One point jumped out at me, which I haven't seen anyone respond to:
> We are concerned about untrained amateurs going directly into trying to solve very difficult global problems. They can then cause harm overall, by lowering the average quality of analysis or launching ill-considered projects due to a lack of experience or understanding. A PhD reduces the risk you’ll accidentally do this.
This point is operating off of a different fundamental paradigm than the comments here that are talking about opportunity cost. Opportunity cost is looking at personal benefit, while the above is looking at the effectiveness at solving hard, large-scale problems.
Obviously, money is important, but for many (often those who are fortunate enough to have reached a certain income level), it isn't the only thing or even the primary thing.
I don't have much insight into how true the point actually is, but I found it to be interesting.
A PhD won't make you smart, but does make you read the literature. You could do that on the job, but who would? ...and who would pay you to? Real world experience won't give you even the breadth of knowledge (unless you're working with all the world-leading experts) - let alone feeling the value of the literature, and the skills to access it.
This helps you avoid well-known stupid mistakes, and concentrate your energies on all-new stupid mistakes.
It's not an unreasonable opinion, but I doubt a PhD can make up for lack of real-world experience. Likely, the best way to get better at solving global problems is by doing just that for years, preferably alongside others with more experience.
I think it depends on the PHD and the real world experience.
There's a quote I kinda forgot, but it goes something like:
"Did you experience the same thing 10 times, or did you experience 10 years worth?"
And I think this is true for a lot of real world experience, you don't necessarily have 10 year of it even though you've worked 10 years, you have 10 times the same experience.
Political science isn't, at all, mindless, and there are focus areas within it that are extremely rigorous either elitist or analytically or both.
OTOH, there's not a lot of places where a PoliSci Ph.D. would substitute for industry experience (or, for that matter, vice versa.) If you want to substitute an academic degree in that general space for “industry” experience, you go for public policy and/or public administration (or, arguably, law), not political science.
Which is funny, because the lack of government <-> academic information transfer is so small that any PoliSci PhD has an immediate competitive advantage in almost any bureaucratic setting (this is true at the federal and local levels).
A PhD in government pretty much guarantees that you'll be retiring after a thirty year career as a GS-15 or SES. A 150k salary plus benefits may not be high for some STEM fields, but it's definitely not something to look down on.
I've often struggled with what "real world experience" is and how a PhD isn't that. A PhD in science is dominated by doing research every day. PhDs often go into industry positions post phd because of their research experience. I think real world experience ends up just meaning "experience with our specific hardware/software/stack/whatever" and that isn't really anything special at that point.
I've been in the industry for 12 years and I've been deeply considering going back to school for a PhD. I love working on fundamental CS and I'd greatly enjoy the sort of work that I'd have access to during and after a PhD. The challenge is that, for folks who haven't won the equity lottery or racked up FAANG RSUs, it's a huge financial challenge to consider going from an engineering salary to PhD stipend. Considering I still have student loans from my BS, and a spouse who doesn't work that is also being supported on my salary, it's almost impossible to figure out the logistics of working on a PhD full time.
I've found a few programs that are open to considering part-time candidates, but they are few and far between, and most of the ones who consider it still strongly recommend against it.
Have you heard of http://www.mrmoneymustache.com ? If you still have student loans after 12 years in the industry, you might want to re-examine your spending/saving patterns
Just discovered this website, this philosophy looks really interesting!
The ultimate goal of early retirement is appealing, so the discipline has a purpose. This might get me somewhere, instead of saving for the sake of getting richer.
Thank you for this discovery!
It’s not the easiest decision. Regarding loans, repayment should stop once you go back. Regarding income, it’s hard at first. You might have to downsize, but also research cost of living. Depending on where you go to, it’s doable. Some universities have housing for people doing their PhDs and for families which is cheaper.
Went (going?) through it. My wife is doing her PhD. From 2 really good incomes to one stipend. We don’t live in campus housing so it’s more expensive (savings helped here initially), but we do know of others there with their spouses and some even with kids that are doing good. I found a job after getting here which helped.
I'm trying it out now, I guess meet you back at this thread in another 10 years?
FYI for others wondering about the same, I wouldn't wait the full 10 years unless you are going to study something very industry adjacent like software engineering methodology. I've had to spend a lot of time relearning anything theoretical.
Not to be argumentative, but there is very little research that I've found on software engineering methodology that is in any way convincing. If you've got a good source, I would be incredibly grateful if you shared it! This is one of those areas where I wonder if a PhD is in any way useful. One the one hand in industry you have a complete lack of rigorous controls to measure your effectiveness and on the other hand in academia there is no way to fund the kind of project that would give you decent data (if you could somehow even figure out a way to introduce rigorous controls). I've thought about this quite a bit, and I think you're better off on the industry side if you are really interested in pursuing that kind of thing. Would love to be proven wrong, though.
I don't think you're wrong, I've seen a few interesting papers that nibble around the edges, usually examining how bugs get introduced and when they are fixed. I mentioned the field as an example of where someone coming from industry would not have to catch up to someone whose knowledge of Turing machines was still very fresh.
You're right, no trials on methodology that resemble something like clinical trials. And of course the reason for that is it would be incredibly expensive to even run a reasonable pilot (two small teams head to head implementing a realistic system), and if any organization has actually gathered data, it would be the crown jewel of proprietary knowledge. Not something you would ever open source.
The only thing I can say is that an immature field is one with opportunities left. If you can figure out how to jump the hurdles you can make a huge impact.
How is this black and white thinking? The sentence was "A PhD reduces the risk you’ll accidentally do this", not "a PhD will prevent this" or "you're an amateur if you only have a BS". It's taking about risks and likelihoods. That's not black and white thinking, that's pretty much the exact opposite.
You do know that most group supervisors have graduate degrees right? (At least at JPL, not sure about other NASA centers). Seniority tends to be based on your highest level of education (obviously there are exceptions - but exceptions are exceptions because they're rare, not the norm).
There's no "single PhDless engineer" who engineers a whole spacecraft.
Imagine a hospital where a new initiative is taken. Nurses follow doctors during their daily patient rounds for 6 months. Then the nurses are expected to do the rounds instead of the doctors. No doctors are left on the premises.
Its not very clear cut, there are some blurry boundaries, but still, would you want people to use that hospital?
That's an example that hits close to home as 2 close family members are Nurse Practitioners. This is a flawed example because Registered Nurses do not do the same job as a physician and are legally not allowed to learn how to do it. What if the nurses were allowed to shadow the physicians for 4 years? That would be called residency.
What if you have Nurse Practitioners with a Nursing BS, 4 years of nursing experience, a Masters degree in NP, and 4 years of experience diagnosing, managing complex conditions, etc.?
I personally know of several examples of Nurse Practitioners who are a drop-in replacement for an MD and that was achieved through work experience.
Not really. As an untrained amateur myself hoping to create great solutions. Thing is its the exception, not the rule.
Also with highly complex things (large scale - global problems) often it is possible for an untrained amateur to create a solution that is easy to communicate why it should be good, and it works in the short term. However awhile later, it is found that long term failure occurs
Yes, but the argument is that these untrained amateurs more often make grave errors from ignorance that make existing situations worse, or at best divert resources from more well-considered approaches.
Which produces net better global outcomes I couldn’t say, though I have my own intuitions.
In certain context - empty frontiers vs discovered/cultivated regions
Take the Internet -no one had any models for it. You could come up with new innovations and then disrupt other industries or make entirely new frontiers.
But in denser more studied areas - advances are incremental. People don’t come up with magic hashing functions for example.
Or new engine designs - there’s only so much you can do without breaking emissions norms or fuel efficiency or
end up having engines which end up getting damaged.
BUT- the dream is always to find that missing or unthought of idea which changes the world.
It’s just relatively easier with an open frontier.
> PhDs are intellectually demanding, and not everyone will be able to finish. In the US between 41 and 78% of people who start PhDs have finished them after ten years, depending on their sub field. Computer and Information Science has the lowest graduation rates (41%)
Right...
I've met a good number of PhD dropouts - I am one myself. None of them dropped out because it was too "intellectually demanding." They all dropped out because the work was unrewarding, the odds of employment in academia were poor, and the work in academia was unappealing.
Looking at Computer and Information Science in particular, I might suggest that high availability of high paying jobs in the private sector is worth looking at rather than the "intellectually demanding" academic program.
I also dropped out of a PhD program in CS after I realized that my advisor had incentives to keep me in the program and no incentives to help me complete my dissertation in a timely fashion. Imagine having a highly-educated, native english speaking computer scientist that you paid $30K / year. Compared to the 150K+ / year that person would make in the private sector.
I went to grad school for chemistry. A PhD in my experience is more an exercise in perseverance than anything. You don’t need to be a genius, just work hard and have some luck as well.
While I suspect you're right, anecdotal evidence must be fairly weak here. Not many people would admit to themselves that they weren't up to the intellectual challenge -- let alone admit it to other people.
My own experience was that part of the problem was that there was no intellectual challenge. I actually suspect that if PhD programs were a little more intellectually challenging drop-out rates would go down.
I think the context that most commenters are missing here is that this is not an argument that you will make more money with a PhD.
The focus of 80,000 hours is on "effective altruism" where the goal is to have the greatest positive impact on the world.
And it's not super hard to believe that there are quite a few jobs that require a PhD, often due to plain credentialism, that have significantly more impact than your average software developer at a well paying company.
I'm still not going to do a PhD, but I'm really not the target audience here, and it's sort of not surprising since 80,000 hours is an outgrowth of academia itself.
I was expecting reasons why now is a better climate to apply then in previous years. That would be interesting, but instead it's a list of generic reasons to apply for a phd. Yes, I know that phd's can open doors.
I read the whole article, including the "why not to get a PhD part." I still thought it was way too kind to the "get a PhD" side of the equation. PhDs represent a tremendous opportunity cost, and I have not seen good evidence that they causally improve outcomes in the vast majority of cases.
This may shock you, but some people prefer doing research to working a normal job and earning lots of money. Those are the types of people that should and will do PhDs (there are unfortunately a lot who think like you and still try and do a PhD for god knows what reason).
You don't need a PhD to do research. All information is free now. It's fully possible for individuals to self teach in a working environment and be capable of performing effective research with 2-4 years of good experience after undergrad. To assume that the only way to accomplish this is via PhD is silly. This is especially the case when talking about computer science fields, where the barrier to entry is basically owning a cell phone.
Hypothetically speaking you're right. You can teach yourself everything you need to know even if that means sitting down and going through MIT OCW with all the textbooks in the syllabi.
But practically speaking, in the real world we reside in, approximately no one authors peer reviewed research without at least being in graduate school. You frequently see "talks" at conferences given by non-PhDs, but it's exceptionally rare to find math or computer science research published in reputable journals or conferences by non-PhDs.
The dirty secret is that the formal barrier to entry was never really the problem. It's extraordinarily difficult to learn how to conduct meaningful research. The structure of a PhD is imperfect, but it places you in direct collaboration with an experienced researcher who can mentor you. Half of the entire work of a PhD program is precisely learning how to learn on your own so you can find interesting research questions.
I don't think it's productive to talk about what's merely possible. A more realistic barometer is what's plausible. Anyone who can bootstrap themselves up to original research in math or computer science through independent study is extremely gifted. You don't want to set a goal that requires you to be extremely gifted right from the start.
Effective research is not always published. There are volumes of private research in technical disciplines you will never hear about. There is not enough data available for you to declare confidently that the PhD is a strong requirement for real research. Furthermore, especially in the field of computer science, you should really run the numbers on how many of ex. the 10 authors on each of the 100 Google papers put out each year had a PhD.
Additionally, nothing has been done to separate the issue of correlation and causation. If research capable individuals pursue a PhD because it interests them, who's to say that their research capability came from the PhD versus it being intrinsic.
One can learn a lot from a PhD. But saying that it's a practical requirement for research in this modern era of information access is without definitive evidence.
Sure you can learn a lot from just the literature but in my experience there is a set of heuristics and common understandings/techniques that are often passed on through instructors or peers you're collaborating with that are absent in textbooks. More importantly, points of emphasis and exigent problems in an era of information deluge often come through only when in regular contact with practitioners and attendance of conferences.
These are thoughts that have crystallized for me over a 4 year period of starting math late in undergrad, struggling through and putting a lot of time into non essential aspects in the process of self study to get up to speed, hitting a road block when attempting to self study graduate level material but ultimately managing upon working in an academic setting.
Not everyone is a certified genius like Grigori Perelman who can do great research while living in a vacuum. You need great advisors who will guide you, and fellow grad students who you can ask for help and bounce ideas off. The academic enviroment is invaluable for someone doing their PhD, and it's pretty much impossible for someone with somewhat above average intelligence (and let's face it, most PhD's are not geniuses) to do good research with just a computer.
The "all information is free and you can study anything using just a computer" is a stupid meme that doesn't work for the vast majority of people.
What is that strawman exactly? You are saying that anyone can do PhD level research with just a laptop and the internet. I am saying that that is extremely difficult and impossible for the vast majority of people.
Maybe I’m misreading, but the strawman seems to be that research can be done in a vacuum without collaborating.
The parent commenter seems to be suggesting that a PhD isn’t a requirement to work successfully on interesting research problems. It’s possible to go online and read the same papers everyone’s reading. You can still talk to people to get guidance and find out known things that are not published.
Ya I guess it was something of a strawman. But it's pretty close to the truth since it's very hard to get good guidance outside academia or places like google with lots of phds around.
>It's fully possible for individuals to self teach in a working environment and be capable of performing effective research with 2-4 years of good experience after undergrad.
It's technically possible, but it's unrealistic. The work that is published in journals often demonstrates results that are known by the community years before the actual publication. There's effectively a "secret club" in many disciplines that you need to have connections to in order to actually understand the current state-of-the-art. Without those connections, it's hard to understand what is high-impact and meaningful. This means the research you would do is likely to be insignificant, or already done and pending publication.
An easy way to avoid this is to start attending conferences and working with establishment players, which is exactly what doing a PhD entails.
It doesn't shock me, but I think a lot of people don't realize what they're giving up until it's too late, and I think a lot of people are deceiving themselves about exactly what they're getting into.
They don’t make sense from a cost/benefit outside of the people actually able to find work in that one specific area and beat out all competition. There are simply more candidates than high paying phd jobs. A masters tends to have a more dependable ROI. But all things considered, I say modern academia is a shadow of what it claims to be.
Of the people I know who have or are finishing PhDs, about half are happy with the choice. The commonality seems to be that the happy ones really love their chosen field - they'd rather play and work on whatever topic than anything else. One is making absurd gobs of money, but she also really loves her niche.
Another, less universal aspect is that many folks I know who went for one have attenuated interest in typically 'normal' life-goals - one or more of things like marriage, kids, 2.4 picket fences, etc. are either not part of the plan or substantially less important to them than to other people. (Though this may be bias introduced by me, because I also fit that description, and family-less folks over a certain age tend to find each other.)
Salary, salary regressed on all factors of things (age,sex,years in industry, etc). Once the equation is properly balanced the marginal impact of a PhD is less than the marginal impact of just those added years in the field.
And don’t forget to factor in the opportunity cost of accruing student loan debt (which can’t be defaulted!) against lost yearly income.
My PhD and the research I do now based on it give me a huge amount of personal satisfaction and are a big part of my quality of life. I'm comparing the work I did before my PhD (because I worked after my degree before a PhD) and seeing what my friends think about work compared to me.
Yeah if you're looking for more money, maybe don't do a PhD. If you're looking for the time and space to find satisfaction by working on a unique problem, then it's hard to find something that gives you that in our industry to the same degree.
So that’s the pitch they make in the PhD programs. But what they leave out are the negative associations with a PhD. You will immediately be over qualified for large swaths of work. And unless you can immediately land that dream job by, amongst other things, beating out all competition for that work - you will enter the work force with 0 applicable work experience. Hiring managers can be downright blind about anything that isn’t direct work experience in something you’d only gain experience from already having experience in that labor market.
> And unless you can immediately land that dream job
I'm not sure it has to be immediately, but your point still stands, and it's often ignored by proponents, which is that merely acquiring a graduate degree does not automatically guarantee all those intangible benefits.
One has to actually secure employment with those benefits. Whether that's easier than finding employment with salary benefits instead (or in addition) seems to be still an open question.
Life is short and the basics of survival require daily purchases. So while someone can take a cruddy job while still looking for a better job, all that time does add up to the opportunity cost of going for a PhD. Agree, it’s very much an open question.
Personally, I'm already happily employed and have just started grad school (masters). My motivation, aside from being more competitive in the job market (I don't want to stay in this job forever), is that I enjoy challenges. But I prefer to meet challenges at least somewhat prepared to take them on. I see grad school as an opportunity to study my field in more depth than I realistically have time or motivation for on the job or after work. I figure this will allow me sort of level up and be able to take on more challenging work in the future.
Nothing in the article about what one really learns in a Ph.D. program. Here are the things you only learn in a Ph.D. program, more or less in chronological order:
1. What kind of information you will find in each section of a research paper in your field.
2. What are the major research programs in your field and where they disagree.
3. How to identify where a paper's author stands with respect to the major disagreements in your field.
4. How to decide that a paper isn't worth your time without reading all of it.
5. How do you write a research paper?
6. How do you decide what research to do? (This decision will often be made for you for the first couple of years.)
7. Where you stand with respect to the major disagreements in your field.
8. How to synthesize half a decade's work into a realtively brief document.
Notably, these skills don't help you do any of the following:
1. Manage a research group in an academic setting.
2. Teach undergraduates.
3. Manage people in industry.
You might pick up applicable skills along the way, but these are incidental. Any specific skill, any particular knowledge, can be obtained without getting a Ph.D. The Ph.D. teaches what you need to know to advance the field.
I'd like to return for a second to talk about teaching undergraduates. It's a very important thing to do. Ph.D. training ought to include pedagogy. Standards for university instruction are shockingly low, and professors are rarely held to account for failing to teach the material well, especially at R1 institutions. However this comment is not about what you should learn, but about what you'll actually learn in a Ph.D. program if you decide to enroll tomorrow.
I agree with your main point here, but I think you're trying to paint and overly broad claim about PhD programs, which are extremely different between institutions. It's like saying "working at a company is like this... you learn this but not that."
Specifically,
1. Senior PhD students in many research groups can become de facto managers of the group, especially when the professor is on sabbatical. PhD students are often involved in the funding, presentations, recruitment, and represent the group. Some PIs/professors empower their PhD students to run a mini research group within their group.
2. I've seen some PhD programs have a strong emphasis on teaching undergraduates, both as in teaching undergraduate courses directly (which many PhD students in the humanities and social science already do), serving as teaching assistants with a substantive educational role (not just grading), and mentoring them in research or more general advising. For example, Brown University and Princeton are known to have a good research-teaching balance for PhD students. Brown has a large number of programs that help PhD students get more involved in undergraduate teaching [https://www.brown.edu/sheridan/programs-services/certificate...]
1-4 can be learned by anyone willing to read enough papers in a great, great many fields. The math needed to understand the articles in the Quarterly Journal of Economics or the American Political Science review is in most cases well within reach for the population of potential applicants for a doctorate.
There’s an enormous amount of valuable non-procedural knowledge that most doctoral degree holders have and most non-holders don’t but the average doctor, lawyer or senior civil servant does research that would be impossible without skills 1-4 without ever doing a Ph.D.
Huge opportunity cost for 5-8 years, think carefully. Also, very very few jobs in academia so don't plan on those. Was interesting and learned a lot, not sure the most useful use of those years.
indeed. a few folks in the FIRE (financial independence / retire early) space talk about being able to save enough money to be retired or semi-retired from 7-8 years of full-time work (this doesn't require FAANG scale salaries, just a combination of working somewhere that pays pretty well, not spending much money, building up investments, and luck to avoid getting clobbered with huge medical expenses. and the luck to start from a context where this might be an option for you at all)
edit: adding a few links to reading material
One of the main things is to focus on maximising your savings rate, i.e. your amount of savings divided by your revenue (for most people this would be their income from their job, minus income tax). You can increase this by making lifestyle changes that let you sustainably decrease your expenses, and by getting a job that pays you more, or other sources of income.
William Bernstein has written a bit about investment and asset allocation. He has a free publication titled "If You Can" geared at young people to help them plan for retirement: http://efficientfrontier.com/ef/0adhoc/2books.htm One of his general pieces of advice is to invest 15% of income every year into low-fee investment funds, i.e. to achieve a savings rate of 15%. If you do this consistently for around 30 or 40 years, you can retire.
Things get much more extreme for higher savings rates. For example, if you hit a savings rate of around 75% year after year (saving three quarters of what you earn after tax) and invest this, then with some assumptions about rates of return on investment, after about 8 years there should be enough income from investments to cover your expenses:
You need to be a bit careful re: assumptions about stock market returns over the next few years or decades, some markets (e.g. the US stock market) are priced very highly compared to historical prices, so assumptions about rates of return from investment that seemed reasonable a decade ago are probably over-estimates now.
Looks like author is focusing on grad school as in PhD only. PhD highly depends on the field. For example in hardcore sciences, to accelerate career PhD is often helpful. However, in applied fields it depends on industry to industry.
I was doing my masters with focus on Human-Computer Interaction (HCI). I was considering PhD but after talking with academic advisor, I realized having PhD will limit my options to either research labs or academia, for which jobs are few and far between. For industry jobs, I felt I didn't need PhD. Yes, you might be few steps below in the ladder but you can quickly catch-up as you climb.
So everyone should seriously consider where they want to end up. If research is something they fancy and their field has ample research opportunities or can be leg up in the industry with PhD then PhD is a great option given the opportunity costs.
I'm very interested in (applied) AI research positions but wasn't quite prepared to commit 5-8 years to a PhD program. In order to validate my interest in research, I've enrolled in a two year Computer Science Master's program at Tsinghua University. This program has a thesis component, allowing me to participate in research and potentially get a publication. Depending on how much I enjoy the content, I plan on applying to AI fellowships, a PhD program, or Machine Learning Engineering positions after graduation.
Thought the program hasn't started yet, I hope that this is an alternative path for those interested in research but not quite ready for a full PhD.
This morning I listened to the Freakonomics podcast from July 18th about “Why we choke”.
About 22 minutes in, a researcher is discussing a study he did where they paid participants a months wage for a single task to see if they would perform better. They didn’t. They actually performed worse! A follow up study at MIT found this to be particularly true for cognitive tasks.
Would this apply to a job over time instead of a single task? Probably not, but I’m also not convinced that salary will increase productivity after some relatively low amount.
I think there's "productivity" from two standpoints. In the worker's eyes, she's more productive if she's paid more, by definition. In the employer's eyes, she's more productive if she produces more for the company, by definition.
A classic "efficient" market should close any gaps between these two definition, but it doesn't work perfectly in practice.
What do you study and how does a citizen with practical experience in the field help? A lot of things I have read about software engineering isn't really well tested. How does your research stack up? Is that something people on HN could help improve on?
I assume this refers to a PHD. The masters degree I got is also post-grad.
I'm not sure if anyone else covered it but you have to have a certain amount of money to get a PHD. I was approved to enter a doctorate program at Stanford. I couldn't afford it so I went to HP to earn some money. After the five-year approval expired I was disappointed that I could never afford it. (We're talking astronomical numbers here).
When I was a student 30 years ago, the rule of thumb was: Don't get a graduate degree unless someone else is paying for it. In my field (physics), a first or second tier school would provide free tuition plus a stipend if they expected to get any students.
Don't know how things are now, or how they are in other fields.
Now, of course free wasn't exactly free, because studying in grad school was potentially at the expense of working a regular job. That's another topic.
Hmm. When I told the professor trying to "hire" me that I couldn't afford it there was no mention of a stipend. Could it have been different in the 1970s?
I'm currently doing a Ph.D. It's standard practice for nearly all PhD programs for the advisor to support a minimal stipend ($15-40k or so depending on the field and university) which goes directly to me, plus university tuition (up to $50k/yr depending on the university) which my lab directly pays my instutition. I only see the tuition bill and that it's paid for me.
Notably, this typically isn't the case for Masters programs, which are cash cows for universities like Stanford and charge $40k/yr+ for students.
Like the other commenter said, any Ph.D. program that isn't covering your tuition and providing a reasonable stipend has huge red flags all over it. They aren't the norm, either.
I'm a MechE so this is probably different compared to software. A masters is far better in terms of ROI compared to PhD. I seriously contemplated to complete my PhD after my masters but an additional ~4 years of school was not worth it financially. Your wage is not much different (though the type of work and responsibility might be different) and sometimes masters + 4-5 years actually make more than a PhD grad. Also, you make a full wage 4-5 year earlier. That meant a lot for me in terms of starting my own family and such.
I can't wait until you can make new strains of E. coli on your desktop/iPad. Time in grad school is often spent cloning. It would be really great if there were more community-accessible Core facilities available on a subscription service.
I know this isn't popular. But I will say it - since I know it goes unspoken in the industry. When I see a CS MS degree I assume the person is terrible. I am always proved correct. I can count on one hand resumes of people that gets a CS BS degree THEN a MS degree in CS [15 years of looking at resumes]. In the US, if you get a BS in CS you usually take off doing cool stuff. CS Master degrees are full of people not majoring in CS for the right reason [eg, never played with computers as a kid]. If computers were your passion you would have got a BS in it.
I mean, if they're coming from a different major, it's the equivalent of hiring people after sophomore year. If they haven't been programming outside that, that frankly isn't enough time to be a good programmer.
This is discouraging. As someone with a BS in something other than CS I don't feel that I get the benefit of the doubt during interviews. My CS fundamentals are weak because I never learned them and they very rarely come up in building things and I have been thinking of going back for a CS Masters
Pick up Discrete Mathematics by Knuth, Foundations of Mathematics by Sibley, Modern Operating Systems by Tanenbaum, and Computer Organization and Design by Patterson and Hennessy.
You internalize those, combined with your experience, and you'll be better off than a lot of CS grads.
Thanks! Currently working through The Algorithm Design Manual, and next on my list is Introduction to Algorithms but I will add your recommendations to my reading list
It's interesting how different things are here. In Germany it's nearly expected to get a Masters degree because of the way things used to be: until not too long ago we didn't have BS/MS but only a diploma which was equivalent to a Masters degree so many companies still expect a Masters degree even nowadays. From what I've seen it also seems more difficult to get a MS in a different field than your BS here. Is it common in the US?
From what I remember of my time in Germany it’s not nearly expected to get a Master’s, it’s expected. Getting a Bachelor’s is still the norm in the UK or US, getting a Master’s is the norm in Germany or Austria.
Getting a Master’s in a different field from your Bachrlor’s is so common as to be unremarkable. The ones which are only available to those with significant prior study in the field are the exception, not the rule.
I've seen it among many managers from several companies. After I got involved in the hiring process I saw why. Like unable to fizzbuzz more than half the time.
By the time I finish my M Eng I will have a cutting edge education in machine learning, robotics, computer vision, and natural language processing, on top of five solid years of paid programming experience.
If someone insists on making a hiring decision based on my ability to whip out fizzbuzz on a whiteboard under a ticking clock, find another person and waste their time. I’ll pass.
Interviewing an experienced and credentialed developer like they’re an undergrad sophomore with a one sentence resume looking for a summer internship shows to me that if an employer is going to waste the time of a candidate (and implicitly insult their intelligence and education), then they will waste the time of an employee, and such they represent an organization I want no part of.
Except it isn't a waste of time. Most of the credentialed, experienced people looking for a job in this current market can't do it. Most of the people actively looking for work in this field don't have a job for a reason.
Have you ever been a part of the hiring process, or are just stating that it's a waste of time because you think it's somehow beneath you to prove that you can do basic work? You know those annoying people in you group projects who don't do any work, but somehow squeak by? They graduate to, and get a job to.
University of Oxford Master’s in Software Engineering
... However, more extensive experience may compensate for a lack of formal qualifications, and a strong, immediately-relevant qualification may compensate for a lack of professional experience.
I think that's the entire point: Gates has honorary doctorates without ever getting a bachelor's. So, yeah, some people have gotten a doctorate while skipping the BS (if we assume BS means Bachelor of Science).
Interesting contrast to the recent wave of posts about how you should absolutely not go for academia in the current market of having 1 job spot for every 3 PhDs. What's one to make of this?
80,000 Hours is a website made by members of a very particular subculture. The “prototypical” PhD they’re imagining is in computer science, probably with a focus on machine learning, probably at a highly-ranked school, by someone who went to a highly-ranked undergrad.
Job prospects are relatively good. Pay is bad but industry internships during summer can compensate. Academic jobs are not quite as scarce, and R&D jobs in industry are plentiful. There is not an air of failure or taboo about moving on from academia after the degree.
The calculation on a PhD is very different compared to most other fields, and certainly compared to humanities PhDs (which actually charge money to attend!)
Don't get a PhD because you think it will be good for random job opportunities. Get a PhD because you really want to spend a lot of time researching X very deeply.
From that article: "Why am I (a tenured professor of physics) trying to discourage you from following a career path which was successful for me? Because times have changed (I received my Ph.D. in 1973, and tenure in 1976). American science no longer offers a reasonable career path. If you go to graduate school in science it is in the expectation of spending your working life doing scientific research, using your ingenuity and curiosity to solve important and interesting problems. You will almost certainly be disappointed, probably when it is too late to choose another career."
A previous comment I made with quotes from those links four months ago with various replies (then the topics was "Time to talk about why so many postgrads have poor mental health"): https://news.ycombinator.com/item?id=16725676
For all that, there are many good STEM-related jobs which seem to require a PhD these days, so obviously there are nuances to the topic and whether a PhD is worth it. If you are getting a PhD and someone else is paying for it (current job or a fellowship) and you already know and like the professor who you will be working for, and you already know of a good industry job prospect (in five to seven years, so hard to predict), and you like academic work, then it may not be quite as financially risky as otherwise (although it still may be emotionally draining). Also, if you are independently wealthy and do a PhD for fun, then again a PhD can make sense.
For people in the USA thinking about getting a PhD, consider schools in Western Europe as they tend to be more research-oriented with less-to-no classroom busywork.
It's sad things have come to this in a country with as much wealth as the USA -- especially given how much advanced research can (in theory) potentially improve people's lives if you are allowed to research such things.
Bottom line: as people have said here, a PhD represents a tremendous opportunity cost with a significant risk of having nothing to show for it.
I'd like to. Do graduate programs accept devs who have been working the last decade?
There are areas I feel I could be helpful with, e.g. creating reproducible experiments or ensuring a codebase could be open sourced in relation to a paper. It's a lot of grunt work that nobody really has time to do.
If that were possible, I'm not quite sure where to start with executing that plan. (Chicago area, FWIW.)
I left another career (military, rather than tech though) to go back and do a PhD. I didn't really find any barriers to it - I just applied like anyone else. I basically cold-emailed some professors doing things I was interested in and went from there.
> There are areas I feel I could be helpful with, e.g. creating reproducible experiments or ensuring a codebase could be open sourced in relation to a paper. It's a lot of grunt work that nobody really has time to do.
Whoah caution there. Don't go into a PhD with the idea that you can be helpful with someone else's work, or doing grunt work. Doing a PhD is very personal - you need your own big idea to work on. You may be part of a lab, your work may even be one part on a larger project, but you must carve out your own very distinct part of that and generate your own research. You can't really do a PhD by helping along with something else. Open sourcing a code base is not a research contribution.
You could make a PhD about how to make science more reproducible, though. You'd need some big idea about how to do that.
It sounds like you might be interested in looking at being a research engineer as well. Prestigious places like Microsoft Research hire quite a few of these I believe.
Yes. I did this a few years ago after working for ~9 years. The trick is getting good letters of recommendation that far out. In my case, I got letters from supervisors/colleagues and applied to Masters programs, then used my Masters as a springboard to a PhD. It's been great, I would highly recommend it.
I'm in the same boat - The PhD admissions proces over-indexes on Letters of Recommendation from other academics. I don't think I can get those because I have not actively stayed in touch with my professors.
I went through the existing process once and promised myself I would never waste that much time again. But if there's a program that genuinely values industry experience, and is willing to look past recommendation letters from academics, I'd jump at it.
A significant number of PhDs drop out, and likely literally everyone undergoes some sort of huge realization of how doing a PhD is completely and surprisingly unlike what they expected (which is still true even with all the advice telling people that doing a PhD will be surprisingly unlike what you expect). Someone dropping out of a lab is a significant loss in time and resource investment which labs would prefer to avoid, and the biggest predictor for staying in a PhD is substantial experience doing research (to have some sense of what the lifestyle is like) and little experience with higher paying jobs (here at MIT, I see many PhD programs prefer people straight out of undergrad, even though older people are orders of magnitude more productive in the early phase of their PhD).
> There are areas I feel I could be helpful with, e.g. creating reproducible experiments or ensuring a codebase could be open sourced in relation to a paper. It's a lot of grunt work that nobody really has time to do.
Nobody has time to do those things because they are valued less than publishing new research, and thus are of lower priority. You'll have a frustrating time in academia if you define yourself around your ability to do things that the people evaluating you believe to be unimportant.
If you can meet or exceed expectations on delivering research and get all those lower-priority things done, you'll be very well-liked. Just remember that the former is a requirement and the latter is icing on the cake.
"Or ensuring a codebase could be open sourced in relation to a paper"
This is required in a 1.7 million dollar grant I'm the PI on. I run a computational lab - making sure my research code base is high quality is one of my highest priorities, and I've seen people have very fruitful careers doing just that.
"If those things are valued, the academics will have already be doing them. Like in your lab, for instance."
We're trying to. Doesn't mean I wouldn't like to hire someone whose better at it. Just like we do a lot of mathematical analysis, but I was still excited about hiring an applied mathematician.
And when I said I knew people with fruitful careers, that implies I know academics already doing them.
That's great to hear. I do think those values are spreading, but they are definitely not universal yet.
If a prospective student wants to be sure that sort of contribution will be appreciated, they should seek out groups who have expressed interest in those skills.
I think your best option is to apply to a masters degree. These are easier to get into, especially for someone out of school for a few years. This is also an opportunity to find professors doing research in topics of your interest.
It has it's plussses and minuses. One thing about a Ph.D. program is that it's relatively easier to get funded. Master's students are more likely to have to pay tuition. On the other hand, you can bang out an M.S. in two years. 1.5 in some cases. If what you're after is knowledge and skills, like a BS++, do a Master's. If you want to learn how to advance the field, do a Ph.D. If you're on a mission to fix undergraduate education, please be aware that the problems are deep and systemic, and you will probably not find your idealism welcome.
At one point i was considering going for a research engineer position. That seems a bit rarer than a straight PhD program, but involves pretty much exactly what you describe. Pay of course won't be what you're used to in industry, but in my experience (several friends hold such positions) you have excellent job security and a very laid-back work life. I think you often are expected to already hold a PhD to guarantee that you'll actually be familiar enough with the things you're building out/for, though YMMV.
You're in very much the wrong part of the country, but I'd absolutely hire a dev who had been working for the past decade. And find them hugely valuable.
Hell, one of the rising stars of my field started as a dev.
> We are concerned about untrained amateurs going directly into trying to solve very difficult global problems. They can then cause harm overall, by lowering the average quality of analysis or launching ill-considered projects due to a lack of experience or understanding. A PhD reduces the risk you’ll accidentally do this.
This point is operating off of a different fundamental paradigm than the comments here that are talking about opportunity cost. Opportunity cost is looking at personal benefit, while the above is looking at the effectiveness at solving hard, large-scale problems.
Obviously, money is important, but for many (often those who are fortunate enough to have reached a certain income level), it isn't the only thing or even the primary thing.
I don't have much insight into how true the point actually is, but I found it to be interesting.