool of supervisors to choose from is going to be quite small, so you should make sure that there is at least one professor there who you would like to work with before joining the program.
This is definitely not the be-all and end-all of advice about doing a PhD. It is very different in different fields, institutions and countries, so make sure to find some specific advice for the area you are interested in.
If you are interested in doing a PhD in physics or AI Safety, feel free to contact me for more specific advice.
“You should not do a PhD just so you can do something else later. Only do a PhD if this is something you would like to do, in itself.”
Why do you think this is the case? For example, I have noticed based on my search that nearly 60% of research roles in think-tanks in Europe have PhDs and that proportion is greater for senior research roles and more academic think-tanks. This does not account for the unmeasurable benefits of PhDs such as being taken more seriously in policy discussions. Isn’t it possible that 4-6 years of PhD work gives you more impressive career capital than the same amount of experience progressing from more junior roles to slightly more senior ones?
So almost half of them don’t. If you want a job at one of those think tanks, I would strongly recommend that you just go straight for that.
If you want to do research, then do the research you want to do. If the research you want to do mainly happen at a company or think thank, but not really in academia, go for the company or think tank.
There are other ways of getting a PhD degree that does not involve enrolling in a PhD program. In many countries, the only thing that actually matters for getting the degree is to write an defend a PhD thesis which should contain original research done by you. For example if you just keep publishing in academic journals, until your body of work is about the same as can be expected to be done during a PhD (or maybe some more to be on the safe side), you can just put it all in a book, approach a university and ask to defend your work.
This may be different in different countries. But universities mostly accept foreign students. So if you can’t defend your independent thesis at home, go some where else.
I did some googling.
In UK there are 4 ways to get a PhD (according to this website) and only one of them is the traditional PhD program.
Here is a discussion on independents PhDs. People are disagreeing on weather it is possible to do a PhD with out a supervisor, pointing towards different practices in different countries.
Several people claim that “The PhD process is about learning, not just publishing.”, but my impression is that this is a very modern idea. A PhD used to be about proving your capability, not monitoring your learning process.
Thanks for this post! I found it quite helpful.
I have a couple of questions about the checklist you linked, though I’m not sure how strongly you endorse it.
First:
and
seem to be indicators of neglectedness, which might make the topics more appealing to EAs. Do you think these are better pursued outside of academia? Or not at all?
Second:
Is there a better option for prospective PhD students who want a career in research outside of academia?
Some of the questions of the checklist, I would endorse more as guidelines, or warning signs than as strict rules.
Both those questions measure how much you can learn from others in academia. If you can’t take advantage of collogues, then going in to academia at all (even if you don’t intend to stay) will be lower value. So you might be more productive elsewhere.
The first one also says something about how easy/hard it will be to publish and generally get recognised. If you do something non-established, you will have a much harder time.
But there are two main reason you might want to step into academia anyway.
1) To influence other academics. (I think this is the main reason FLI chooses to be an academic institution.)
2) To get paid. (In cases where there are no other options.)
Lot’s of places out side academia does research. Companies, non-profits, think tanks, independent AI Safety researchers with Long Term Future Fund grants.
What is the better option depends on what research you want to do. The more abstract the more likely academia is a good choice. The more concrete the more likely it is not. E.g. charity evaluation is a type of research that I don’t think would do well in academia (though this is not my field at all, so I might be wrong).
Great response—thank you!
Very nice post! As a late-stage CS PhD student, I agree with pretty much everything. I wish more people would read this before deciding whether to get a PhD or not.
One extremely minor thing:
>From what I have heard, these are some problems you might have:
>[...] your supervisor wants to be co-author even if they did nothing to help.
In computer science (at least in AI at top institutions in the US), it is the norm for PhD supervisors to be a co-author on most or all papers that their students write, even if they contribute very little. One can debate whether this is reasonable. (I think there are various reasons why it is more reasonable than it may appear on first sight. For example, it’s good for the supervisor’s incentives to be aligned with the students publishing papers. Supervisors should get credit for causing their students to do well, regardless of whether they do so by contributing object-level results or not. Since the main way to get credit in academia is to be a co-author on papers, the simplest way to do this is for the supervisor to be a co-author on everything.) In any case, because this is norm, these co-authorship listings are, I believe, inconsequential for the student. People will typically expect that if the authors listed are a PhD student and their PhD advisor, the PhD student will have done the vast majority of the work. This is definitely different in other disciplines. For example, in economics papers that require a lot of grunt work, the PhD student author often does the grunt work and the PhD advisor does the more high-level thinking.
Hi Linda,
I am seeking advice for an eventual PhD thesis in AI ethics, safety and alignment.
I am about to finish my B.Arts (Philosophy Major) in June and am considering the next step in applying for masters programs to fulfil PhD thesis ambitions.
I note you studied at Grenoble and that it offers several masters courses in English (this is the only language I currently speak. I am willing to learn french, but am 46 now and don’t learn languages quickly—so I need English courses until I am better). I am seriously considering the English masters courses at Grenoble Alpes University and would like recommendations of which ones offered by Grenoble are best to pursue as the next step as foundations towards PhD.
My intuitions suggest based on feedback, that I need solid mathematics, logical and computer science foundations to pursue these PhD ambitions. There is a Grenoble course in English that is Masters in Mathematics and applications. Out of the Grenoble courses offered, is this the one you would recommend those like me apply for? or is an alternative more suitable?
Additionally, I am getting a PhD math tutor in order to prepare for validation of skills and knowledge requirements as an alternative entry pathway for graduate studies. Are you able confirm that Grenoble allows informal mathematics education that is later validated through tests, to be acceptable for entry into Grenoble graduate programs like the one I mentioned?
thanks and regards
Ed
I do not recomend going to France if you don’t already know some Frecnch. I got though my PhD ok in English, and leanring enough french to be able to by food and similar is not hard. But I did not have a social life for over 2 years and it was terrible, and eventually I left to finish my PhD from Sweden (my home country).
My plan was to learn french when I got there, and I tried. But I’m also slow at languges, and never got good enough to have a real conversation.
I recomend going to an English speaking country, or go to one of the small western Europe countries, (Nordics, Netherlands, etc) where most peopel speek good English.
If you decide to go to Grenoble anyway, I can’t help you with courses. My PhD program required very few coursers, and I think all the ones I took where for PhD students only. And the only good one whas a one-time course about particle phsics given by a German post-doc, who is probably not there anymore.
I don’t know much about masters programs in general. I did a undergraduate and master roled in to one program, which is common in Sweden, so I never had to look for a master.
I recomend joining this slack and ask in the applying-for-granschool channel.
Here’s more AI Safety grand school advise