5) No, but I am obviously biased due to not having a PhD. I doubt you are going to get an unbiased answer from anybody; everybody wants to believe they made the right choice.
> I doubt you are going to get an unbiased answer from anybody; everybody wants to believe they made the right choice.
I'm not sure where choices come in here. I specifically mentioned a mixed team; the choices have already been made, so the performance of the existing team is what I'm asking about.
As far as not getting an unbiased answer, that's why I'm asking HN at large -- hopefully there are enough people in enough environments to give an interesting and informative combination of answers. :)
I work at NVidia as a GPU architect doing ML applied research. I'm not a hiring manager, but I have interviewed a bunch of people.
1) Having a PhD will make it easier to get an interview, but it is not necessary. Relevant experience counts just as much. I only have an MSc.
2) That will vary a ton from one team to another and from one position to another. The interviewers are not going to change their questions depending on your education.
3) The same requirement as for PhDs. It will depend on the role, but in general I expect people value hands-on experience more than theory.
Except for some particular hiring managers with strong opinions, a PhD is not going to be a hard requirement for nearly any job. However, a PhD in a relevant area is going to be useful, just like any other relevant experience you may have.
> 1) Do FAANG companies hire non-PhDs for machine learning positions? Most seem to require a MS or PhD
That depends on what you mean by "machine learning positions" and what your bar for normalcy is. For research roles - these generally have distinct titles like "Research Scientist" or "Quantitative Researcher" - it is extremely difficult to get an interview without a PhD, let alone an offer. It happens, but rarely, because there are many capable people with PhDs and other relevant experience applying for the same roles.
If instead you relax the bar to also include software engineers who work with research scientists on model implementation and optimization then yes, people with "only" an MSc are routinely hired for these positions. These engineers are still credited on papers published as a result of their collaboration, and they still need to have a firm understanding of how the models work. The difference is that they don't tend to have leadership roles and don't develop novel theory - they are responsible for supporting the core research team and helping the research output become production ready software.
Both of these types of roles require strong coding skills, but the "hard" research roles require significantly stronger mastery of linear algebra and probability theory. As an example, compare the roles for Research Scientist[1] and Research Engineer[2] at Facebook. You'll find a similar bifurcation at other industry labs like Google, Microsoft and IBM.
If you have the opportunity to do either and you're optimizing your career for wealth maximization or research impact, it's better to obtain a role as a research scientist. That being said most PhDs do not end up at Google Brain or FAIR, so it's not a cut and dry decision. You can't just choose to trade n years of your life and an easier-to-obtain role on the periphery of research for the ability to do theoretical research at the best tech companies in the world later on.
Thanks. Can you list what courses one should focus on as a baseline to do research in ML? You mentioned Probability theory & Linear algebra, but what specific topics in those areas should one learn? Any other fields of math?
Probability, statistics and linear algebra are recommended because all of machine learning uses these as foundations. Any graduate level course in these areas will cover what you could possibly need. For people working in the more theoretical side of machine learning, maybe some analysis (functional specifically) might be useful.
A little disappointed with the one-sided point of view here - the author states
"...I deeply admire everyone I’ve listed, and I am not arguing that a PhD is never useful or never works out well" but never really gives examples of skills that PhDs do provide.
And there are absolutely career paths where a PhD is not required, but since many of the practitioners have one, can often be selected for (data science, biotech, biostats, lots of engineering research etc.) So without a PhD you might have a harder time rising as far as you would like in one of these positions (again being realistic that it's not all about talent, it can often be politics/perceived competence, which a PhD can augment).
It's just important to be honest about both pros and cons when writing advice articles like this.
Sadly, there is a weekly post on HN that hates on PhDs and grad school. I have tried my best to comment on each one and try to give a different perspective:
In CS you can earn over 50k a year with your stipend plus summer internship, you have a lot of unstructured time to explore your interests, and there are a lot of tenure-track positions without doing a postdoc.
This really depends on the role, the company and the culture. I am arguably the most senior engineer where I work (out of over 500 engineers) and I don’t have a college degree. We have many PhDs on staff. What matters at many companies and in many roles are results.
This gets it right, I think. And I say that as someone who does not regret taking five years to get a PhD.
A PhD is too long, narrow, and frustrating to do just to get a slightly fatter paycheck.
For some reason, that deep dive clicked with me, and I’m grateful for all the personal growth that came from that. But for the vast majority of people, I think it could easily end as an exercise in frustration.
Yeah. As a chem PhD, the average tends toward 4.5-5.5ish.
Between classes, teaching, and genuinely getting things done, it's not a fast experience.
Also, you probably arent making a ton more unless you go for an industry job...which is a little bit the opposite of the Platonic independence supposedly at the heart of the training method.
There are fields in the sciences (e.g., biology) where the PhD routinely takes 5-6 years on top of a bachelors, and you won't be hirable at the PhD level in academia and some parts of industry until after at least two multi-year postdocs.
When I finished up in the 90s, the market was flooded with applicants from the former soviet union as well as locals. I remember speaking to people I knew on hiring committees who told me of 1000+ applicants per open tenure track position at tier 2 and tier 3 schools.
Around that time I was looking at the postdoc train, saw where it (didn't quite) led, and chose a different path.
In computer science in the US, this is just false. Four years is the usual "officially expected" length of a PhD straight out of a bachelors' (which is the most common way to do a PhD in CS in the US). In practice it usually takes longer, sometimes much longer (seven+ years is not unheard of).
In the UK and Europe, shorter PhDs are much more common, in part because you're expected to do a master's before a PhD.
Even for two or three years, I don't think getting a computer science PhD is a good bet if your primary goal is making more money.
> In the UK and Europe, shorter PhDs are much more common, in part because you're expected to do a master's before a PhD.
That’s true of Europe but it’s quite common to go straight from a Bschelor’s to a doctorate in the U.K. and many other former British Empire countries.
I took eight years but I also got married, worked full time, and started a company before I finished. By the time I graduated, the CS department had started taking a much firmer stance on timelines, with a desire for most students to graduate within six years or sooner.
My PhD in experimental physics took seven years, which is also the average time to completion in that field.
The discrepancy here may be that "PhD time" in the US usually includes a masters, while it is counted separately elsewhere. That said, even if you break it down, my MS took 2 years and the PhD took another 5.
My BS took 4 years.
My MS took 2 years.
My PhD (computational physics) took 7 years.
I was the fast one at my school. Some of my peers (high energy physics, nuclear, etc.) took 9+ years.
Then again, I met my former business partner (not at the time) while in grad school. I was 2 years into my research, and he was a fresh new assistant prof in CS. About 9 weeks younger than me. Ph.D. in CS in 3.5 years.
I grossly underestimated how much I could learn by working in industry. I believed the falsehood that the best way to always keep learning is to stay in academia, and I didn’t have a good grasp on the opportunity costs of doing a PhD. My undergraduate experience had been magical, and I had always both excelled at and enjoyed being in school. The idea of getting paid to be in school sounded like a sweet deal!
Wholeheartedly agree. Aspiring PhDs discount what industry can teach them. The problem is compounded by undergrads who have zero industry experience when they graduate.
If you spend time working in a lab with grad students, listen to them. Heed their points about the field and their boss.
Research is research, and maybe you'll have some idea of what the technical details of the field are. But the only way you know what your life will be like is to pay attention. It's not bad, but you're trading something real to have "Dr." on your magazine subscriptions.
My biggest concern is more people getting PhDs, and the process becoming the New Bachelor degree, particularly in STEM.
I've dealt with a lot of people in industry, and so far, only those in research labs at the FAANG are of the quality that you would expect in academia.
In the majority of industry "science" appears to be a dirty word and "evidence" means "oh my buddy did this so it must work". The bar is depressingly low.
I have a phd in a Communication-realted field. I like my job at a community college but I am over-educated for it. However, I burned out on trying to get published. My advice is if you have doubt about going to grad school then don't. There are tons of people competing for every job. There is no guarantee you will get a job and there are fewer and fewer tenure track jobs. I was lucky to get out of adjuncting hell and I was lucky that I don't have a spouse or kids so I could afford to live on 24k for a few years.
> There are tons of people competing for every (faculty) job.
This is not at all the case in Computer Science. Especially permanent positions at CCs and lecturing at universities, there are more roles than people at the moment. Mostly because the worst-case alternative is taking one of the plentifully available 100k-200k industry jobs. Especially in ML.
Generally, "don't get a Ph.D. planning that you'll be a professor" is good advice. CS at the moment is the exception that proves the rule. Especially outside of R1.
If you don't have doubts about ANY major life choice, then you are the exception.
My experience has been great. Straight out of grad school I got several tenure-track offers from R1 universities. I'm not a super star and graduated from an unranked department.
This is total anecdata, but out of my closest college friends, five of them went to get a PhD, while I went to industry. When they graduated, I was already making more than of all of them, with five years of industry experience. And I made more them them for the next thirteen years too. I never hit any magical ceiling where a PhD was necessary.
In other words, don't do a PhD if you're in it for the money.
It may open a few more interview doors for you, but honestly, at least in my experience, I wasn't even aware of which of my coworkers had PhDs. When I eventually found out, they were all just slightly older than me working at the same pay grade.
I was in a PhD program, I dropped out. I have been working in R&D at a large company for a while now; many of my best colleagues have PhDs, many others do not. As far as my personal experience is concerned, a PhD has no bearing whatsoever on your capacity to do solid research.
That said, I am seriously considering going back on the PhD route, because I think I’d like to spend more time teaching down the line. Kind of silly, but I have only a master’s and it seems like most higher ed institutions do not consider hiring you as a professor unless you have the magical piece of paper.
> a PhD has no bearing whatsoever on your capacity to do solid research
Having worked in research at a few major tech companies now, when hiring research staff I’ve found they typical want someone in a strong position in the research community, typically evidenced by a strong publication record in the field they are recruiting in. While this can certainly be done without a PhD, I find it to be a bit rare. Our current research group is around 10 people (in a ~2K person company), 9 have PhDs.
> PhD has no bearing whatsoever on your capacity to do solid research
In my experience this is not true except unless you are super genius. Most folks without PhD often keeps making same naive mistakes, for example, not studying previous state of the art, not recording experiments properly, heuristics instead of rigorous analysis and so on. PhD trains to avoid all these. It allows you to build network, identity great researchers in the field as role models and understand what scientific scrutiny entails. It is not unusual to identify paper written by someone not experienced vs someone experienced. For example, a person without PhD would often neglect to mention scale in the graphs, compute variances in findings, describe figures properly and so on. These might look minor cosmetic things but it often goes long way in overall rigor.
As far as my personal experience is concerned, a PhD has no bearing whatsoever on your capacity to do solid research.
That is all well and good, but I can't name a single person in my research field who has a decent publication record without a PhD (finished or in progress).
But it’s so sad to be in it just for the money. If a phd will allow you to eventually do the things that you love, that should be the priority.
Of course you can do what you love also without a PhD, but I’d argue that a PhD will eventually allow you to explore fields and things that are not economically viable.
In fact I would say that it’s just plain sad that we have to choose something based on how economically viable it is (industry) rather than purely for its interesting properties.
If you’re well off (from family or something like that) then sure you can of course do it in your own.
However I’d rather have a system that would still allow me to get some sort of basic income (to afford a living) while doing, say, archeology studies in Ancient Rome. Such system is academia right now.
> If you’re well off (from family or something like that) then sure you can of course do it in your own.
I grew up in a below-average income family in a below-average city in America, and I carved this out for myself despite zero undergraduate degree. This meme that you have to be well off or that everyone who does it had benefactors is kinda silly.
I realize I am a sample size of one, but I really don't care for being wholly discounted in these arguments as if such a path is impossible. Was it exceptionally hard? Yes. Would it have been easier if I just finished my BS and an accepted-to MS? Also yes.
But I didn't for various reasons, and it still worked out because most of the factors that make a successful entrepreneur, scientist, and employee are exogenous to formal education anyway.
I advise people regularly to continue on with formal education; it is not like I think this is a very good track to take. But it is one that is available to those who self-study their ass off and don't mind working menial jobs to put food on the table for themselves and their families with great sacrifices.
You can, but you won’t necessarily get the support that you can from a stipend, adviser, upper division courses, student colleagues, conference travel, and so on.
I've turned down a handful of offers to teach classes at university and I'm about 34 semester credits short of a Bachelor's Degree.
So I guess the land you need to be in is the United States of America, at least in my case anyway. Might be true elsewhere but I haven't tested my luck.
Being a professor also involves doing original research, leading and managing larger research projects, advising and mentoring grad students and publishing original work. Teaching undergraduate classes is just one small part of being a professor, and in many cases the part many professors find the least interesting.
I happen to have more than a handful of peer-reviewed papers under my belt, including some first authorship credits. That's pretty much how I got the offers.
Oh indeed. If your main reason to want to be professor is that you want to teach some classes and publish some papers then there are definitely options to do that via working in industry (depending of course a lot on your field). However if you actually want to be a professor and work full time within the university world, with all the pros and cons that brings, then it's very hard to do without a doctorate.
But I absolutely agree if you 'just' want to be involved with some research projects, publish a bit on the side and perhaps teach a few courses then there are many ways to reach that goal that don't involve becoming a professor.
Sounds like an Adjunct position which I hate to break it to you is no where near Professor. You will likely make less than minimum wage, I have seen salaries of $3,000 per class taught, and have essentially no room for advancement.This is a common end point for PhDs that didn't get a Tenure track position and have no other choice.
If you aren't a professor and you are teaching classes you are an Adjunct/Lecturer which is a dead end career-wise. Maybe nice as a side gig or when you retire but a horrible position to be in otherwise.
Edit: I actually don't know what you are arguing, I think the point is that you won't get a job as a Professor without a PhD, getting teaching offers is completely different and happen to people in industry without PhDs all the time.
there is definitely a naive perception in some circles right now that a doctorate studying deep learning is a path to riches and fame. Due to misleading reports of million dollar salaries for new PhDs and so on.
It's a degree with a high opportunity cost that won't pay off financially in the long run, even with no tuition or debt. In the past I think everybody understood this but now I'm slightly more worried.
The vast majority of PhDs aren’t in ML, however. Even in ML, most are probably not in it for the “more money,” so the minority that is might actually be justified looking at all the JDs for more advanced ML jobs.
However eventually the next AI winter will come and the old lessons will be relearned again.
You should not go into a PhD program in any field if you're in it for the money. This is because your department will have at least a few grad students who are in it for the love of the field, and they will work insane hours. They will have no life, and will be fine with that. They will be nearly broke and will be fine with that. You will be in competition with those people (for thesis advisors, access to equipment, grants, etc) and you will have a very miserable experience.
Ironically this is even true in investment banking. If you don’t love making deals above all else, you will never even make it to the point of being able to make the big money deals, you will wash out in the first few years.
I have a PhD in physics. At the time that I finished college, and I think it's still somewhat the case today, the relevance of a PhD varied from one field to another. Maybe scientists just take longer to ripen. As an example, I've noticed that startup founders tend to be older in science than in computer programming.
Don't do a PhD if teenagers are getting rich in your field. Instead, decide if you belong in that field, i.e., if it's the kind of work that you actually want to do.
You obviously haven’t seen the other side :). If you are working as engineer/management/sales in software products, PhD rarely have advantage. However if you walk in to the places like FAIR, Google Brain, DeepMind, you will quickly find that your background as engineer/management/sales has little value compared to having a PhD in relevant area. These places are buzzing with innovations and creativity advancing the state of the art that rest of the world looks upon. Researchers are leading this front everyday and without PhD you won’t even get that job.
Several of the examples in the article of people without PhDs work at Google Brain so you are literally wrong in this case. They are obviously outliers but either way you aren't right.
Total anecdata too - when me and my friends finished undergrad, I went to work in the industry and two of my friends doing PhDs were much better paid than me throughout their entire PhD programme - the combination of grants, paid hours for teaching and marking made their salary about 150% of mine. Only after they finished their PhDs 4 years later my salary has matched what they were making at the start.
I suppose it could be unique in a way that their PhDs were sponsored by large companies and hence well funded, but I worked 4 years a C++ programmer and never caught up with their pay.
This would never happen in the us where average PhD stipends are $30,000 at great schools and even a two bit programmer in the middle of nowhere USA could make $50k easily.
A PhD is a license to do deep research. Deep research careers are very rare, but for the right kind of person, they're great. However, deep research by definition means it probably won't work, and there's just a lot more money in doing things that will probably work vs things that probably won't. So it helps to be extremely talented at deep research if you want to pursue the PhD.
Also, most engineering PhD's are bogus because most engineering "research" is actually not deep research - it's building prototypes that aren't quite useful but not quite that novel or interesting either. If you're in a PhD program and you're not doing something really interesting and fundamental, you're definitely in a tough spot.
What do you consider to be deep research? I agree that engineering research is not actually deep, but once you eliminate that almost all of the current (applied, and maybe theoretical too) Machine Learning work for instance is also dismissed.
As a guiding principle, I think deep research is answering questions that no one has approached in quite the same way before. In my own PhD work, I used the time to learn how to answer questions sufficiently (methodology) as well as how to recognize shape/distill questions in a way that they can be approached. Developing these skills don't require a PhD, but the dedicated time helped me.
I've used these skills to learn deeply and answer questions to great extent in business roles and drive some solid change in a few large orgs. It doesn't mean I do a job better than someone without a PhD. Also, the PhD signals that I have answered some deep questions to the satisfaction of others that answer deep questions (PhD committee).
That's why it's called Machine Learning and not Artificial Intelligence. It was an intentional differentiation to avoid the pure research academics who start every presentation with 'assuming infinite compute resources'.
ML is an Applied Research discipline and all the better for it.
There's plenty of academic research in machine learning.
I've never been to any academic presentation where they start like that. In-fact, more often they complain about the huge compute resources in industry.
I don't have a PhD, but I have run research programs.
The deeper into a field you get the more you realize that the parts which seem deep research aren't, and the parts which seem incremental improvements are actually very deep.
I can think of a number of things in machine learning which appear hard which are easy, and vice versa.
Theoretical justification for GAN improvements (eg, the WGAN paper): elegant but obvious, even though I'm not a mathematician.
Generative models for text including entities that remain coherent for longer than a sentence? We barely know how to even start thinking about this problem.
I'm not who you asked, but here's my current take:
Deep research maximizes uncertainty reduction (or information gain in other words). Uncertainty here could be model uncertainty if you are developing models, or a shift in the probability distribution for a particular question more generally. E.g., "Does P=NP?" would be an example of the latter.
It might be very general, applicable in many fields. Or it could be targeted at a particular field, but in a way which answers many questions.
Bayesian experimental design can do what I think is the easy part of the problem: maximizing information gained for a particular experimental problem statement. In my view, most of the time you can reasonably guess what Bayesian experimental design might tell you by looking at a state space of your exprimental data. So the math may not be strictly necessary. Unfortunately, not all research is experimental. And it won't tell you, for example, if you are missing a variable.
Framing the problem (which questions to ask, and how to answer them) seems like the most important part to me. Or at least it has been in my almost complete PhD.
These thoughts are in flux. I may have a different view in a year.
The phrase "most engineering PhD's are bogus" is astoundingly ignorant of the excellent work that goes on in the systems community (assuming by "engineering" you're referring to application-oriented CS research?). "Building prototypes that aren't quite useful but not quite that novel" is bad systems research, not all systems research.
I agree with you, but I do think it's incredibly hard and almost unnatural to do great engineering research, and that's why it's so rare. It's straightforward to solve someone's problem, it's also straightforward to study a phenomenon in nature (or computers) and learn something about it. It requires a very special type of person to create a new way of thinking that sets the groundwork for building systems that solve people's problems.
Microsoft encourages its research people to help product teams ship things based on their research. I had two interactions with PhD types on two different products.
The first was a researcher who talked to us a few times (just a few hours of meetings) then returned a few months later with a pile of MatLab code and a "problem solved!" attitude -- his simulations showed that things were working great. We looked at the code and while it taught us some things, it was obviously not shippable. None of his stuff wound up in the product, though it did point us down some interesting paths (some good, some bad).
Another researcher got a desk smack in the middle of the product team and spent 18 months sitting with us full-time, porting and dramatically improving his algorithms. I'd say that he learned just as much from us as we learned from him. His first few months were rocky, but he eventually became a productive and supportive member of the team. I hope that MSR treated him well upon his return.
Personally I think that research is great, but it's fantastic if you can occasionally ship your work to real customers.
> Also, most engineering PhD's are bogus because most engineering "research" is actually not deep research.
Hello! Engineering researcher here. I toy with things that are both wildly theoretical (information bounds for algorithms, inference in stochastic dynamical systems, etc.) to things that are fantastically and directly useful (design of photonic structures for LIDAR, (much) better AR/VR lenses, etc.).
While you're possibly right that I might be "building prototypes that aren't quite useful but not quite that novel," I disagree that they're "[not] interesting." In fact, I'd be absolutely surprised if in a few years, much of the applied "engineering" work our lab does (in contrast to the theoretical work) is not in constant use for on-chip photonics and fabrication of optical structures.
Don't do a PhD for the money (although it is more than adequate in STEM fields).
Do a PhD for the jobs that it unlocks (professor or researcher, mostly) or the type of freedom that it provides (it was 6 years of mostly unstructured time that I got to explore things that interested me while being paid). If all else fails, you can still go join a big tech company and make more than enough money to live a good life.
I'm a professor now and love it! Couldn't have happened without a PhD first.
I'm envious of that. I went through a master's program and I sorely miss that unstructured time to research and genuinely explore a topic. All of the learning I do these days is by force of business. I'm OK with that but it would be cool to get back to learning sometimes.
Could you talk about what life is like as a professor? I looked at your research interests and they lean heavily on the applicable-to-industry side - is that by purpose?
BTW - I think it's so cool that a professor is posting on HN. I think back to my professors and I couldn't imagine anyone of them being nearly as hip.
Thanks for the kind words! My research interests are definitely on the applied side on purpose. Some of the major criticism of research is that it is too theoretical or doesn't help people, so I wanted to do the opposite. I study actual engineers and the problems they face while working. This was actually my interest long before grad school. I wanted to make tools, plugins, and languages that were easier for me to use. Being very applied has worked out well for me so far since many people can relate to what I do!
I am a new professor but I can give you a short summary of what it is like. It is very unstructured. No one tells me how to spend my time, but I have to balance many different things: teaching a course, working on multiple research projects, writing multiple papers, writing grant proposals, reviewing papers for journals/conferences, traveling to conferences, recruiting and working with student researchers, etc. Some people like to describe it like running a startup.
This is a rather narrow perspective focused on (and giving examples from) one subfield (deep learning) at one point of time (year 2018). Have factors like (i) commoditization of software + hardware, (ii) the limited mathematics required, (iii) many open problems, and (iv) a lot of industry funding made a PhD unnecessary for doing research in deep learning in year 2018?: Yes. But does this mean that a PhD (with several advanced courses and a few years of struggle solving hard research problems) won't be useful for doing Computer Science research for the next 30 years? The answer is probably "No". If you want a long-term, intellectually satisfying research career, whether in academia or industry, a PhD is extremely useful.
Also, this field is quite easy to get into without any external support. You can become an expert in deep learning with a laptop and an internet connection. Most research fields will require a well equipped lab. You can usually find good labs in industry, but you won't have the same freedom to play with all those toys as you will have when pursuing a PhD.
I loved grad school. Loved loved loved it. I wouldn't give up the worst day I had in grad school for almost anything. And I love having the skills it taught me. I am a much better engineer and researcher than I could possibly have been had I taken almost any other route. I was given freedom in grad school that just would not have been present in most industry jobs. I was given a hardware project - a submersible robot - that I was completely in charge of on day 1. I had to teach myself machining, how to do electrical and electonicd works, how to do embedded programming, how to tune a PID loop. How to work with other students. How to give a persuasive presentation. How to come uo with my own ideas, how to convince other people that they were worth pursuing, how to quickly become an expert in a topic.
That being said, I am not under any illusions about the financial loss I experienced. I spent ten years making a pauper's wage, when if I had chosen to go into industry I would have been a software engineer... in the Valley... in 1993.
I also was in a remarkable lab in grad school. It made top-ten lists of “coolest college lab”. And that wasn't hype. The caliber of student and of professor was off the charts. And they were not only smart, the vast majority of them were good people. A lot of places aren't like that. And as the srticle correctly said: a toxic graduate school environment is worse than most toxic work environments. In any practical sense, students don't have HR protections. If you can't get your advisor to write you a recommendation, getting into a different program is nearly impossible. You can be worked 100 hours a week. You can be blackballed for getting sick, for taking vacations, for taking maternity or paternity leave (even if it is - and it almost certainly will be - unpaid).
The key is to find an advisor who is doing good work and who is sane and moral. If you can find that you're golden. If you can't you may be completely screwed.
As a current PhD student in CS -- seconded. I'm fortunate to have a great advisor, and my research career so far has been largely fun and productive as a result. I have also seen equally (or more) talented and hard-working students have worse experiences because of worse advisors. It's just hard to become a good researcher without good mentorship.
As far as practical advice for finding good advisors, track records of previous students can be a helpful first filter once you've made a list of people who wrote papers you like. Probably the most helpful thing is talking to current students at the admitted students day and listening to them. Bad advisors usually have at least one notable case, and there will probably be at least one person who'll take you aside and tell you about it. Listen to them carefully. Even great advisors can have at least one unsuccessful and bitter student through no fault of their own...but they usually don't show up to the admitted student events unless there's a real cause for animus.
Grad school was one of my best decisions. I went to a great school that was what I call the Goldilocks size, big enough to have a great faculty, equipment, and decent funding but small enough so that collaboration was the norm and the crazy horror stories of maniacal hours and/or cutthroat competition were normally self induced. My PI was an incredibly good guy and still a close friend. I met my business partner and co-founder while working with him the lab and we're now building a company that expands on the work we did in grad school.
That being said, I saw plenty of people not having the experience I did. This was almost always because i) they didn't really like research and didn't know it until they were there or ii) they picked a PI (PI = professor/boss) that was a really bad match for their work style and personality. Finding a lab & PI that matches your personal expectations about the PhD I would say is more important than the research focus. Don't choose something you'll hate learning about but ultimately the PhD can be more about learning how to teach yourself than the skills you learn during research.
What would interest me much more is if y'all liked studying or university life before that? Because for me.. I originally started to study CS because I wanted to learn stuff (was already working as a programmer) and hadn't ruled out pursuing a PhD afterwards per se. But the longer I was at university (German Diploma, 13 semesters for me, 9 minimum) the more I couldn't wait to leave - so even the thought of staying there went away quite quickly, although I wasn't "in academia" per se. I'm still not sure if I'm just more on the practical and pragmatic side of problem-solving and less in research.
I didn’t mind being an undergrad. But that was largely because of my friends and being an adult away from home for the first time. My friends and I built robots in our dorm room with our own money because the university didn’t have a program and wasn’t willing to accommodate us in any way. I enjoyed learning in class but it was always so rushed and stressful - you were always working up to an exam, then making it past and preparing for the next one. There was never any time to breathe.
Grad school was far, far better. Completely different league.
I had a similar experience during my physics PhD. Going in to the degree I thought I'd just learn physics more deeply. What I ended up learning is how to learn itself, not because of any course I took but because I spent years surrounded by incredibly smart people who asked the right (often difficult) questions about my work. Eventually I learned to ask myself the right questions about my work.
I spent my undergrad learning the theory, and then spent my PhD applying it to real problems, working with actual geniuses, being paid a reasonable salary to do it, and getting to discuss it with external colleagues in nice places. The experience will stay with me forever.
I can't agree with this enough. And, in industry, in a truly toxic culture, Human Resources ironically supports and defends the very toxicity that it's designed to protect employees from. "Oh, you went to HR, huh? So you think you're better than us?" Or "Let's have a one-on-one sit-down with your boss, and air our grievances," Or, "hey, there's this great new job opening up in our <2+ hr away> branch." Even promotions in a toxic culture can end up taking the course of, "oh, you want to be promoted, and you're in a technical role? That's nice and all, but we really value line management employees." My point is, things can be equally bad in industry.
This is offtopic, but I have to point out that these sort of cynical pronouncements are somewhat dangerous. You are probably coming from a good place and trying to warn people about being too trusting, but I see a risk in these sort of statements in making younger readers believe that this is a normative statement (as opposed to the positive sense I think you meant it in). Ideally HR does have something to offer the employee, balancing their interest with the company's. Let's not retreat entirely from that ideal.
Sure, I wasn't trying to suggest people turn sociopaths the moment they become employed by a large corporation. I expect people to be decent human beings and that includes compassionate interaction whatever the situation.
It's just good to remember HR is not an agent of the employee but of the employer, and you can't lean on the HR representative as you would for example on your own lawyer.
I second (by now probably seventh or eight) most of your post: I loved most of my time in my Math PhD program.
> A toxic graduate school environment is worse than most toxic work environments
I agree with this 100%. I would even extend this to overall academia pre-tenure (and maybe tenure as well). Moving to a different employer is usually easy and non-traumatic; in PhD program you are stuck -- leaving will mean abandoning your half-written thesis and starting from scratch, etc. Thus it is critical to avoid bad programs, either with a toxic environment or those that treat grad students as long term slaves.
> A lot of places aren't (expand) ... with vast majority of smart and good people. ... You can be worked 100 hours a week. You can be blackballed for getting sick, for taking vacations, for taking maternity or paternity leave
With this I disagree. I had friends in many schools and while there were a few exceptions by and large the environment was very good: supportive and enabling without kid gloves. I was on the theory side, which surely made things easier (no expensive hardware or purchases to pay for), but I basically wasted third year of 5 of my PhD on aimless wandering: I could get no traction on any problem, would try something and drop it at the first challenge, etc. And I heard no complaints from my adviser -- he was checking in, asking if I want feedback or suggestions on problems to look at, but otherwise let me be. No 100 hour weeks, etc. I knew he was not thrilled, but he let that disease (or growth) run its course.
I much later spoke to other folks doing theoretical PhDs and found that this is not that uncommon: transition from doing great in classes (learning along an externally designed sequence) to planning and doing your own research may not go smoothly.
> transition from doing great in classes (learning along an externally designed sequence) to planning and doing your own research may not go smoothly
I would speculate that this is largely because students spend most of their attention on pre-planned classroom-like work and do almost no research-like work during the first 16+ years of their formal schooling.
I echo everything said above. Mine was in theoretical chemistry/biophysics. I was also married through most of grad school, which was a big reason I chose not to do a postdoc (the norm in my area) and joined a startup instead. I was lucky that I was able to do that.
This might be the most accurate description of grad school I've ever read. It's almost too poignant. It's this balance between freedom and toxicity that makes graduate school such a polarizing experience.
Recovering from a difficult graduate school experience, like the one you perfectly described, is insanely difficult. Financially, emotionally, and temporally.
I only have a bachelors of science, but I went to a top 10 school in my degree field (material science).
The caliber of student and professor was off the charts like you said. Everyone was infinitely smarter than I was. They had some of the nicest equipment available. I worked in a 3 research labs for 3 years in college, under 3 different professors.
My first grad mentor, I took everything for granted. Didn't really realize how great he was and how much of an effort he put foward.
My second mentor ... didn't give a shit about me at all. Actually, this lab was the most interesting too, which made it worse. He saw me just as someone who clearly was wasting his time everyday.
My third mentor, I started to appreciate the value in a good teacher. Not the greatest but made efforts in making the concepts much easier to understand. I cared way more about my final honors dissertation in this lab than previous research I did. I wouldn't even call the other 2 labs research, I just read papers all day and did lab tech work.
In the end I realized research is not for me. Its not fun, and painstakingly slow to see results. Still learned a lot.
I am considering doing a Phd in a completely unrelated area when I turn 50 (45 now).
Is there anyone in this thread done a phd that late? Obviously my motivation is different now to study - to really learn the subject. I am financially self sufficient and will continue to be, and hence making a living out of my phd is not a consideration.
Edit: I would like to be able to study in a university setup (not distance education). Main reason is to soak in all the related conversations / workshops and also I like being in a young environment.
I’m doing this with my masters right now. Taking distance courses in Operations Research at GA Tech. I’m 36 and have two years of part time grad school to go but it’s been an awesome experience. If it’s for you, im not sure why age matters?
A PhD is something I plan on doing much later in life too (35 now), a bucket list sort of thing once I've achieved financial independence and my kids are a bit older.
The hard part I imagine will be hanging out with mid-20-year olds everyday when I'll be in my late 40s already :)
>>I am financially self sufficient and will continue to be, and hence making a living out of my phd is not a consideration.
I'm curious about this scenario from the advisor or departmental perspective: If a PhD candidate is financially self-sufficient, does this mean that a potential advisor has one less mouth to feed when competing for grants? How does the power dynamic change between advisor and financially independent PhD candidate? What if the PhD candidate is capable of funding a whole (or good portion) of a lab themselves - is there a conflict of interest somewhere there between donor vs. principal investigator vs. PhD-candidate roles? Do financially independent PhD candidates have more, less, or no competitive advantage during the selection and admissions process for an R1 institution?
I often get asked about what my views on doing a PhD are (I am more-or-less finished with one now), and one of the ways I frame it is the following:
You know how you've take a course before where the professor was just surprisingly awful at teaching? These professors are often some of the most knowledgeable people in a subfield of the subject you are taking, yet their teaching ability is severely lacking and you have to scramble to learn the material some other way (or just never learn it).
During a PhD, there is a decent chance that your adviser is similarly a bad manager. Unfortunately, having a bad manager for 5-7 years of your life can be a fairly awful experience. You will work with someone who you, on the one hand, look up to, but on the other hand, who seems to not care at all about your mental health, your possible career desires outside of academia, your work/life balance, or the exact reason why this week was a rough week for research in your (human) life.
I have a lot of other thoughts on the matter, but I thought I'd try to keep this post more concise =).
To be fair by the time you have finished a PhD you should be capable of learning / doing research by yourself. I don't have a PhD but learning by yourself is a vital skill for any competent software engineer.
not care at all about your mental health, your possible career desires outside of academia, your work/life balance, or the exact reason why this week was a rough week for research in your (human) life.
I wouldn't call that being a bad manager, but rather being an asshole.
As a professor, I often think that one of my biggest weaknesses is indeed management skills. After all, we suddenly find ourselves having to manage people without any training in the matter, and when our true call is typically science, not management.
First principles of doing a PhD and taking up an industrial jobs are quite different, which this article sidesteps. I am talking from the perspective of someone who did a PhD, postdoc and migrated to be a founder/CEO.
A PhD system trains you to think about unsolved problems in an given domain deeply with a larger time runway. The end goal is not a tangible product that reaches millions of people, but rather a set of ideas that can take a crack at the unsolved problems in your field in a novel way. A good work should inspire others in the field, and eventually a larger audience to pick them up and expand and build on top of it. To give a small example, a majority of the fundamentals of machine learning was charted out by many, many PhD works over the last 40 years. Implementing a linear classifier is 2 lines of code in 2018, but many Bothans died to bring us this information :-) .
The goals of industry are more immediate. Expect for a privileged few research labs in industry, your work is expected to be monetized, and rightly so. The goal is for you, if you run the business, else your management team to first figure out a problem of high relevance and monetary value. Build products/solutions for that problem, that can be used by someone who is less versed/ambivalent of your technical solutions. Efficacy of solving that particular problem often defines the merit of your contribution.
The fundamental of choosing the PhD or industry should be taking stock of what kind of contribution you want to make as an individual. If it is a few set of ideas to science, which on a later date might become something fundamental in our understanding of the world, then PhD is a good path. If it is a set of contributions towards a product/solution that eases the pain of many users then go into the industry first.
This is quite an unfortunate article. Author goes on to list several folks who don’t have PhD but have “made it” and asserts that many people can do fascinating and cutting-edge work without PhDs. There are always outliers in stuff like this but ask yourself: How many people you know who don’t have PhD and have freedom to explore at work full time what truely interests them? Author has rather twisted view of the selection bias.
People should do PhD if they are genuinely interested in doing scientific research. If you are doing PhD under pressure or in hope of getting better paying jobs you will be dissopointed. It is an arduous process and taking up your precious years but it gives you opportunity to have freedom to explore and work on your interests for rest of your life. You won’t be coming to office everyday doing assigned task on your backlog and reporting your status in scrum meeting. Instead you will be reading about new creative work that was literally published yesterday, mulling over that in lunch with colleagues and apply your original ideas to actually get published under your name. The downside could be lower pay and/or no stock bonuses for many outside of hot areas like AI. But in general, you have much better chance of doing cutting edge work that you are truely passionate about if you have PhD in that area.
Your second paragraph sums everything up perfectly. I am a tenure-track lecturer, and love my job for all the reasons you mention, despite the relatively poor pay. I essentially have the freedom to work on whatever the hell I want; and that to me is priceless.
I would also mention teaching: some academic staff hate it because it takes time away from research. This is true but my experience is that if you put time into being a good teacher then you get to take your pick of the brightest students to help you with your research, which can lead to interesting Masters projects which lead on to PhDs.
> How many people you know who don’t have PhD and have freedom to explore at work full time what truely interests them? Author has rather twisted view of the selection bias.
I don't have a degree and I have that. Also I know a lot of people with a PhD who do not have anything remotely resembling freedom.
The author might be biased but it's good to hear it from that side once in awhile, rather than be deluged with "you can't do science without a PhD" and similar sentiments (I work parallel to academia and get plenty of shit from their ivory towerism, hence my comment).
> The author might be biased but it's good to hear it from that side once in awhile,
In what way is the author "biased" exactly? She did this course, was in both a PHD and tech, and tells people why she thinks it's worth considering not getting a PHD, listing various (valid IMO) reasons.
You won’t be coming to office everyday doing assigned task on your backlog and reporting your status in scrum meeting
Including former colleagues I probably know a couple of hundred PhDs who do this exactly the same as mere Bachelors. Outside of a few niches a PhD is barely recognised in industry. Sorry, but that’s the truth.
I think they're referring to life doing a PhD, not day-to-day industry (although they do earlier say 'rest of your life' - which I assume means continuing into an academic career)
> People should do PhD if they are genuinely interested in doing scientific research.
This condition is necessary, but not sufficient.
After finishing my PhD in quantum information, I turned to data science. I couldn’t be happier about this transition! Compared with academia, data science world looks to me like a wonderland.
tl;dr: faster pace, more freedom (sic!), way less bureaucracy and politics, etc
I have a PhD, worked as a post doc and found I never had the freedom at work to explore full time what truly interests me. I had to spend a significant portion of my time seeking funding or jobs. I had to work on what my advisor wanted me working on. A job focused on teaching college level physics, which is what I got into it to do, seemed to stretch further and further away as I moved forward. That said, I'm still glad I got a PhD. I just think your experience might be a bit closer to the ideal than what many experience.
The first graph (titled "THIS IS YOUR EDUCATION, THIS IS YOUR SALARY") in the following article will disabuse one of the misdirected desire to get a PhD:
"Career Guide for Engineers and Computer Scientists"
by Philip Greenspun
The Author is unfortunately right about the academic PhD program today, but mixes up two different things. A PhD is meant to teach you about what I call the "body of human knowledge" and to initiate you into it. Period. It does not matter if you can make a learning out of it, nor it helps you earn loads of money. All those hours of reviewing done by the DC members, journal reviewers etc. are done without the expectation of a single penny. It is done to place the PhD research in context among the world of human knowledge. Once you have done a PhD, you are supposed to get more confident dealing with unknown problems in general. Granted, other pressures like career, money and sometimes mere survival! take precedence after PhD but those are beyond the scope of your PhD.
Of course, you can work on cutting edge problems without a PhD but probably do it better with PhD. A PhD teaches a lot of intangible things like being comfortable with unknown problems , working patiently towards an end goal, get over the fear of failure etc. One thing it does not teach is how to earn money :).
In this context I would like to highlight unique cases where you do an "Industrial PhD" where you are working in a company and chose a relevant problem for a PhD. These have the "best of both worlds" and one is not bogged down with typical pressures in a regular PhD like those pointed by the author.
That said, the final word is that a PhD is merely a conduit. Its more of what you chose to do with it than what it is , which matters at the end.
Mind the survivorship bias when reading all the comments from grad students and PhD's here. Drop outs are much less likely to open this comment section or post a comment.
I'll try to give an answer and a response
more generally to the material in the OP.
Background. I got a BS in pure math with
nearly a second major in physics and
worked in computing and applied math for
problems in US national security around
DC. Jobs were very easy to get; at one
time my annual salary was 6 times what a
new high end Camaro cost; much of the work
was challenging for both the computing and
the applied math; I was learning a lot of
both computing, e.g., algorithms in Knuth,
and applied math on the job and also
especially math in independent study on
evenings and weekends. Soon I got a call
from a college friend to join FedEx. I
used my background in computing, with a
little applied math, to write some
software, in six weeks, to schedule the
fleet. The results pleased the BoD,
enabled crucial funding, and saved the
company. Later the BoD wanted some
revenue projections. I formulated and
solved
y'(t) = k y(t) (b - y(t))
for time t, revenue y(t) at time t, rate
of growth y'(t) = d/dt y(t) at time t, and
full revenue potential b. The BoD was
pleased, and ..., to make a long story
short, I saved the company again.
Then I got a Ph.D. in applied math from
the engineering school of a famous, high
end research university.
Now I'm doing an Internet startup, a Web
site, basically a new and very different
search engine -- for the, IMHO, very large
part of search handled at best poorly by
the existing Web sites and well known
techniques.
For the startup, my applied math
background is crucial: The crucial,
enabling core of the startup is some
original applied math I derived based on
some advanced pure math prerequisites I
got both in grad school and in independent
study. I already knew enough computing
except had to learn how to bring up a Web
site that has been easy for the user
interface but due to the core math
somewhat tricky on the server side. I
learned Microsoft's Visual Basic .NET (for
the programming language -- I like it),
ADO.NET (for the Web pages), and ASP.NET
for the (relatively meager) use of SQL
Server. The main difficulty was working
through 5000+ Web pages of documentation.
The first code is the first production
code, 100,000 lines of typing, 24,000
programming language statements, and lots
of documentation. The code seems to run
as intended, but I need to add some data.
Some lessons:
(1) Math. IMHO, the key to powerful,
valuable, new applications of computing is
applied math. That is, if we accept that
the big opportunity is to exploit and
apply current computing, then we might
notice that whatever we code to put out
data users will like, as information,
entertainment, whatever, is necessarily
mathematically something, understood or
not, powerful and valuable or not. So, in
some sense, from 100,000 feet up, it
should help to proceed mathematically,
with possibly advanced prerequisites, some
new results focused on the application in
mind, and complete with theorems and
proofs. Just IMHO. But I don't know
anyone with a yacht over 100' long that
did that; I suspect that very few people
agree with me. Maybe what I am saying is
a hopeless wild goose chase or a great
green field opportunity -- you judge.
(2) Getting a Ph.D. In a nutshell, the
three most important parts of getting a
Ph.D., in no particular order, are
research, research, and research. Yes,
there can be courses, credits, grades,
teaching assistant positions, weekly
research seminars, qualifying exams, etc.,
but at a research university what can "cut
through", dominate, and trump all or
nearly all of that is research. The
research should be publishable in a good
peer-reviewed journal of original
research; if there is any question, then
send it in.
The criterion for a Ph.D. dissertation may
be something like "An original
contribution to knowledge worthy of
publication." -- so, in case of some
doubt, publish the thing.
The usual criteria for publication are
that the work be (i) new, (ii) correct,
and (iii) significant.
My wife was fatally injured in her Ph.D.
program. The OP outlines a lot of just
what happened to her. To have time to try
to help her, for a while I took a slot as
a B-school prof. It didn't work -- lost
her anyway.
I never for even a milli, micro, nano,
pico, femto second wanted to be a prof.
Instead, I wanted to be solving problems
in business, the money making kind.
(5) Non-Academic Career. Then I tried to
get my career going again, outside
academics. Bluntly, that didn't work very
well.
I made a mistake: I should have returned
to DC and gotten back into applied math
and computing for US national security. I
guessed that there would be opportunities
as an employee in business; I was wrong.
Bluntly, my view is that US business and
Ph.D. holders mix less well than oil and
water.
Part of why:
(A) Business is still a lot like Ford in
Henry's day: The manager knows more, and
the subordinate is there to add muscle to
the work of the manager. A manager has no
use for a subordinate who knows much and
resents or feels threatened by such a
person.
Supposedly lawyers have a solution: A
working level lawyer should work only for
a lawyer. Period.
Well, a working level Ph.D. should work
only for another Ph.D., and that criterion
would eliminate nearly all jobs for a
Ph.D. in business.
Not even a CEO wants a Ph.D. around except
maybe tucked away in some side
organization, out of the main work of the
business. E.g., the CEO is plenty sure
that he is the only really important
person in the company and, thus, certainly
doesn't need a Ph.D. or some academic
background he (the CEO) doesn't have!
(B) Business regards Ph.D. holders as blue
sky dreamers out in the ozone who refuse
to contribute to the business, who really
want to publish a lot of papers and get a
prof slot in academics.
(C) If a Ph.D. person does anything
original relevant to anything in business,
usually the business will regard this
person as a threat.
(D) Suppose a Ph.D. takes on a practical
business problem:
(i) If the Ph.D. successfully uses their
advanced knowledge to get a good solution,
e.g., one that makes a lot of money for
the business, then everyone else in the
business, even the CEO and the BoD, will
feel threatened and/or jealous.
(ii) If the Ph.D. fails to get a good
solution, then everyone else will take the
opportunity to denigrate both the person
and the Ph.D. degree -- "I always thought
that a Ph.D. was just a useless, hopeless,
worthless impractical dreamer out in the
ozone, and now we know for sure.".
(6) A Ph.D. in a business research
division. Yes, some businesses, say, ones
with some loose cash, might set up a
research division, hire a Ph.D. as the
director, and hope for something good. If
nothing good happens, well, the company
could afford the wasted money.
Generally, connections about the actual
business between the research division and
the rest of the company are more awkward
than a skunk at a Victorian garden party.
The rest of the company doesn't want to be
bothered, sees various threats, etc.
Here are some of the reasons for such a
research division:
(A) Luster. Use the research division to
impress the public, for good PR, to
impress customers, to cover the rear
exhaust port of both the CEO and the BoD,
etc.
(B) As a patent shop. So, the research
division can develop a patent portfolio,
maybe dozens, hundreds, thousands of
patents. Then some specialized lawyers
can use that patent portfolio as a, call
it, battering ram against any would be
competitors. There can be cross licensing
deals, revenue, etc.
(7) Career direction. It's your career.
In this career, there will necessarily be
some directions you will be pursuing.
Some directions are good; most are not.
It's up to you, and maybe your family,
closest, trusted friends, etc. to pick, at
least try to pick, a good direction(s).
If you just look for a job, get some
offers, and take the best offer, then
likely you will be following the direction
...
TLDR: Not every boss appreciates an employee with a Ph.D. If you want to go your own way, start your own business, otherwise you are expected to follow the lead of your boss.
To be more clear and blunt, if your boss gets fired or leaves or if the whole group is fired, you can be on the street, with an advanced education and some recent, specialized experience (that likely didn't make big bucks for all concerned), and 40+ and with a tough time ever again holding a job relevant to your education, experience, interests, capabilities, or potential -- EVER.
My guess is that there are a lot of Ph.D. electronics engineers over 40 who are essentially unemployable at anything close to their past who want an electrician's license so they can install AC wiring in houses.
Maybe the best job they can get is to be a clerk in the electronics department at Wal-Mart.
Imagine, two guys just out of high school. They both mowed lawns as teenagers. Joe wants to continue that as a business, at first still living at home. Tom goes to college, continues, gets an electronics Ph.D., and gets a job in an electronics company.
Then they are both 35: Joe now has five lawn and garden crews, each with a late model, crew cab truck, a $5000 trailer with 4 riding mowers each worth $15,000. He has clients in upper class residential areas and small to medium commercial lots. He does weeding, soil testing, fertilizer applications, mowing, edging, shrubbery trimming, landscaping, etc. Tom's employer has a policy: By age 35, promoted into management or fired. Tom gets fired.
Joe is now much better off than Tom. Tom would do well to look for a job with Joe and rise to managing one of the crews, planning and marketing for higher end parts of the business, etc.
> Ph.D. dissertation may be something like "An original contribution to knowledge worthy of publication."
So true - now most students are happy with, "I came into the office every [most] days and did [the thing]. I wrote it up so where's my PhD."
They don't evaluate their own data critically for interesting information - they did the experiment or simulation as told and they think this is their job.
I'm in a similar boat thinking about getting a PhD in AI/ML. I have a masters degree and in 2 years of my current job (research based) have already published 2 conference papers one of which I'm a lead author. My work colleagues who aren't research oriented suggested me to get a PhD instead because of the effort I put in reading papers and spending hours on solving a problem. It makes sense to me at times because not many people in my company have a hardcore ML background and when I'm stuck reading a paper or solving a ML problem, it's difficult to find someone to discuss it with. I end up spending most time in figuring it out or eventually giving it up. This when I miss academia the most.
I'm on H1B visa and have money constraints so giving up a 6 figure salary and living on stipend is something freaks me out.
I'd really appreciate if anyone reading this comment have something to suggest. Thanks a lot HN community, you've been a great support.
It is even worse in Europe where you are often required to have Master's degree before you can even start a PhD.
People in my country usually start PhD. at 25 and take at least 6 years to finish, because the universities use them as cheap workforce and aren't incentivised to allow students graduate quickly.
> People in my country usually start PhD. at 25 and take at least 6 years to finish,
uh ? which country are you in ? In france like you said you have to do a master before BUT a phd is only 3 years. I'm 26 and I finished my phd - had my defense earlier this year when I was still 25.
My husband is a German who did a Mech E PhD, and indeed was 31 when he finished, despite going through the pipeline completely on schedule. Back then, German students finished high school (Gymnasium) at 19, then did a Diplom (think combination bachelors/masters) which took 5 years.
In his particular institute, the better you were, the LONGER it often took because the professors found stuff they wanted you to help with unless you were really good at boundaries. My husband took 7 years; a very kind, very bright friend of his needed 8.5.
This used to be the situation in Europe but it got out of hand, as a result all new PhDs in Europe must be advertised as three year programs, and completed in four at a maximum.
Students apply for extension after four years. The only difference is that after 4 years you no longer receive the PhD. stipend and have to be paid from project grants. This might further reduce your income or force you to work even more on projects related only tangentially to your thesis.
I have worked in a few academic institutes (and still know many friends in the same places). 5 years seems to be the norm, with some people taking up to 7.
I am university lecturer, and the rules have been tightened dramatically in the last few years, precisely to prevent 7 or 8 year-long PhDs, which were indeed getting increasingly common, and benefited nobody.
Now, students must submit their thesis within five years (one of my students very nearly failed his PhD because of this), and there are funding restrictions for continuing beyond three.
I didn't know this was a Europe-wide rule, but as a professor in Spain I can attest that the 3 years + 1 year extension rule is now enforced and met. If your time elapses and you haven't completed your thesis, you're out.
I am in Spain. Maybe it has changed but a few years back it was the norm. The institute I worked for had a 5 year rule and you got kicked out (if you were doing research), but a few people managed to get extensions around that.
Depending on how cheap that may not be such a bad thing. You are effectively having a "real job" from 25 (if a slightly worse paid one than in industry), and most importantly you get the PhD without accumulating any more debt. This is an important factor in equality and recruitment meaning not only those that can afford grad school will do it.
In e.g. Sweden (that has this PhD system), the max student debt allowed (unless you go to a private lender which is more or less unheard of) is for 5 years full time studies. Meaning you already "maxed out" your student debt before starting tour PhD.
Most or many German phds are intended to be 3 years long, while in the US many people do 5 years, but it's basically a master's and phd combined. For instance a German phd will almost always be only research, no classes. So really it's not that different, master's and phdare just rolled into one in the US.
I am sad to hear that this is still going on. I got my PhD 25 years ago and it was the norm then. Postgrads were used as IT support as well as teaching assistants and anything else going.
At least I knew this when I started. And it motivated me to get finished in 3 years.
Someone did give me the advice in this article before I started. Which was good. I wasn't in it for the money. It was the freedom to work on things that interested me that mattered.
What no-one told me later, when I started a post-doc, was that the chances of getting a full-time academic job were close to zero... now that did feel like 3 years wasted.
As I posted in response to the author on Twitter - while generally agreeable, it would be nice if the post had a disclaimer that this is "What (bad stuff about doing PhDs) You Need to Know Before Considering a PhD". There are plenty of positives that one also needs to know when considering a PhD this article does not list... see the ensuing discussion here https://twitter.com/andrey_kurenkov/status/10342151109678407...
To me, this is whether you have good people around or not. PhD usually comes with good advisers who help you think in different ways than you have done before. However, such good advisers can be also at a company + you work on practical an modern problems (usually). For that you don't need PhD. In any case, once you ask whether you should do PhD or not it feels like you don't want to be left out. In that case, just do it.
I was once told by a professor who I respected very much how he gives advice to prospective PhDs. When someone asks "should I get my PhD?" the answer is always "no". Because the only people who will enjoy grad school are the ones who will do it anyways.
I have found this to be good advice for people who are thinking of starting companies too.
Yes, this is a good advice.
I never asked anyone whether I should do a PhD. I always knew that I want to do it. And I'm doing it right now, and I really enjoy it. (Speech recognition, Translation, Deep Learning; RWTH Aachen University, Germany.)
My answer is always "no" as well - for many of the reasons outlined in this essay. But really, most of the students who have asked me this question just want someone to say "yes" and don't end up listening to my reasons for "no" anyway.
However, I can't help but think that this mentality does scare away certain personality types that would otherwise make for great researchers. I fall into the category of people who ignored advice and did it anyway, but I try to be welcoming to other types too.
(PhD here) An ex-colleague in grad school caught me once on a bad day and I told her she shouldn't do a PhD, stick to a Master's. To this day she credits me with convincing her to graduate at the Master's level - and she is grateful and thanks me for it. None-the-less, I have always felt guilty because my PhD experience was great and that sort of responsibility on someone else's life is kinda heavy.
194 comments
[ 2.4 ms ] story [ 132 ms ] thread1) Do FAANG companies hire non-PhDs for machine learning positions? Most seem to require a MS or PhD
2) What are the interview questions like at FAANG companies for machine learning positions? Is the interview different if you don't have a PhD?
3) For non-PhDs applying, what are the math requirements for the job?
4) For people that have a PhD working in ML at a FAANG, do you feel like you use your PhD level skills day-to-day?
> 4) For people that have a PhD working in ML at a FAANG, do you feel like you use your PhD level skills day-to-day?
5) For people who work with a mix of PhDs and non-PhDs in the same field, do you notice a difference in output quality?
I'm not sure where choices come in here. I specifically mentioned a mixed team; the choices have already been made, so the performance of the existing team is what I'm asking about.
As far as not getting an unbiased answer, that's why I'm asking HN at large -- hopefully there are enough people in enough environments to give an interesting and informative combination of answers. :)
No, individual differences outweigh any pattern that I've seen.
1) Having a PhD will make it easier to get an interview, but it is not necessary. Relevant experience counts just as much. I only have an MSc.
2) That will vary a ton from one team to another and from one position to another. The interviewers are not going to change their questions depending on your education.
3) The same requirement as for PhDs. It will depend on the role, but in general I expect people value hands-on experience more than theory.
Except for some particular hiring managers with strong opinions, a PhD is not going to be a hard requirement for nearly any job. However, a PhD in a relevant area is going to be useful, just like any other relevant experience you may have.
That depends on what you mean by "machine learning positions" and what your bar for normalcy is. For research roles - these generally have distinct titles like "Research Scientist" or "Quantitative Researcher" - it is extremely difficult to get an interview without a PhD, let alone an offer. It happens, but rarely, because there are many capable people with PhDs and other relevant experience applying for the same roles.
If instead you relax the bar to also include software engineers who work with research scientists on model implementation and optimization then yes, people with "only" an MSc are routinely hired for these positions. These engineers are still credited on papers published as a result of their collaboration, and they still need to have a firm understanding of how the models work. The difference is that they don't tend to have leadership roles and don't develop novel theory - they are responsible for supporting the core research team and helping the research output become production ready software.
Both of these types of roles require strong coding skills, but the "hard" research roles require significantly stronger mastery of linear algebra and probability theory. As an example, compare the roles for Research Scientist[1] and Research Engineer[2] at Facebook. You'll find a similar bifurcation at other industry labs like Google, Microsoft and IBM.
If you have the opportunity to do either and you're optimizing your career for wealth maximization or research impact, it's better to obtain a role as a research scientist. That being said most PhDs do not end up at Google Brain or FAIR, so it's not a cut and dry decision. You can't just choose to trade n years of your life and an easier-to-obtain role on the periphery of research for the ability to do theoretical research at the best tech companies in the world later on.
________________________________
1. https://www.facebook.com/careers/jobs/a0I1H00000Mp2ZCUAZ/
2. https://www.facebook.com/careers/jobs/a0I1H00000LJm3MUAT/
"...I deeply admire everyone I’ve listed, and I am not arguing that a PhD is never useful or never works out well" but never really gives examples of skills that PhDs do provide.
And there are absolutely career paths where a PhD is not required, but since many of the practitioners have one, can often be selected for (data science, biotech, biostats, lots of engineering research etc.) So without a PhD you might have a harder time rising as far as you would like in one of these positions (again being realistic that it's not all about talent, it can often be politics/perceived competence, which a PhD can augment).
It's just important to be honest about both pros and cons when writing advice articles like this.
In CS you can earn over 50k a year with your stipend plus summer internship, you have a lot of unstructured time to explore your interests, and there are a lot of tenure-track positions without doing a postdoc.
A PhD is too long, narrow, and frustrating to do just to get a slightly fatter paycheck.
For some reason, that deep dive clicked with me, and I’m grateful for all the personal growth that came from that. But for the vast majority of people, I think it could easily end as an exercise in frustration.
Between classes, teaching, and genuinely getting things done, it's not a fast experience.
Also, you probably arent making a ton more unless you go for an industry job...which is a little bit the opposite of the Platonic independence supposedly at the heart of the training method.
But very few people I know want to be a PI.
When I finished up in the 90s, the market was flooded with applicants from the former soviet union as well as locals. I remember speaking to people I knew on hiring committees who told me of 1000+ applicants per open tenure track position at tier 2 and tier 3 schools.
Around that time I was looking at the postdoc train, saw where it (didn't quite) led, and chose a different path.
In the UK and Europe, shorter PhDs are much more common, in part because you're expected to do a master's before a PhD.
Even for two or three years, I don't think getting a computer science PhD is a good bet if your primary goal is making more money.
I tell students to expect 5-6 depending on whether you want to go to industry or academia.
That’s true of Europe but it’s quite common to go straight from a Bschelor’s to a doctorate in the U.K. and many other former British Empire countries.
I took eight years but I also got married, worked full time, and started a company before I finished. By the time I graduated, the CS department had started taking a much firmer stance on timelines, with a desire for most students to graduate within six years or sooner.
The discrepancy here may be that "PhD time" in the US usually includes a masters, while it is counted separately elsewhere. That said, even if you break it down, my MS took 2 years and the PhD took another 5.
My BS took 4 years. My MS took 2 years. My PhD (computational physics) took 7 years.
I was the fast one at my school. Some of my peers (high energy physics, nuclear, etc.) took 9+ years.
Then again, I met my former business partner (not at the time) while in grad school. I was 2 years into my research, and he was a fresh new assistant prof in CS. About 9 weeks younger than me. Ph.D. in CS in 3.5 years.
Wholeheartedly agree. Aspiring PhDs discount what industry can teach them. The problem is compounded by undergrads who have zero industry experience when they graduate.
If you spend time working in a lab with grad students, listen to them. Heed their points about the field and their boss.
Research is research, and maybe you'll have some idea of what the technical details of the field are. But the only way you know what your life will be like is to pay attention. It's not bad, but you're trading something real to have "Dr." on your magazine subscriptions.
My biggest concern is more people getting PhDs, and the process becoming the New Bachelor degree, particularly in STEM.
In the majority of industry "science" appears to be a dirty word and "evidence" means "oh my buddy did this so it must work". The bar is depressingly low.
This is not at all the case in Computer Science. Especially permanent positions at CCs and lecturing at universities, there are more roles than people at the moment. Mostly because the worst-case alternative is taking one of the plentifully available 100k-200k industry jobs. Especially in ML.
Generally, "don't get a Ph.D. planning that you'll be a professor" is good advice. CS at the moment is the exception that proves the rule. Especially outside of R1.
My experience has been great. Straight out of grad school I got several tenure-track offers from R1 universities. I'm not a super star and graduated from an unranked department.
In other words, don't do a PhD if you're in it for the money.
It may open a few more interview doors for you, but honestly, at least in my experience, I wasn't even aware of which of my coworkers had PhDs. When I eventually found out, they were all just slightly older than me working at the same pay grade.
That said, I am seriously considering going back on the PhD route, because I think I’d like to spend more time teaching down the line. Kind of silly, but I have only a master’s and it seems like most higher ed institutions do not consider hiring you as a professor unless you have the magical piece of paper.
Having worked in research at a few major tech companies now, when hiring research staff I’ve found they typical want someone in a strong position in the research community, typically evidenced by a strong publication record in the field they are recruiting in. While this can certainly be done without a PhD, I find it to be a bit rare. Our current research group is around 10 people (in a ~2K person company), 9 have PhDs.
In my experience this is not true except unless you are super genius. Most folks without PhD often keeps making same naive mistakes, for example, not studying previous state of the art, not recording experiments properly, heuristics instead of rigorous analysis and so on. PhD trains to avoid all these. It allows you to build network, identity great researchers in the field as role models and understand what scientific scrutiny entails. It is not unusual to identify paper written by someone not experienced vs someone experienced. For example, a person without PhD would often neglect to mention scale in the graphs, compute variances in findings, describe figures properly and so on. These might look minor cosmetic things but it often goes long way in overall rigor.
??? They taught all of us that in undergrad.
a person without PhD would often neglect to mention scale in the graphs
... and they taught us about the scale of a graph in secondary school ...
That is all well and good, but I can't name a single person in my research field who has a decent publication record without a PhD (finished or in progress).
Of course you can do what you love also without a PhD, but I’d argue that a PhD will eventually allow you to explore fields and things that are not economically viable.
In fact I would say that it’s just plain sad that we have to choose something based on how economically viable it is (industry) rather than purely for its interesting properties.
However I’d rather have a system that would still allow me to get some sort of basic income (to afford a living) while doing, say, archeology studies in Ancient Rome. Such system is academia right now.
I grew up in a below-average income family in a below-average city in America, and I carved this out for myself despite zero undergraduate degree. This meme that you have to be well off or that everyone who does it had benefactors is kinda silly.
I realize I am a sample size of one, but I really don't care for being wholly discounted in these arguments as if such a path is impossible. Was it exceptionally hard? Yes. Would it have been easier if I just finished my BS and an accepted-to MS? Also yes.
But I didn't for various reasons, and it still worked out because most of the factors that make a successful entrepreneur, scientist, and employee are exogenous to formal education anyway.
I advise people regularly to continue on with formal education; it is not like I think this is a very good track to take. But it is one that is available to those who self-study their ass off and don't mind working menial jobs to put food on the table for themselves and their families with great sacrifices.
What kind of research do you do?
So I guess the land you need to be in is the United States of America, at least in my case anyway. Might be true elsewhere but I haven't tested my luck.
But I absolutely agree if you 'just' want to be involved with some research projects, publish a bit on the side and perhaps teach a few courses then there are many ways to reach that goal that don't involve becoming a professor.
Edit: I actually don't know what you are arguing, I think the point is that you won't get a job as a Professor without a PhD, getting teaching offers is completely different and happen to people in industry without PhDs all the time.
Well is that because you aren’t in a job track that has anything that would need a PhD?
You do a PhD to scratch that itch, hopefully once you graduate you can keep scratching those itches, even if often that doesn’t work out.
It's a degree with a high opportunity cost that won't pay off financially in the long run, even with no tuition or debt. In the past I think everybody understood this but now I'm slightly more worried.
However eventually the next AI winter will come and the old lessons will be relearned again.
I have a PhD in physics. At the time that I finished college, and I think it's still somewhat the case today, the relevance of a PhD varied from one field to another. Maybe scientists just take longer to ripen. As an example, I've noticed that startup founders tend to be older in science than in computer programming.
Don't do a PhD if teenagers are getting rich in your field. Instead, decide if you belong in that field, i.e., if it's the kind of work that you actually want to do.
I suppose it could be unique in a way that their PhDs were sponsored by large companies and hence well funded, but I worked 4 years a C++ programmer and never caught up with their pay.
That was in the UK btw.
Also, most engineering PhD's are bogus because most engineering "research" is actually not deep research - it's building prototypes that aren't quite useful but not quite that novel or interesting either. If you're in a PhD program and you're not doing something really interesting and fundamental, you're definitely in a tough spot.
As a guiding principle, I think deep research is answering questions that no one has approached in quite the same way before. In my own PhD work, I used the time to learn how to answer questions sufficiently (methodology) as well as how to recognize shape/distill questions in a way that they can be approached. Developing these skills don't require a PhD, but the dedicated time helped me.
I've used these skills to learn deeply and answer questions to great extent in business roles and drive some solid change in a few large orgs. It doesn't mean I do a job better than someone without a PhD. Also, the PhD signals that I have answered some deep questions to the satisfaction of others that answer deep questions (PhD committee).
ML is an Applied Research discipline and all the better for it.
I've never been to any academic presentation where they start like that. In-fact, more often they complain about the huge compute resources in industry.
The deeper into a field you get the more you realize that the parts which seem deep research aren't, and the parts which seem incremental improvements are actually very deep.
I can think of a number of things in machine learning which appear hard which are easy, and vice versa.
Theoretical justification for GAN improvements (eg, the WGAN paper): elegant but obvious, even though I'm not a mathematician.
Generative models for text including entities that remain coherent for longer than a sentence? We barely know how to even start thinking about this problem.
Deep research maximizes uncertainty reduction (or information gain in other words). Uncertainty here could be model uncertainty if you are developing models, or a shift in the probability distribution for a particular question more generally. E.g., "Does P=NP?" would be an example of the latter.
It might be very general, applicable in many fields. Or it could be targeted at a particular field, but in a way which answers many questions.
Bayesian experimental design can do what I think is the easy part of the problem: maximizing information gained for a particular experimental problem statement. In my view, most of the time you can reasonably guess what Bayesian experimental design might tell you by looking at a state space of your exprimental data. So the math may not be strictly necessary. Unfortunately, not all research is experimental. And it won't tell you, for example, if you are missing a variable.
Framing the problem (which questions to ask, and how to answer them) seems like the most important part to me. Or at least it has been in my almost complete PhD.
These thoughts are in flux. I may have a different view in a year.
The first was a researcher who talked to us a few times (just a few hours of meetings) then returned a few months later with a pile of MatLab code and a "problem solved!" attitude -- his simulations showed that things were working great. We looked at the code and while it taught us some things, it was obviously not shippable. None of his stuff wound up in the product, though it did point us down some interesting paths (some good, some bad).
Another researcher got a desk smack in the middle of the product team and spent 18 months sitting with us full-time, porting and dramatically improving his algorithms. I'd say that he learned just as much from us as we learned from him. His first few months were rocky, but he eventually became a productive and supportive member of the team. I hope that MSR treated him well upon his return.
Personally I think that research is great, but it's fantastic if you can occasionally ship your work to real customers.
Hello! Engineering researcher here. I toy with things that are both wildly theoretical (information bounds for algorithms, inference in stochastic dynamical systems, etc.) to things that are fantastically and directly useful (design of photonic structures for LIDAR, (much) better AR/VR lenses, etc.).
While you're possibly right that I might be "building prototypes that aren't quite useful but not quite that novel," I disagree that they're "[not] interesting." In fact, I'd be absolutely surprised if in a few years, much of the applied "engineering" work our lab does (in contrast to the theoretical work) is not in constant use for on-chip photonics and fabrication of optical structures.
Do a PhD for the jobs that it unlocks (professor or researcher, mostly) or the type of freedom that it provides (it was 6 years of mostly unstructured time that I got to explore things that interested me while being paid). If all else fails, you can still go join a big tech company and make more than enough money to live a good life.
I'm a professor now and love it! Couldn't have happened without a PhD first.
Could you talk about what life is like as a professor? I looked at your research interests and they lean heavily on the applicable-to-industry side - is that by purpose?
BTW - I think it's so cool that a professor is posting on HN. I think back to my professors and I couldn't imagine anyone of them being nearly as hip.
I am a new professor but I can give you a short summary of what it is like. It is very unstructured. No one tells me how to spend my time, but I have to balance many different things: teaching a course, working on multiple research projects, writing multiple papers, writing grant proposals, reviewing papers for journals/conferences, traveling to conferences, recruiting and working with student researchers, etc. Some people like to describe it like running a startup.
I loved grad school. Loved loved loved it. I wouldn't give up the worst day I had in grad school for almost anything. And I love having the skills it taught me. I am a much better engineer and researcher than I could possibly have been had I taken almost any other route. I was given freedom in grad school that just would not have been present in most industry jobs. I was given a hardware project - a submersible robot - that I was completely in charge of on day 1. I had to teach myself machining, how to do electrical and electonicd works, how to do embedded programming, how to tune a PID loop. How to work with other students. How to give a persuasive presentation. How to come uo with my own ideas, how to convince other people that they were worth pursuing, how to quickly become an expert in a topic.
That being said, I am not under any illusions about the financial loss I experienced. I spent ten years making a pauper's wage, when if I had chosen to go into industry I would have been a software engineer... in the Valley... in 1993.
I also was in a remarkable lab in grad school. It made top-ten lists of “coolest college lab”. And that wasn't hype. The caliber of student and of professor was off the charts. And they were not only smart, the vast majority of them were good people. A lot of places aren't like that. And as the srticle correctly said: a toxic graduate school environment is worse than most toxic work environments. In any practical sense, students don't have HR protections. If you can't get your advisor to write you a recommendation, getting into a different program is nearly impossible. You can be worked 100 hours a week. You can be blackballed for getting sick, for taking vacations, for taking maternity or paternity leave (even if it is - and it almost certainly will be - unpaid).
The key is to find an advisor who is doing good work and who is sane and moral. If you can find that you're golden. If you can't you may be completely screwed.
As far as practical advice for finding good advisors, track records of previous students can be a helpful first filter once you've made a list of people who wrote papers you like. Probably the most helpful thing is talking to current students at the admitted students day and listening to them. Bad advisors usually have at least one notable case, and there will probably be at least one person who'll take you aside and tell you about it. Listen to them carefully. Even great advisors can have at least one unsuccessful and bitter student through no fault of their own...but they usually don't show up to the admitted student events unless there's a real cause for animus.
Grad school was one of my best decisions. I went to a great school that was what I call the Goldilocks size, big enough to have a great faculty, equipment, and decent funding but small enough so that collaboration was the norm and the crazy horror stories of maniacal hours and/or cutthroat competition were normally self induced. My PI was an incredibly good guy and still a close friend. I met my business partner and co-founder while working with him the lab and we're now building a company that expands on the work we did in grad school.
That being said, I saw plenty of people not having the experience I did. This was almost always because i) they didn't really like research and didn't know it until they were there or ii) they picked a PI (PI = professor/boss) that was a really bad match for their work style and personality. Finding a lab & PI that matches your personal expectations about the PhD I would say is more important than the research focus. Don't choose something you'll hate learning about but ultimately the PhD can be more about learning how to teach yourself than the skills you learn during research.
YMMV
Grad school was far, far better. Completely different league.
I spent my undergrad learning the theory, and then spent my PhD applying it to real problems, working with actual geniuses, being paid a reasonable salary to do it, and getting to discuss it with external colleagues in nice places. The experience will stay with me forever.
The main function of HR is not to serve employees. The main function of HR is to protect the employer from legal liabilities.
[0] https://www.amazon.com/Corporate-Confidential-Secrets-Compan...
It's just good to remember HR is not an agent of the employee but of the employer, and you can't lean on the HR representative as you would for example on your own lawyer.
> A toxic graduate school environment is worse than most toxic work environments
I agree with this 100%. I would even extend this to overall academia pre-tenure (and maybe tenure as well). Moving to a different employer is usually easy and non-traumatic; in PhD program you are stuck -- leaving will mean abandoning your half-written thesis and starting from scratch, etc. Thus it is critical to avoid bad programs, either with a toxic environment or those that treat grad students as long term slaves.
> A lot of places aren't (expand) ... with vast majority of smart and good people. ... You can be worked 100 hours a week. You can be blackballed for getting sick, for taking vacations, for taking maternity or paternity leave
With this I disagree. I had friends in many schools and while there were a few exceptions by and large the environment was very good: supportive and enabling without kid gloves. I was on the theory side, which surely made things easier (no expensive hardware or purchases to pay for), but I basically wasted third year of 5 of my PhD on aimless wandering: I could get no traction on any problem, would try something and drop it at the first challenge, etc. And I heard no complaints from my adviser -- he was checking in, asking if I want feedback or suggestions on problems to look at, but otherwise let me be. No 100 hour weeks, etc. I knew he was not thrilled, but he let that disease (or growth) run its course.
I much later spoke to other folks doing theoretical PhDs and found that this is not that uncommon: transition from doing great in classes (learning along an externally designed sequence) to planning and doing your own research may not go smoothly.
I would speculate that this is largely because students spend most of their attention on pre-planned classroom-like work and do almost no research-like work during the first 16+ years of their formal schooling.
If I had to do it all over again, I would.
Recovering from a difficult graduate school experience, like the one you perfectly described, is insanely difficult. Financially, emotionally, and temporally.
The caliber of student and professor was off the charts like you said. Everyone was infinitely smarter than I was. They had some of the nicest equipment available. I worked in a 3 research labs for 3 years in college, under 3 different professors.
My first grad mentor, I took everything for granted. Didn't really realize how great he was and how much of an effort he put foward.
My second mentor ... didn't give a shit about me at all. Actually, this lab was the most interesting too, which made it worse. He saw me just as someone who clearly was wasting his time everyday.
My third mentor, I started to appreciate the value in a good teacher. Not the greatest but made efforts in making the concepts much easier to understand. I cared way more about my final honors dissertation in this lab than previous research I did. I wouldn't even call the other 2 labs research, I just read papers all day and did lab tech work.
In the end I realized research is not for me. Its not fun, and painstakingly slow to see results. Still learned a lot.
Is there anyone in this thread done a phd that late? Obviously my motivation is different now to study - to really learn the subject. I am financially self sufficient and will continue to be, and hence making a living out of my phd is not a consideration.
Edit: I would like to be able to study in a university setup (not distance education). Main reason is to soak in all the related conversations / workshops and also I like being in a young environment.
The hard part I imagine will be hanging out with mid-20-year olds everyday when I'll be in my late 40s already :)
I'm curious about this scenario from the advisor or departmental perspective: If a PhD candidate is financially self-sufficient, does this mean that a potential advisor has one less mouth to feed when competing for grants? How does the power dynamic change between advisor and financially independent PhD candidate? What if the PhD candidate is capable of funding a whole (or good portion) of a lab themselves - is there a conflict of interest somewhere there between donor vs. principal investigator vs. PhD-candidate roles? Do financially independent PhD candidates have more, less, or no competitive advantage during the selection and admissions process for an R1 institution?
So many questions...
You know how you've take a course before where the professor was just surprisingly awful at teaching? These professors are often some of the most knowledgeable people in a subfield of the subject you are taking, yet their teaching ability is severely lacking and you have to scramble to learn the material some other way (or just never learn it).
During a PhD, there is a decent chance that your adviser is similarly a bad manager. Unfortunately, having a bad manager for 5-7 years of your life can be a fairly awful experience. You will work with someone who you, on the one hand, look up to, but on the other hand, who seems to not care at all about your mental health, your possible career desires outside of academia, your work/life balance, or the exact reason why this week was a rough week for research in your (human) life.
I have a lot of other thoughts on the matter, but I thought I'd try to keep this post more concise =).
I wouldn't call that being a bad manager, but rather being an asshole.
As a professor, I often think that one of my biggest weaknesses is indeed management skills. After all, we suddenly find ourselves having to manage people without any training in the matter, and when our true call is typically science, not management.
But at least I'm not an asshole.
A PhD system trains you to think about unsolved problems in an given domain deeply with a larger time runway. The end goal is not a tangible product that reaches millions of people, but rather a set of ideas that can take a crack at the unsolved problems in your field in a novel way. A good work should inspire others in the field, and eventually a larger audience to pick them up and expand and build on top of it. To give a small example, a majority of the fundamentals of machine learning was charted out by many, many PhD works over the last 40 years. Implementing a linear classifier is 2 lines of code in 2018, but many Bothans died to bring us this information :-) .
The goals of industry are more immediate. Expect for a privileged few research labs in industry, your work is expected to be monetized, and rightly so. The goal is for you, if you run the business, else your management team to first figure out a problem of high relevance and monetary value. Build products/solutions for that problem, that can be used by someone who is less versed/ambivalent of your technical solutions. Efficacy of solving that particular problem often defines the merit of your contribution.
The fundamental of choosing the PhD or industry should be taking stock of what kind of contribution you want to make as an individual. If it is a few set of ideas to science, which on a later date might become something fundamental in our understanding of the world, then PhD is a good path. If it is a set of contributions towards a product/solution that eases the pain of many users then go into the industry first.
People should do PhD if they are genuinely interested in doing scientific research. If you are doing PhD under pressure or in hope of getting better paying jobs you will be dissopointed. It is an arduous process and taking up your precious years but it gives you opportunity to have freedom to explore and work on your interests for rest of your life. You won’t be coming to office everyday doing assigned task on your backlog and reporting your status in scrum meeting. Instead you will be reading about new creative work that was literally published yesterday, mulling over that in lunch with colleagues and apply your original ideas to actually get published under your name. The downside could be lower pay and/or no stock bonuses for many outside of hot areas like AI. But in general, you have much better chance of doing cutting edge work that you are truely passionate about if you have PhD in that area.
I would also mention teaching: some academic staff hate it because it takes time away from research. This is true but my experience is that if you put time into being a good teacher then you get to take your pick of the brightest students to help you with your research, which can lead to interesting Masters projects which lead on to PhDs.
I don't have a degree and I have that. Also I know a lot of people with a PhD who do not have anything remotely resembling freedom.
The author might be biased but it's good to hear it from that side once in awhile, rather than be deluged with "you can't do science without a PhD" and similar sentiments (I work parallel to academia and get plenty of shit from their ivory towerism, hence my comment).
In what way is the author "biased" exactly? She did this course, was in both a PHD and tech, and tells people why she thinks it's worth considering not getting a PHD, listing various (valid IMO) reasons.
Where does bias come in exactly?
Including former colleagues I probably know a couple of hundred PhDs who do this exactly the same as mere Bachelors. Outside of a few niches a PhD is barely recognised in industry. Sorry, but that’s the truth.
This condition is necessary, but not sufficient.
After finishing my PhD in quantum information, I turned to data science. I couldn’t be happier about this transition! Compared with academia, data science world looks to me like a wonderland.
tl;dr: faster pace, more freedom (sic!), way less bureaucracy and politics, etc
Longer version: https://p.migdal.pl/2015/12/14/sci-to-data-sci.html
"Career Guide for Engineers and Computer Scientists" by Philip Greenspun
http://philip.greenspun.com/careers/
Of course, you can work on cutting edge problems without a PhD but probably do it better with PhD. A PhD teaches a lot of intangible things like being comfortable with unknown problems , working patiently towards an end goal, get over the fear of failure etc. One thing it does not teach is how to earn money :).
In this context I would like to highlight unique cases where you do an "Industrial PhD" where you are working in a company and chose a relevant problem for a PhD. These have the "best of both worlds" and one is not bogged down with typical pressures in a regular PhD like those pointed by the author. That said, the final word is that a PhD is merely a conduit. Its more of what you chose to do with it than what it is , which matters at the end.
I'll try to give an answer and a response more generally to the material in the OP.
Background. I got a BS in pure math with nearly a second major in physics and worked in computing and applied math for problems in US national security around DC. Jobs were very easy to get; at one time my annual salary was 6 times what a new high end Camaro cost; much of the work was challenging for both the computing and the applied math; I was learning a lot of both computing, e.g., algorithms in Knuth, and applied math on the job and also especially math in independent study on evenings and weekends. Soon I got a call from a college friend to join FedEx. I used my background in computing, with a little applied math, to write some software, in six weeks, to schedule the fleet. The results pleased the BoD, enabled crucial funding, and saved the company. Later the BoD wanted some revenue projections. I formulated and solved
y'(t) = k y(t) (b - y(t))
for time t, revenue y(t) at time t, rate of growth y'(t) = d/dt y(t) at time t, and full revenue potential b. The BoD was pleased, and ..., to make a long story short, I saved the company again.
Then I got a Ph.D. in applied math from the engineering school of a famous, high end research university.
Now I'm doing an Internet startup, a Web site, basically a new and very different search engine -- for the, IMHO, very large part of search handled at best poorly by the existing Web sites and well known techniques.
For the startup, my applied math background is crucial: The crucial, enabling core of the startup is some original applied math I derived based on some advanced pure math prerequisites I got both in grad school and in independent study. I already knew enough computing except had to learn how to bring up a Web site that has been easy for the user interface but due to the core math somewhat tricky on the server side. I learned Microsoft's Visual Basic .NET (for the programming language -- I like it), ADO.NET (for the Web pages), and ASP.NET for the (relatively meager) use of SQL Server. The main difficulty was working through 5000+ Web pages of documentation. The first code is the first production code, 100,000 lines of typing, 24,000 programming language statements, and lots of documentation. The code seems to run as intended, but I need to add some data.
Some lessons:
(1) Math. IMHO, the key to powerful, valuable, new applications of computing is applied math. That is, if we accept that the big opportunity is to exploit and apply current computing, then we might notice that whatever we code to put out data users will like, as information, entertainment, whatever, is necessarily mathematically something, understood or not, powerful and valuable or not. So, in some sense, from 100,000 feet up, it should help to proceed mathematically, with possibly advanced prerequisites, some new results focused on the application in mind, and complete with theorems and proofs. Just IMHO. But I don't know anyone with a yacht over 100' long that did that; I suspect that very few people agree with me. Maybe what I am saying is a hopeless wild goose chase or a great green field opportunity -- you judge.
(2) Getting a Ph.D. In a nutshell, the three most important parts of getting a Ph.D., in no particular order, are research, research, and research. Yes, there can be courses, credits, grades, teaching assistant positions, weekly research seminars, qualifying exams, etc., but at a research university what can "cut through", dominate, and trump all or nearly all of that is research. The research should be publishable in a good peer-reviewed journal of original research; if there is any question, then send it in.
The criterion for a Ph.D. dissertation may be something like "An original contribution to knowledge worthy of publication." -- so, in case of some doubt, publish the thing.
The usual criteria for publication are that the work be (i) new, (ii) correct, and (iii) significant.
Now...
My wife was fatally injured in her Ph.D. program. The OP outlines a lot of just what happened to her. To have time to try to help her, for a while I took a slot as a B-school prof. It didn't work -- lost her anyway.
I never for even a milli, micro, nano, pico, femto second wanted to be a prof. Instead, I wanted to be solving problems in business, the money making kind.
(5) Non-Academic Career. Then I tried to get my career going again, outside academics. Bluntly, that didn't work very well.
I made a mistake: I should have returned to DC and gotten back into applied math and computing for US national security. I guessed that there would be opportunities as an employee in business; I was wrong.
Bluntly, my view is that US business and Ph.D. holders mix less well than oil and water.
Part of why:
(A) Business is still a lot like Ford in Henry's day: The manager knows more, and the subordinate is there to add muscle to the work of the manager. A manager has no use for a subordinate who knows much and resents or feels threatened by such a person.
Supposedly lawyers have a solution: A working level lawyer should work only for a lawyer. Period.
Well, a working level Ph.D. should work only for another Ph.D., and that criterion would eliminate nearly all jobs for a Ph.D. in business.
Not even a CEO wants a Ph.D. around except maybe tucked away in some side organization, out of the main work of the business. E.g., the CEO is plenty sure that he is the only really important person in the company and, thus, certainly doesn't need a Ph.D. or some academic background he (the CEO) doesn't have!
(B) Business regards Ph.D. holders as blue sky dreamers out in the ozone who refuse to contribute to the business, who really want to publish a lot of papers and get a prof slot in academics.
(C) If a Ph.D. person does anything original relevant to anything in business, usually the business will regard this person as a threat.
(D) Suppose a Ph.D. takes on a practical business problem:
(i) If the Ph.D. successfully uses their advanced knowledge to get a good solution, e.g., one that makes a lot of money for the business, then everyone else in the business, even the CEO and the BoD, will feel threatened and/or jealous.
(ii) If the Ph.D. fails to get a good solution, then everyone else will take the opportunity to denigrate both the person and the Ph.D. degree -- "I always thought that a Ph.D. was just a useless, hopeless, worthless impractical dreamer out in the ozone, and now we know for sure.".
(6) A Ph.D. in a business research division. Yes, some businesses, say, ones with some loose cash, might set up a research division, hire a Ph.D. as the director, and hope for something good. If nothing good happens, well, the company could afford the wasted money.
Generally, connections about the actual business between the research division and the rest of the company are more awkward than a skunk at a Victorian garden party. The rest of the company doesn't want to be bothered, sees various threats, etc.
Here are some of the reasons for such a research division:
(A) Luster. Use the research division to impress the public, for good PR, to impress customers, to cover the rear exhaust port of both the CEO and the BoD, etc.
(B) As a patent shop. So, the research division can develop a patent portfolio, maybe dozens, hundreds, thousands of patents. Then some specialized lawyers can use that patent portfolio as a, call it, battering ram against any would be competitors. There can be cross licensing deals, revenue, etc.
(7) Career direction. It's your career. In this career, there will necessarily be some directions you will be pursuing. Some directions are good; most are not. It's up to you, and maybe your family, closest, trusted friends, etc. to pick, at least try to pick, a good direction(s).
If you just look for a job, get some offers, and take the best offer, then likely you will be following the direction ...
My guess is that there are a lot of Ph.D. electronics engineers over 40 who are essentially unemployable at anything close to their past who want an electrician's license so they can install AC wiring in houses.
Maybe the best job they can get is to be a clerk in the electronics department at Wal-Mart.
Imagine, two guys just out of high school. They both mowed lawns as teenagers. Joe wants to continue that as a business, at first still living at home. Tom goes to college, continues, gets an electronics Ph.D., and gets a job in an electronics company.
Then they are both 35: Joe now has five lawn and garden crews, each with a late model, crew cab truck, a $5000 trailer with 4 riding mowers each worth $15,000. He has clients in upper class residential areas and small to medium commercial lots. He does weeding, soil testing, fertilizer applications, mowing, edging, shrubbery trimming, landscaping, etc. Tom's employer has a policy: By age 35, promoted into management or fired. Tom gets fired.
Joe is now much better off than Tom. Tom would do well to look for a job with Joe and rise to managing one of the crews, planning and marketing for higher end parts of the business, etc.
This is not a joke. Besides, it's not funny.
So true - now most students are happy with, "I came into the office every [most] days and did [the thing]. I wrote it up so where's my PhD."
They don't evaluate their own data critically for interesting information - they did the experiment or simulation as told and they think this is their job.
I'd really appreciate if anyone reading this comment have something to suggest. Thanks a lot HN community, you've been a great support.
I am in a similar position as you. My PhD. topic is focused on the same research area as my job in tech company so I can do both together.
If it is not possible I would advise not pursuing PhD. and continue self-learning. The opportunity cost of PhD. is not worth it in my opinion.
People in my country usually start PhD. at 25 and take at least 6 years to finish, because the universities use them as cheap workforce and aren't incentivised to allow students graduate quickly.
uh ? which country are you in ? In france like you said you have to do a master before BUT a phd is only 3 years. I'm 26 and I finished my phd - had my defense earlier this year when I was still 25.
It was wrong from me to generalise to whole Europe. I didn't want to be too specific in my comment.
On the other hand, the upside is that we do not have to pay a single dollar for education here as opposed to US.
In his particular institute, the better you were, the LONGER it often took because the professors found stuff they wanted you to help with unless you were really good at boundaries. My husband took 7 years; a very kind, very bright friend of his needed 8.5.
Now, students must submit their thesis within five years (one of my students very nearly failed his PhD because of this), and there are funding restrictions for continuing beyond three.
At least I knew this when I started. And it motivated me to get finished in 3 years.
Someone did give me the advice in this article before I started. Which was good. I wasn't in it for the money. It was the freedom to work on things that interested me that mattered.
What no-one told me later, when I started a post-doc, was that the chances of getting a full-time academic job were close to zero... now that did feel like 3 years wasted.
I have found this to be good advice for people who are thinking of starting companies too.
However, I can't help but think that this mentality does scare away certain personality types that would otherwise make for great researchers. I fall into the category of people who ignored advice and did it anyway, but I try to be welcoming to other types too.
It depends of course in hwo the "no" is worded.