This comes up a lot: everyone has some "skills" that transfer from one domain to another. Read an interview with a titan of industry and they will invariably bring up the willingness to get up early and work hard that s/he "learned" from the paper route. The CEO of the studio "learned" to network from being in the mailroom and so on. In theory, almost any job can give you some transferable skills. A PhD may give you many more than most.
I know if you have PhD you (probably) worked really, really, really hard, were able to focus on a single problem for years, can communicate, analyze and so forth. I have enormous respect for PhDs, and believe they can deliver enormous value. But they are not the same kind of hire as an MBA.
What is the difference between a PhD and an MBA? I think to me it captures the single most important attribute that you cannot pick up directly: what people care about.
The thing about an MBA is, they really enjoy this "business stuff." They like thinking about markets, the customer, costs, finance, how to cross-sell, how to avoid dilution and the like. Its not just that the skills are different (though they are), it is that the interests are different.
Source: comp sci PhD student who switched to an MBA.
I don't think these articles are written for CS PhDs, who are an exceptional class of PhDs.
First, CS PhDs can ALWAYS get tenure-track academic jobs. Maybe not at a top R1, but getting a university teaching job in CS is not some sort of prize. Quite the opposite. Most non-phd-granting institutions with sub-billion endowments struggle to hire CS faculty (they pay sub-100K, sometimes as low as 65K... if you go that route in CS, your undergrads are making 3x your income at their first gig). So there's no "oh no plan B" fear. You don't need to be reassured your PhD wasn't a waste after failing to get any sort of academic post, because if you lower your expectations enough you will get an academic post. This is NOT true in nearly any other field.
2. CS PhDs, with a few exceptional subfields, are in high demand. It's pretty reasonable to expect 300K out of a top CS PhD program; the total comp number for top-tier MBAs is about half of that. CS PhDs who choose industry don't need to be reassured that their PhD has value. It's reflected in their compensation.
CS PhDs can ALWAYS get tenure-track academic jobs. Maybe not at a top R1, but getting a university teaching job in CS is not some sort of prize.
You got to be kidding. Competition for CS tenure track positions is insane. At my school, which is ranked around 50 in the nation, you would have to be pretty outstanding to get it (most candidates were from top-5 schools, in hottest fields, with many strong publications). I’ve heard such numbers as 100 candidates for a single spot.
> At my school, which is ranked around 50 in the nation
Top 50 on US News & World Report or CS Rankings means you're at a very good R1 institution. At #50, your institution is ranked ABOVE places like Vanderbilt, Notre Dame, RIT, Syracuse, Clemson, ...
You do realize that the USA has nearly 4,000 colleges, right?
If you scroll all the way down to the bottom of the CS rankings in US News, you reach Walden University at #186. Which means the colleges that US news even bothers to rank in CS constitute less than 1% of the total number of colleges in the US. And you're at a place that's in the top third of the <1%!!!
When I say academic CS jobs are easy to get, I'm referring to jobs at the >3,500 US higher ed institutions that aren't even included in US New's CS rankings.
Again, in fields like Mathematics or Biology even TT jobs at unranked/low ranked places are non-trivial to secure.
If I had to guess, OP is at an institution in the northeast/west/Chicago. Academic CS recruiting in the south and in mid-tier cities is typically more difficult (some of that might be preference, but the big thing is two body problems. I love New Orleans but can't imagine solving an academic two body problem there is particularly easy).
For anyone wondering: the two body problem is a phenomenon that happens when two academics marry. They both want an academic job together, at the same institution preferably, but there's usually only enough room for 1. What many couples do to solve this is accept postdocs at different universities (because there's a two body problem for post docs too, so you can't usually find 2 postdoc positions at the same institution), and then wait around until they can find two assistant professorship openings at the same institution.
Sometimes if the candidate is really good, one department will ask another to make room for another faculty there. But that can be a political nightmare. Or if both researchers are in the same field e.g. CS, then the department might have to hire both even if they really only would take the one.
To give a different anecdote: at my school which is ranked closer to 100, we got 40 applicants, of which about 15 meet basic qualifications we're looking for (have a PhD, in the right field, have any kind of publication record, have ever taught a class). And I think we are doing pretty well compared to some places, since we're in a major coastal city and have a relatively light teaching load for a non-R1 place (2/2). A lot of places have been outright failing their CS faculty searches in the past few years, and I think more than usual will fail this year.
> 40 applicants... about 15 meet basic qualifications we're looking for... 2/2 load... major coastal city
Wow. Things are even worse than I thought. Your institution sounds like the rare type of place that shouldn't have a problem hiring. Good luck with your search.
> have a PhD, in the right field, have any kind of publication record, have ever taught a class
Most places with a teaching load higher than 3/2 dropped all three of those requirements from their job ads years ago.
They need CS lecturers all the way down to the worst college in the country. Hundreds and hundreds of departments. Everyone applies to the top, but they filter down, and if you look at who’s teaching at the lowest they clearly aren’t superstars and it wouldn’t take much to compete with them.
Even in non-CS STEM fields, there are a ton of industry jobs. Many of the students in my doctoral cohort went for industry jobs, and almost all of them are making good money (over 120k annually) and are generally happy in their careers. The problem from what I hear from them is the tendency to be slotted into super-technician roles where you are in charge of a single piece of specialized equipment. People tend to stagnate in such jobs (even though the compensation is often really generous), and such people often find themselves struggling after a decade.
However, I think it's not the Ph.D. that is not the problem - instead, it's the postdoc. While a Ph.D. is a terminal degree associated with prestige and career advancement, the outcome of postdoctoral training is far more diffuse. It ostensibly prepares you for academia yet often fails to teach essential academic skills like writing grants as the sole PI. The funnel is also really narrow, and many postdocs transition to industry after a few years - often in a very similar role and salary that they would have got straight out of their doctoral training anyway.
As a postdoc, I really feel this comment. However, a lot of people I know do postdocs since its essentially become a requirement for many industry jobs, and getting a job right after PhD is becoming harder & harder. I know people who got industry jobs (pharma) at the salary band you mentioned after 6 years of postdoctoral experience.
What’s the stigma against postdoc in the US? I don’t get it?
In the UK a postdoc is the first job after your PhD. You obviously aren’t going to get a professorship for a couple of decades, and will be too junior for a lectureship as well, so you have to do something in between. What do you do if not a postdoc?
The problem is we train way more PhDs than there are tenure track positions available. Most PhDs won’t be lucky enough to get one of those academic positions, and a postdoc is just delaying the inevitable transition to an industrial job. Postdocs are for all-star students with a good academic pedigree and publishing track record who have a good shot at tenure. People who were less then that (such as myself) are often better served starting their career outside academia.
From a pragmatic perspective that is how the calculus worked for me. It’s probably where the stigma arises from as well. Although I wish it weren’t like that, I never thought the purpose of a college/university education should be so limited to ‘job training’ (that’s what trade schools are for). That’s the American perspective, anyways.
I think the point parent is trying to make is that the majority of postdocs will move to an industry job simply because that's where most of the jobs are, even if they'd prefer to remain in academia.
While postdoctoral jobs aren't age limited, they are defined in the US as a temporary training position, and it's not possible to remain a postdoc for very long. The vast majority of postdocs in the US are foreign nationals, on a visa scheme that limits their stay to a theoretical maximum of 5 years. The salary is low enough that it dissuades postdocs from remaining in these positions for very long anyway. Some institutions may support transitioning senior postdocs to a more semi-permanent "associate" position, but that's not usually the case. The visa scheme also disallows postdocs from applying for a green card (while maintaining status), which complicates matters further.
The thing is, a postdoc is a classic case of credential inflation. You don't really need a postdoc, especially if an industry job is your goal. The purpose of an academic postdoc is to strictly increase the items in your CV to make you more marketable for academic employment. Industries are hiring senior postdocs strictly because more postdocs are applying for such jobs. And this is a side effect of poor mentorship in academia in general, with freshly minted PhDs drifting into postdocs just because.
Very few graduate students think seriously about their destination during their doctoral program and are happy to be in the lab all day meeting goals their advisor sets out for them. And advisors are also happy to let this be the state of affairs since you are getting highly trained, motivated workers for pennies. This passivity starts dissipating only during a postdoc and not always. Poor mentorship can be excused in industry, but this is inexcusable when academics are paid to be mentors; it's literally in a professor's job description. NSF/DOE/NIH grants all have significant mentorship sections - and they are there for a reason.
Exactly. But this is very variable across fields. In CS or Math, it is feasible to get a tenure-track position or the equivalent in industry without a postdoc.
In Biology for instance, inflation is so insane that most tenure-track positions demand you have significant postdoctoral experience. Same applies to many industry jobs.
The issue is the inflation of requirements in some fields due to competition. A postdoc is generally understood to be further training in running your own lab as a PI - you might manage and guide doctoral students, for example, or learn how to write and apply for grants. With that in mind, a lot of people are happy to commit to a few years of postdoctoral work. In some fields, however, the extreme shortage of tenure track positions has made very long postdoctoral stints a requirement, rather than as something to sweeten the deal. Like others have pointed out, in CS a PhD is enough to get you a tenure track academic position, while industry jobs abound. Meanwhile, in biology I commonly see 5+ years of postdoc experience being typical just to be in the consideration for an industry job that 10 years ago would have happily hired a fresh PhD. The general thinking is that a postdoc only makes sense if you want an academic position, but the glut of PhDs in the market has pushed up the requirements for all kinds of jobs.
I still don’t really get it sorry - why isn’t being a postdoctoral research assistant a valid career by itself?
I know people who don’t want to teach or lead a group or work in industry so they do postdoctoral RA work. That’s their career. Why’s that not seen as a valid thing to do in the US? People are saying ‘because you might not get tenure.’ Ok but not everyone wants that career.
Because it's time limited. National lab postdocs are valid only upto 5 years from PhD. Many grants won't fund you if you are X years since PhD.
That's why the position of Research Assistant Professors exist - the way around the postdoc time limit loophole.
However, RAPs are significantly more expensive than postdocs, which is why faculty are more eager to phase out senior postdocs and hire fresh blood rather than convert them to RAPs.
In my experience, if it's research that gets you up in the morning then a postdoc is the sweet-spot - you're experienced enough to make good progress but not too senior that you're sat in meetings all day. It's just a shame the salary is so bad you're almost forced to move on.
> It's just a shame the salary is so bad you're almost forced to move on.
UC Berkeley, for example, pays its postdocs $60k to $65k per year. It's challenging to live in the Bay Area for that money. Also, remember many postdocs have kids or are planning to have one. Child care is at least $2000 a month in the Bay Area. Asking bright, motivated individuals to delay their life decisions in pursuit of ill-defined scientific ideals is something I have a tough time defending.
This is why I am so pissed off at academic hypocrisy currently. Labs were shut down for months. For what? To protect whom? The senior faculty who did their Ph.D. in the early eighties and refuses to transition to emeritus status?
Junior researchers were absolutely shortchanged in the pandemic shutdown, and nobody is bothered about it. For all the talk about keeping people safe, why were younger scientists denied access to experimental facilities, and yet almost no funding agencies extended grant contracts? Almost zero institutions extended their timed postdoc contracts too. And to add insult to the injury, most academic institutions had hiring freezes in 2020, effectively kiboshing careers of young scientists caught in this trap.
Honestly, I am disgusted with university faculty right now - all fancy talk, no action.
An example of the hypocrisy: a prominent university in the Boston area that announced that they will suspend on campus childcare from January. Labs are still open, and the population of lab workers most likely to have benefited from on campus childcare were postdocs. Now they have to find at home childcare at their own expense. Postdocs also tend to be overwhelmingly foreign nationals who want to avoid rocking the boat.
Some hiring freezes have actually continued even to today. I applied to one postdoctoral fellowship in the Bay Area in March 2020, and the organization still sends me a monthly update that they have suspended all hiring due to the pandemic and will resume hiring once the pandemic is over (as ill-defined as that may be). That's almost two years at this point. That is a lifetime for young scientists. I agree that decisions like these do effectively end the scientific careers of many young scientists, or at least push them out of science as a career choice. It's sad but true.
Thing is - we don't really know at this point whether this wilful destruction of the scientific careers of a cohort were avoidable or not. And zero scientific administrators will face zero consequences for this.
Whatever it is, the reverse will actually be trotted out, scientists saved lives and pat themselves on their backs.
Higher ed is an industry in crisis. Get out and be happy that you realized this before it was too late to leave that hell hole of a sector. The grass is greener. I promise.
If CS PhD's are an exceptional class then I think it's only by virtue of supply & demand. There are a larger # of opportunities in industry, and they often pay better.
As for students earning 3x your income, it seems like you're taking the lowest paid professors and comparing to the highest paid grads. The truth is, if you're at a school where the CS program only pays faculty $65k then you're probably not walking into a $200k+ job on graduation except as an extreme outlier. Those students aren't going to a school with the name recognition required to easily open doors for entry-level $200k positions.
Many liberal arts schools pay everyone like they are liberal arts professors. So they have impossible time hiring people in valuable professional fields like CS and statistics. They still can get solid students though. Disproportionately locals and by offering scholarships. Also prior hot areas like "data science" and security indeed led to a lot of insane starting salary stories ($300k+) for merely-above-average students from meh schools with exactly the right skills. Even the median starting salaries in hot areas have been surprisingly high for a while.
> If CS PhD's are an exceptional class then I think it's only by virtue of supply & demand. There are a larger # of opportunities in industry, and they often pay better.
Well of course. What else would it be? Doesn't it always work out in life that the jobs you can get are the ones where they need you more than you need them?
> Many liberal arts schools pay everyone like they are liberal arts professors.
Presumably, you mean something like “arts and humanities” the second time you say “liberal arts”, since the liberal arts include the natural sciences (both life and physical), social sciences, mathematics (including, among other things, computer science), arts, and humanities.
Nah, I think "liberal arts" fits the bill. Including natural sciences. In contrast to the professional colleges: Law/Business/Medicine/Engineering (which CS should be at-parity with).
Of course LACs don't have law schools or med schools or (usually) engineering colleges, so the difficulty of properly compensating CS is understandable with respect to those fields.
However, LACs do tend to have Finance/Accounting, and often pay folks in those departments better, but for whatever reason don't treat their CS faculty like their Finance faculty. My own pet theory for this market mismatch is that finance/business types tend to dominate those college's boards. LAC boards have a lot of older folks who still think of CS as code monkeying.
Absolutely. I never meant to imply otherwise. Worth mentioning that this will be true for a while though.
> The truth is, if you're at a school where the CS program only pays faculty $65k
I have first-hand knowledge about 3 such institutions, all of which place at least one student in a position that pays >$150K every year. Perhaps shy of 3x but higher than 2x for sure. Averages tend to be around 90K (pulled down by people who choose to go to grad school).
> Those students aren't going to a school with the name recognition required to easily open doors for entry-level $200k positions.
But they do receive fantastic educations, because they are at teaching-oriented institutions and get hand-held through their pre-career years (internship placement, interviewing skills, etc. are all coached 1:1).
The basic issue is that administrators at lower-tier institutions haven't yet flipped the switch where "CS = Finance/Accounting". They continue to hope and dream that their mathematics faculty can pick up the slack, as if it's still the 80s/early 90s and CS hasn't blossomed into its own highly specialized field.
>place at least one student in a position that pays >$150K every year.
That's pretty much what I mean by it being an outlier: it's not the norm.
>But they do receive fantastic educations, because they are at teaching-oriented...
That's hit or miss: Anyone that's highly self-motivated will do fine at most schools. One prof I had was at best disinterested. They basically summarized chapters on the chalkboard. Their apathy in answering questions went a long way towards discouraging questions in the first place. It was a shame too because their personal research was interesting: 3-dimensional computer vision. Unfortunately this was a foundation course too. I did fine, but other students that were excited about CS yet had no exposure to it before college ended up with C/D/F grades and were either setup for ever tougher times when the work got harder or simply left the program. (Computers were still expensive at that time and the dot com bubble was still building, so it was pretty common for students to arrive enthusiastic about the new tech field but completely without prior exposure to it. Maybe things are better today).
> The truth is, if you're at a school where the CS program only pays faculty $65k then you're probably not walking into a $200k+ job on graduation except as an extreme outlier.
Only because they're not located in the "$200k+ starting" job market. If they move to SF then they're immediately worth that much.
Thankfully the jobs are diffusing across America now.
Most academic professors in CS are mentoring much more than one PhD students over the course of their academic careers, yes? Therefore, as with other academic disciplines, the ratio of academic tenure-track job openings to new PhD's looking for academic jobs is >> 1, even if you look at industry demand for PhD's.
No. The trope you're repeating here has some truth in the humanities and some natural sciences, but it's not true in CS.
Industry has a veracious appetite for CS PhDs. Only 40% of PhDS stay in academia, and of those only a fraction take jobs at universities that grant PhDs. Tenured CS professors somewhat leave for industry at much higher rates than other fields.
Meanwhile, undergraduate enrollments have surged without a corresponding increase in PhD enrollment.
Multiple sources of demand are surging, but supply is waning.
From a CRA report on the subject:
The demand for PhD’s in computer science (CS) in the US continues to outpace the supply. Both industry and academia struggle to fill positions. Since 2014, approximately 2000 CS PhDs have been awarded annually and about 60% of new PhD graduates take jobs in industry [ZwBi19]. The rapid growth in undergraduate CS enrollments during the last decade has significantly increased the number of open faculty positions at academic institutions, but has so far not led to an increase in the number of PhDs graduated.
Historically, US PhD programs have relied heavily on international students. Figure 1 shows the number
of CS PhDs awarded annually to domestic and international students since 1985. The percentage of
domestic PhD students graduating decreased from 69% in 1985 to 37% in 2018.
The reliance on international students to drive innovation and leadership in computing research in the US
has become unstable as international students increasingly face obstacles or disincentives to study in the
US, and an increasing number return home to attractive opportunities after graduation from a US
university. Moreover, some areas of computing have security implications that make positions in those
areas inaccessible to non-US nationals. The continuing demand for PhDs in computer science combined
with this instability of international student participation requires bold action to increase the number
of domestic students completing a PhD in computer science.
So... Why don't CS PhD's get together to do what medical doctors do and form a cartel to further limit the number of PhDs that are granted each year?
The current supply/demand imbalance sounds like the perfect leverage to get the ball rolling. Institutions that don't play ball could be boycotted.
A good number of specialties in medicine have a median income in the upper two thirds of the six digit range. It seems like limiting production of CS PhDs could be equally lucrative.
> A good number of specialties in medicine have a median income in the upper two thirds of the six digit range. It seems like limiting production of CS PhDs could be equally lucrative.
Probably not. Industry would route around the PhD system. Academia would just hire folks with masters degrees as ad juncts.
In order for such a scheme to work, you need a licensing regime. If anyone could practice medicine at any hospital by passing an on-site interview, the AMA would have substantially less power.
> Most academic professors in CS are mentoring much more than one PhD students over the course of their academic careers, yes?
Let's take a look at this oft-repeated claim. If we consider the Taulbee CS department survey [0], we can see the breakdown of the CS faculty population. Based on the survey, we see 32% are old-guard full professors, 16% associate professors, 20% assistant professors, 20% teaching faculty, 5% research faculty, and 5% postdocs. These are all positions with a PhD as a requirement.
This tells us that about 49% of the CS department faculty has tenure, while 51% is either working toward it or will never have it. Within that 51% there's a lot of churn. Many in that group will fail to achieve tenure. Others will work for a 3-5 year contract and move on afterward.
That's also not to say that tenured faculty never quit. Some leave academia for industry. Others go on sabbatical. Others leave for the NSF or some other institution. Others are poached by another department for a chair position. Some will leave once their spouse gets tenure at a superior institution. Others will go home to their native country.
So while it is true there are only so many positions, openings are happening all the time.
If you look at where CS PhDs are going after they graduate, we see that ~55% are going straight into industry, while only 30% are sticking around to find an academic job. This is a testament to the draw of industry in the CS field, and doesn't even capture the draw for faculty, which can be much stronger (much higher comps offered for experienced researchers as opposed to newly minted PhDs).
To your second point, most of the people I know who got a PhD (myself included, focus in operating systems) and went into industry didn’t clear $300k. Maybe it depends on your definition of top school though, this was around #14, plus or minus a bit.
> your undergrads are making 3x your income at their first gig
Why are you making stuff up?
These sorts of careless lies, which are all too common, can be hard on young people in college, or considering college, or otherwise. They're either trying to value their potential education or wondering why the fuck they aren't getting the job offers they should be getting because so many are telling them their degree is the gateway to instant riches, and you're not helping.
> the total comp number for top-tier MBAs is about half of that.
Wrong in a very different way here. You should stop.
Serious question: how many students and faculty currently at low-tier institutions have you talked with in the last two weeks? Two months? Two years? Do you sit on any boards or advisory panels at these types of institutions? Do you actively recruit from these types of institutions? I do. All of those things.
I know what I'm talking about. Stop being mean to me.
LACs pay junior faculty $65K and those faculty routinely place students in positions with >$150K total comp. These are facts, even at low tier colleges.
Every student? No. Some students every year? Absolutely. If you teach CS well and make $65K, almost all your students will make more than you do at their first position and many will make 2x-3x. More than enough that you'll start asking "why the hell am I here?"
> These sorts of careless lies...
This rant sounds extremely personal. Not going to touch this.
>> the total comp number for top-tier MBAs is about half of that.
> Wrong in a very different way here. You should stop.
Total comp out of a top CS PhD programs lately is around 200K-300K range with some outliers.
Total comp out of a top MBA program lately is around 100K-200K range with some outliers.
Average outcomes vary considerably. That's why my original post is properly conditioned on "top".
> LACs pay junior faculty $65K and those faculty routinely place students in positions with >$150K total comp. These are facts, even at low tier colleges.
You're leaving out a key fact, which is salary for academics is reported for 9-months, not 12.
A 9 month salary of $65k would put you at the bottom 10% of teaching faculty in the nation. The 50th percentile is more like $82k, which is $109k annualized. If you start at $65k, I think by the time you actually graduate any students you should be making a lot more than that. And if you're not, there's got to be some other reason why you're not making a more representative salary.
But yes, in general academics can make less than the students they graduate. Many academics are okay with that because:
1. It's really hard to put a value on not having a boss in the traditional sense.
2. It's also hard to put a value on getting 3 months off in the summer and 1 month off in the winter every year.
Then again, I guess it's not hard to put a value on that: it's whatever they forego in extra salary working in industry. In that sense while the students earn more, they don't 10 paid weeks off + 11 weeks unpaid vacation in the summer.
Well, yeah, my entire point is that those $65K places fail to hire/retain, precisely because they can't/won't pay and consider the bottom of the CRA range insanely generous/competitive. Maybe I wasn't clear enough about that point. The whole thread descended into pedantry about a multiple when everyone concedes the basic point.
What the CRA survey doesn't tell you is that a huge percentage of CS departments just have failed searches year after year. (And, actually, there is some CRA-E report somewhere that laments the incredible difficulty of hiring in CS.)
I can see a scenario where $65k would be attractive. If it were some small college in the middle of nowhere with a 3/3 course load, that would be a very attractive position for some people. I bet when you look into it though, those schools want something like a 5/5 load with min 3 preps. No thank you.
> The whole thread descended into pedantry about a multiple when everyone concedes the basic point.
What else would you expect from a thread that has attracted a bunch of academics? :P
> If it were some small college in the middle of nowhere with a 3/3 course load, that would be a very attractive position for some people.
Can't imagine who. Maybe the childless? Or perhaps folks with trust funds.
> I bet when you look into it though, those schools want something like a 5/5 load with min 3 preps. No thank you.
3/3 was available pre-COVID. But the pandemic turbo-charged the MBAification of higher ed.
Today? Maybe you can find 4/4, but the high prep count is real. Oh, and you're definitely the chair at some point before tenure. These are like 50-60 hour weeks if you're doing the job well. Not worth the short summer off (during which you will... sit around in the middle of nowhere and pinch pennies if you happen to have a family)
>> The whole thread descended into pedantry about a multiple when everyone concedes the basic point.
> What else would you expect from a thread that has attracted a bunch of academics? :P
I visibly winced when a dean told me about the unique allure of life immersed in academia ;-D
> If you teach CS well and make $65K, almost all your students will make more than you do at their first position and many will make 2x-3x
"Many" is a weasel word that adds a different flavor to your original assertion that "your undergrads are making 3x your income at their first gig." What you're saying is still crazy hyperbole. I simply cannot imagine what your source of data is here. A randomly chosen google hit shows that the average starting salary for a new CS undergrad is around $68k, which seems about right.
It is still, if we're being honest, probably a bit humbling for a junior professor to be making the same as a new graduate. But you had to lean into the "2x-3x" hyperbole...
>> These sorts of careless lies...
> This rant sounds extremely personal. Not going to touch this.
Oh, touch it. Young people trying to gauge the profession, higher education, and its' costs are going to read your comments. They are entering what is often a lucrative profession but they are not going to be making three times as much as their teachers. Why make stuff up? It's not helping anybody.
> Total comp out of a top CS PhD programs lately is around 200K-300K range with some outliers.
> Total comp out of a top MBA program lately is around 100K-200K range with some outliers.
This would be a lot more compelling if you provided citations. The salary numbers someone else provided for the Stanford MBA program, numbers which you were dismissive of, included range, median, and mode, and those numbers were higher. But the numbers I (and I suspect, anyone involved in the profession) are likely to be most skeptical about are the numbers you're talking about for CS PhD new grads. Those people will often gravitate towards postdoc and teaching positions, while the Masters students will often gravitate towards FAANG jobs.
To be honest I would be delighted to learn that newly graduated CS PhDs from top programs are making, on average, as much as those entering industry with a Masters, since it would probably signal a lot more money being put towards research. I'm pretty sure the numbers you're talking about would have them making twice as much as the new Masters grads, which, again, great! I would be delighted to learn it's true.
Dude. Sometimes is also a qualifier. It's right there in my original quote. WTF is this conversation even about anymore?
This is getting unbelievably pedantic.
Look, it's a thing at the 3 institutions I advise for their CS faculty to make below-market wages and for their graduates to make above-average wages. I think this pattern of facts is uniquely common at lower tier LACs with CS PhD on faculty. Those institutions pay uniquely low rates for CS faculty, but the value of LAC-style 1:1 mentorship from a CS PhD is enormous and consistently results in better than average placements.
I didn't claim this is the base case. I used conditioning words. Sometimes. Up to. From first post on-wards.
you can call me a liar. Whatever. This is real phenomenon. These situations exist. You've even conceded my fundamental point about this subset of CS PhDs AFAICT: that the jobs are plentiful and not particularly attractive.
Next, you claimed Stanford MBAs make more than Stanford CS PhDs because some Stanford CS PhDs choose academia (and, yes, the Stanford CS PhDs who choose industry make more than their MBA counter-parts; go look at levels.fyi). That's all fine and well. Probably true. here's the thing, though. My top-most post in this thread explicitly preconditioned this branch of the conversation on "CS PhDs who choose industry". So at this point I'm not really sure what point you're trying to make. Clearly, we're so far down-thread that the plot is completely lost.
The pedantry is trying my patience. Go back and read my original post. Is there any aspect of that basic thesis that you actually disagree with in a substantive way?
> the salary numbers someone else provided for the Stanford MBA program, [...] and those numbers were higher.
Standford is widely-regarded as one of the very tippy-top MBA programs. The Economist ranks them #5 (but #1 in post-MBA earnings). FT ranks them #2 and US News and World Report ranks them #1.
I don't consider a PhD to be in any way equivalent to a MBA. As far as training goes, a PhD is about how to do research. The result often in fact is a step down in salary. Plus it seems to be very difficult to find any recent data on salaries.
Right. But are there are "hard" skills from an MBA that transfer? Not to offend, but MBAs just seem like networking and drinking degrees to me, for people that love business. I've seen MBAs syllabuses: I've literally seen very basic things like "profit = revenue - cost" on slides. It seems like with MBAs it's almost entirely soft skills, no?
> are there are "hard" skills from an MBA that transfer?
You are asking a genuine question that has a lot of backing research. Before I became a Quant, I was a student at the University of Chicago, getting the Masters in Financial Math. So the professor who taught the Options course at MSFM told us one day - Now I have to teach this same material to the MBA class, but at a 10,000 feet level.
One of the students asked him to elaborate. So he says - If you have a scale of 1 to 100, with 100 being the hardest, then a UChicago Finance PhD is like 100. A UChicago MSFM is about 50. A UChicago MBA is like 10. A regular (non-UChicago) MBA is a 1.
More concretely, someone like me, the 50, can put up the Black Scholes on the blackboard from memory. I can derive it, solve it, code it up in C++ and price your options on a real equity & tell you whether you should buy/sell the damn thing. But that's as far as I was taught.
The Finance PhD i.e. the 100, can derive and solve not just the BS but a whole family of models - the Heston, Derman, SABR, BDT, HJM, HW etc etc - there's like a dozen of these PDEs & it gets seriously complicated very soon. So supposedly this Finance PhD can do all of that & more.
The UChicago MBA, which is the 10, knows Black Scholes, knows what the Greeks are & can eyeball the value of the Greek & tell you whether the option is overpriced or not, but can't derive/solve/code up the PDE.
The non-UChicago MBA has heard of something called the Black Scholes but that's about it.
Ofcourse that's his opinion & you should take it with a grain of salt. That said, during my time in the investment banking industry, I have worked with a boatload of MBAs, some quants(MSFMs) & a few PhDs - his opinions bore true.
This sounds like something a biased physics professor who values his own field more than others could also say about teaching quantum computing to financial math students. You don't need more than an overview and it's hard. What percentage of MBAs will need to analytically solve PDEs?
Good MBA programs have courses in linear programming and optimization, complex financial modeling, and supply chain that are quantitative in nature.
There are also technology components that review basic IT setups for a business.
They teach you how to think about markets, and how to review financial statements and understand a business’ profit and operations, and offer opportunity to get capital. The org behavior courses teach about how to actually get an organization to achieve its desired outcome. And that can be way more powerful than a solo person working alone in a lab.
An ideal MBA with experience post-school would be able to look at an idea, figure out if it’s a viable product with a good market fit, have the contacts to get financing, the connections to get it produced at the right quality with the lowest possible cost, hedge against forex risks, work with lawyers to negotiate the associated contracts (or know enough to not get screwed), record and accurately report the profits to governments and stakeholders, and get the right people onboarded (these are the soft skills) to get the work done.
I wrote the above, but I think the problematic ones are the “text book MBAs” (I just coined that) who go to mediocre, money-grabbing programs and think they are good to go because they read text books and have the degree. There are too many of these in the marketplace.
The other skill a good MBA imparts is solid written and oral communication.
I don’t have an MBA, but considered it and chose another MS program, cherry picking MBA courses I felt were relevant.
But to discount an MBA as two more years of undergrad partying is missing out on a lot of skills they can bring.
"I wrote the above, but I think the problematic ones are the “text book MBAs” (I just coined that) who go to mediocre, money-grabbing programs and think they are good to go because they read text books and have the degree. There are too many of these in the marketplace."
This is essentially true for all college degrees at this point.
I have friends who earned MBA's, typically through "executive" programs attached to major universities. Their training was being paid for, and they were looking for the more prestigious programs.
These programs did not accept students straight out of college, and so the people going into them already tend to have some business experience, and have already sorted themselves according to their math ability. They're already known to be diligent, organized, and satisfactorily literate.
The ones who went in with strong math skills found the quantitative courses to be a breeze. Everybody learned accounting and finance principles, which are useful for being conversant in that language. They took a course in business communications, and came out with a fairly standardized approach to writing memoranda and reports. I think they learned how to write and critique a business plan. There were one or more courses on business law, including HR.
I call these "hard" skills because they are technical in nature. I believe that these are all good skills for entry into middle management.
Just as a layperson often underestimates the true depth of a PhD's expertise in their field of study, many people often underestimate the skill of someone who has spent a similarly long time meditating on the "soft skills" that actually glue a good idea into a successful business.
That's not an argument against MBA's having hard skills. It's an argument for a finely standardized product.
MBAs are a commodity, but they do or at least can have hard skills. I've wasted plenty of time and money on subpar market entry and expansion that could have been saved had I worked with people with some theoretical background on those.
You could write an encyclopedic book of counter intuitive organizational dynamics. While I don't know that this is the case with all of them, perhaps MBAs could be understood as internalizing the combination of such an encyclopedia with a few others besides and gaining insight into how they interplay with supply line logistics.
The insight mostly comes from experiences; the knowledge of what to reference to find where to find where to find what to find could be (boringly) boiled down to rote memorization and distributed through your standard classroom practices.
I have heard some version of this argument frequently over the years: there is this "hard" skill in engineering or physics or computer science that has real lasting value, and there there are these other disciplines that don't really have much of value.
Now, let's unpack what people mean by "hard" skills. Do you mean they are difficult to learn? Do you mean they are useful in today and tomorrow's economy? What exactly does the hard in hard skill mean.
I think you can see how the deconstruction of the argument works here. If by hard you mean difficult to learn, then I can list a bunch of hard skills that are not directly useful. If someone can rattle of all the major proofs in quantum mechanics on a blackboard, they may be a genius, have worked hard, and can do amazing things with their brain but that effort may not directly make them more employable.
If by hard skills you mean things that are useful to the economy, then that changes. I could have spent 4 years learning WAMP and then when I come out the economy has moved on to some other framework.
Maybe by hard skills you mean some skills that make it easy to pick up the other computer skills. So I may have learned a particular computer language, but along the way I know conditionals, loops, computational expense and so on. Are these harder to learn than presentation skills? Are they more useful than financial statements?
The example here of a slide that says "profit = revenue - cost" is hardly fair. That's like saying I once saw a piece of code on github that was very poor. Picking individual examples of something you don't find challenging is easy in any field.
Explain what you mean by hard skills, and then we can talk about whether a particular discipline has them or not.
> Now, let's unpack what people mean by "hard" skills. Do you mean they are difficult to learn? Do you mean they are useful in today and tomorrow's economy? What exactly does the hard in hard skill mean.
Hard skills are what is the scientific core of your degree course or PhD program.
So knowledge about theoretical physics of you study physics or do your PhD in it. Knowledge of mathematics and ability to do hard proofs for mathematics.
For computer science, the situation is more subtle: Here the hard skill is knowledge of computer science, being able to understand papers about this topic (degree course) or doing research (PhD program). In this sense, I would consider programming knowledge as a (central!) soft skill of a CS program.
What I'd expect in an MBA : accounting, finance, marketing, basics in legal, management, business strategy, etc.
Sure. It's not higher maths. But they are real skills that are 1) hard 2) are not 'natural'.
Yes. It's also a good way to meet other people like you. But then the same thing goes for any 'skill'. How many businesses weren't founded by comp Sci students from the same uni?
MBA's are generalist degrees. You'll see ex-military, social workers, English majors in addition to technical people get MBAs.
As such, the content is generalist in nature. You don't do a deep dive on any particular topic (well you can in 2nd year, but it's still nowhere as deep as a specialist). A technical person could do the statistics class in their sleep (it's freshman statistics) while social worker might struggle. But the technical person would learn something new in a negotiations class, but the English major who studied debate might not.
The goal of the MBA is to give students a broad understanding of business, not specialists in any one area. So when the new graduate is given the job of determining if a business line is profitable, they can go and talk to finance and know enough that they can ask the right questions.
So the hard skills is a difficult thing to pin down - for some students no (because they had them already) for others? Yeah. And a big part of the skill set (hard? soft?) is an understanding of the different parts of a business and how they interact. That way when the CEO asks your opinion on whether or not the company should acquire another you don't say "yeah their technical team is great!", you can talk about finances, talent, product lines, competitors, etc.
Yes, it's entirely soft skills. But they work very hard at using their soft skills to solve trivial business problems.
Every worked a marketing or revops job? It's mind-numbingly trivial, the likes of you and I wouldn't last a day. But these MBAs do it all without blinking, and do it passionately. And I, for one, appreciate them for that.
an MBA involves a lot of case studies. that's breaking down what happened with a given business over a period of time in some detail. at least back in my day, MBA programs had a lot of reading and writing them
look up some examples of MBA case studies some time. they can be very interesting and informative
High level MBAs are for networking. But the degree is in Business Administration which is a bunch of broad spectrum lightly technical work in accounting, logistics, etc, which lots of businesses or entrepreneurs need and not everyone can self teach like individualist programmers do.
if you have PhD you (probably) worked really, really, really hard, were able to focus on a single problem for years, can communicate, analyze and so forth
Correction: if you have multiple publications in top conferences as the first author. Most phds don’t have that. Part of my job is to hire ML researchers. 90% of phds who apply are quite pathetic as far as their publication record.
> I know if you have PhD you (probably) worked really, really, really hard, were able to focus on a single problem for years,
Focus, yes.
Hard work - not a requirement. Depends entirely on your professor's requirements to get the PhD. Most of what I saw in my time in engineering/physics: Just consistently do the work and stick to it without getting too many distractions. Hard work merely made you get the PhD quicker.
Of course, if your advisor is fussy, this strategy won't work.
The one other observation is that when you look at the output and career trajectory, there isn't much correlation with the value of the PhD and how hard the advisor makes you work. Most of the value of the hard work slave driving professors make you do goes to your professor, not you,
Do you feel like the "search space" of MBA problems is somewhat finite?
Like, if you read 5 books and 20 blogs on VCs you are unlikely to be mystified by anything that happens to you as you seed/VC/pivot/exit whatever. There are only so many possibilities.
(MBA stuff is still hard, and sensitive to luck/timing/networking/cultural factors..)
Having obtained a PhD myself, I would say its definitely helped in my career, though that is largely because my industry work is related to my PhD; people tend to assume I'm an expert in this super complicated thing and they just need someone around who understands it as some kind of insurance or person to turn to in a crisis, even though with the PhD itself I was working in a niche within a niche within a niche which won't really help the company. Obviously during the PhD I got a working knowledge of the whole field, but I often feel as if a lot of what the company uses me for is pretty obvious and googleable. Very, very rarely do I actually use my PhD expertise. I definitely think if you work as a software engineer in tech a PhD would have too large an opportunity cost vs just working for it to be worth it, especially in the US where PhDs typically last ~5 years.
Aside from that, I think I was lucky to have a supervisor who taught me how to make good presentations. I definitely think my presentations are a lot better than most peoples in industry and that's benefitted my career too. Though, I've also seen very many terrible presentations in academia so I'm probably just lucky there. I'm not sure its the case PhDs typically give better presentations. My LaTeX CV looks incredible though :) (though I highly doubt that will ever affect whether or not I get a particular job).
I feel going from a CS PhD to working as a standard software engineer is probably a waste of the person’s time, while also resulting in lower pay.
It’s not like other fields where your area of research might be very different from what you get employed for; in CS, your area of research probably has a startup that could use your domain skills. Eg: a CS Security PhD can find gainful employment on the security teams of any of the big software giants doing work that’s not too dissimilar from their research.
Not a PhD just a MSc, but the presentation skill I picked up along the way was definitely the most important thing I got out of it.
To paraphrase the words of my professor: if I really understand something, I must be able to explain it to any audience in 2 seconds, 2 minutes, or 2 hours.
The one thing I learned as a mcdonald's cashier was how to compute change from a $20 in my head, quickly.
There was nothing about the workflow system of my mcdonalds that delivered hot food fast. There was a ton of inefficiency in it. I made a number of recommendations to my manager who told me I needed to attend Hamburger University lectures of why what McDonalds did was so great.
Same here. There is something about practicing customer service in a human-facing job that conditions ones mind for immediate agressive problem solving.
Perhaps I'm definitionally-challenged, but none of the skills listed strike me as "hard" skills. I see a hard skill as being one for which a person receives domain-specific training, e.g. CS major at $SCHOOL learning to program in Rust. In contrast, the soft skills are those obtained in the process of achieving the hard skills, e.g. aforementioned CS major developing teamwork and time management skills while working on a group project.
In this vein, none of the skills listed in this article (grant-writing, data analysis, information synthesis, data presentation) are domain-specific.
The hard/soft distinction is a bad metaphor that we should stop using, especially when talking about science. Both soft and hard have other meanings that come with so many implicit cultural assumptions. Domain-specific vs domain-independent is far more precise. And if you want to refer to social/communication skills, why not just say that?
Hard means it is easy to measure as they have rigid shape and form. Soft means it is hard to measure as soft items doesn't have shape or other properties similar to hard items.
So a hard skill could be "this person knows how to code basic programs that works". A soft skill could be "this person knows how to write code that others find easy to understand and modify". The first is much easier to test and measure, the second is really hard to test but is still valuable.
Note that hard/soft has nothing to do with how difficult they are to acquire.
Thus the hardest of skills are when the job requires licensing or similar. If you want to hire a doctor then you need someone who is licensed for it. This requirement is 100% rigid and it is very easy to test if they have a licence, the candidate absolutely needs it so there is no compromise to be done.
Yes. I think the point is that the hard skills required to be a lawyer (e.g. case review, brief writing, contract interpretation, knowledge of the bar exam) are different from the soft skills required (e.g. client management, arguing at trial...).
But not that one is more difficult to acquire than the other.
It borders on saying, "it takes general intelligence to do this, and it takes general intelligence to have a job, therefore people who can do this often can do a job."
"Specific training", as in learning to program in $LANG, is essentially worthless, and if that's all you got from a CS degree, I'd recommend asking for a refund. Teamwork and time management may well be skills that you learn while getting a CS degree, but they're not the focus and not going to be taught.
Coming out of academia, I would say the biggest thing I had to internalize was that nobody in the for-profit world cares about your interesting findings, unless those interesting findings result in more money. So, apart from just being able to write to a format, I don't think grant proposal skills are transferable, and in some sense they may hurt you if you forget that the case you're making is not about research, it's about generating profits.
FWIW, I don't think this a bad thing in many cases. I left academia after I realized that our 4-year NSF project to build collaborative software was never intended to actually improve anyone's life, or even to ever be used outside of our very limited study. It didn't have to be good, or useful, or get adopted by any community in order to succeed. The motivations were completely different: it was to write more papers, and get the next grant, and fund the next set of grad students, so they could build something else that would never be used. Grant proposals all the way down.
Yes, basic research is valuable, obviously, my point is just that it's an entirely different, and largely non-transferable worldview from the private sector.
IMHO an important point here is that in much of the story as you've described it, there are incredibly valuable skills that just need to be re-dressed. As an academician who was lured to industry by an interesting startup and has worked the gamut of positions from IC to executive, I'm reasonably confident in telling folks that it's all about perspective. If things look upside down, then turn yourself to match! For example:
> I don't think grant proposal skills are transferable,
Well, they actually probably are. If you work as a domain expert on the business or technical side of an organization, you may need to directly engage with funding agencies, other companies, or some other agent where your ability to (a) understand the audience and what is required to persuade them to support your idea, and (b) quickly execute on the content required to generate the proposal. Extra points if you were one of those pre-tenure faculty who needed to submit a half-dozen grant proposals every quarter, because you've likely built some intuition as to where it's safe to cut corners in your writing and pitching. Sure, in most jobs you may not be writing NSF proposals every week, but the general skill of effective and succinct technical communication is beyond valuable across a wide swath of applications you'll find in industry.
> The motivations were completely different: it was to write more papers, and get the next grant
... but that's, in a sense, the same thing as the start-up hustle, is it not? Build the MVP, get that one extra client - not because this next iteration of your product is going to change the world, but it's one more notch on your belt that will look good in the next funding round. Hell, maybe the majority of what you're building on any given iteration is throw-away work, but it's towards an explicit goal. I don't really see the start-up rat race as much different than the grant rat race (for better or for worse - I'm not making a judgment call here).
> the general skill of effective and succinct technical communication is beyond valuable across a wide swath of applications you'll find in industry.
I did not feel that academia rewarded succinct technical communication in any way, and the criteria for evaluating effectiveness were very different than in the private sector. The kind of cases you need to make to pitch to investors (or even product owners) are different than the kinds of cases you need to make to pitch to a grant committee.
In both domains, there is a language you need to learn to communicate effectively, and I suppose that's partly what I was thinking of when I cryptically mentioned "writing to a template". But, the languages are not the same, and in fact you have to unlearn some ingrained concepts when you move from academic to private sector — or vice versa, I'm sure. At least I did.
> ... but that's, in a sense, the same thing as the start-up hustle, is it not? Build the MVP, get that one extra client...
No, I don't think it's the same. They both have incentive structures, but the incentives are different. It's good for private businesses to be judged on whether they make more money than what was invested in them, but it would be pretty catastrophic to our culture if academic research projects were.
I think it's just too hard to generalize. PhDs are still relatively few and far between, come from different disciplines, have different interests, strengths, weaknesses. We're not exactly children when we start a PhD program, so our trajectory is influenced by what we've already done. We come from 100 countries.
An odd thing about physics is that we don't have a lot of our own tools or techniques. We're opportunists and we borrow ideas from everywhere. My thesis involved nascent laser technology, mechanics, electronics, theory, programming, data analysis, and a bit of chemistry. Historically, physicists were often the pioneering users of those things before they became branches of engineering.
Today, my work still tends to be multidisciplinary. I do a lot of programming, but I've never wanted a software development job per se. I'm willing to work on problems that are ill-defined, that don't fall into anybody's field, and that might not be solvable. Almost all of my work is heavily quantitative. I interact regularly with management, customers, academics, etc. The people at my workplace who are doing that kind of stuff are, by and large, PhDs. They are not necessarily the highest paid employees -- the coders probably have us beat.
What I don't know (if I were to anticipate reasonable skepticism) is whether this is a special feature of PhDs, or if we move into this role by necessity due to the lack of marketable hard skills. I don't know if any of my skills, taken in isolation, would be enough to get me a job.
Premise: A person spends 5 - 6 years doing a focused task; and the resulting skills are transferable to other fields.
This premise is not surprising. As others pointed out, you can be doing anything and learn something that is valuable and transferable.
The real question is whether a PhD program is the most time- and energy-efficient way of learning those transferable skills. If you were doing your own startup for 5 - 6 years, would you have learned less transferable skills? How about 3-year tenures at 2 of FAANG? What is the opportunity cost of a PhD?
It’s really best to think about a PhD as an apprenticeship to run a research lab. If you go into it with that expectation, you’re not bound to be disappointed. Any other perceptions may leave you a bit disillusioned.
The true value of a PhD is that you get to spend about 8 years or so learning under the close tutelage of a distinguished individual in some field. And really, that’s how you should pick your program — more than the school or the program or the research topic, you have to pick a researcher you want to work with; your brain is going to end up working a lot like theirs by the time you defend your dissertation.
Because what they are going to do is teach you everything they know. It’s incredible, you will feel like a novice at the beginning, but there comes a point a few years in when you realize that actually, you know about as much about this topic now as someone considered an expert in the field. In fact, you are writing their papers that they are publishing. And actually, you disagree with some aspects here and there. Then suddenly, you start having ideas of your own, ideas that you know no one else has thought of before. So you write them down, and that is how a researcher is born.
At this point the idea is that you are kicked from the nest and you will set up a research startup at some other university. This is much like running any other startup but the product is research papers and the revenue is federal grant money.
So to the questions at hand:
1. Are there skills you learn during a PhD that you can learn elsewhere?
Of course. And maybe sooner. A great deal of time is spent on the dissertation. You may have acquired the necessary skills only a few years in. Many people in CS acquire these skills outside of a PhD program (and wrongly assume they have therefore obviated the degree entirely).
2. What are the opportunity costs of getting a PhD?
Well, quite a lot of money although not as much as you think (due to stipends and tuition remission). But there is a lot of financial upside afterwards depending on the field. But I think if you look at peer groups in their early 30s, PhDs would be on average much poorer but some ramping up to much higher earnings.
3. What are the opportunity costs of not doing a PhD?
You will be locked out of some opportunities for good. Or at least, a PhD is an easy way of unlocking some opportunities. For example being an expert witness. It’s easier to immigrate to some countries if you have a PhD. It’s easier to get visas. There are certain grants you can only qualify for if you have a PhD.
Also people treat you like you know everything about everything. This is a blessing and a curse, because if you’re like me you are wrong a lot. Then people will make remarks like “you have a PhD? How do you not know this?” The flip side is your relatives will love to brag about it and it will make your aunts very proud to tell their peers.
And finally, there is really no other way to get that apprenticeship experience. If you really want to get into a topic, spend the better part of a decade obsessed with it, with no other responsibilities or distractions, then I know of no other place to do that than a PhD. Some startups could be close but really would need a lot of funding and be okay with little to show for 7 years. I don’t know many investors that would be cool with that but maybe there are.
The parent post has some great info, but it's worth noting that it's very specific to lab sciences. PhDs in lots of other fields are quite unlike this.
The weirdest thing for me was that after my Ph.D., people started coming up to me on the street and asking directions. The scary thing was I usually knew the answer. Even when I was in a new city.
I think folks are getting hung up on what is or is not a "hard skill".
It doesn't matter.
Lots of people leave academia with/without a PHD and go to work in places where their degree and the skills/knowledge gained by that pursuit aren't specifically required. How relevant your degree actually is (how much it "helps" you) depends on where you end up and how much you're willing to adapt.
Whatever the case, with some things you're going to be just as novice as a 22 year old entry level candidate (in my case, probably even worse).
I recall leaving my phd program and months later in "a job" encountering my first "project manager". I literally didn't know that this job title even existed. I blurted out to some co-workers "Who is this person and why did they ask everyone to estimate on the spot when they would be done with their piece of the project, and anyway, how could we even estimate something so uncertain?" Everyone just laughed!
...and so began my years long reality-check where I discovered exactly how green the grass was on the "other side" working in corporate jobs.
Public speaking, writing, networking, reading primary sources, and independence are all useful.
I’ve met really good PhDs, and awful ones. Awful ones wanted respect because they had the PhD. Really good ones realized how much they didn’t know and acknowledged that.
A PhDs can be a requirement to get very highly specific jobs (ml at google or robotics at Argo AI) but it’s rubbed me the wrong way when PhDs are a bonus for stepping into leadership, which sometimes seems to occur in government / research labs tied to academia.
I think PhDs often experience a dysfunctional work environment in their formative years during the PhD and then try to translate that to the work place, implementing almost a caste system with an over-emphasis on education and power concentrated at the top.
Also I’ve seen some PhDs discount MBAs or other degrees that aren’t hard science because it’s not what they chose to study (physics or pure CS or Chemistry). I guess the point is that someone can get a PhD and be missing some basic core competencies.
The only people that have ever had problem with me (no degree) training them were PhDs. Most have had a superiority complex of their education over mine. I’m like “sorry they feel that way, but I built most of the software you’re gonna use”. Some PhDs have been super humble and they tend to be the smarter people I’ve met.
Yeah. I’m not sure how an additional 3-4 years in an academic environment early in a career makes a difference in one’s leadership ability 20-30 years down then road.
And humble PhDs realize their limitations, which is required for continued learning.
Most important skill from my PhD is the realization that always being wrong and nothing working for months at a time is actually par for the course instead of something to stress over.
When I hire a Ph.D., I hire someone that has undergone something that I know from own experience requires a lot of curiosity, perseverance, drive, self-motivation, self-management, and an ability to communicate well in writing and orally, to convince adversarial audiences, systematic thinking, analytical problem solving skills - each one of them priceless skills.
Any PhD worth their salt would know that a much more likely reason why their hard skills were valuable had nothing to do with PhDs and everything to do with them as a person being able to complete a PhD.
This is no different than anyone who spends all their time producing a great album, or building a company or programming a game and I could go on.
The PhD if you want to talk about it in generalized terms has no unique properties that can't be accomplished through other means.
> Any PhD worth their salt would know that a much more likely reason why their hard skills were valuable had nothing to do with PhDs and everything to do with them as a person being able to complete a PhD.
I often equate my Ph.D. with the ability to teach myself how to do things. Ex. going from a wet/dry lab biologist with zero experience in C-style languages, to learning Arduino's flavor of Cpp and the PID control library to run process controls for your wet lab biology experiment.
That makes it sound like a pretty terrible value, honestly. That kind of "fake it till you make it" plunging into new domains is something people often learn without spending 6 years fighting academia. In fact, your exact example is the kind of project you'd buy for e.g. a teen that likes computers.
I still the think the main value of a PhD is learning to wade through adversarial bullshit and bureaucracy and actually deliver something unless your career overlaps heavily with your research.
Lots of snarky comments in this thread are missing the point: a lot of academics get tunnel vision.
If you've spent 10 years studying ion transporters in mushrooms, it can feel like you are way over specialized and that getting a job outside of your narrow expertise in academia is impossible. It is not. A trained scientific mind is a valuable commodity in the job marketplace, regardless of your speciality.
As a current PhD student in computational neuroscience, I think most of the academics I interact with tend to drastically overestimate the value of their “data science” skills to real jobs, and underestimate the value of software skills beyond prototyping.
I spent my Ph.D. coding in MATLAB, as my Ph.D. was experimental and MATLAB was mainly used for data analysis and plotting. On the other hand, I had a full conda package with complete documentation up and running by the end of my postdoc. Though this was a side project, the time spent on this was incredibly valuable. I did not have to prove to anyone I was proficient in Python; my GitHub was enough. Additionally, the vast majority of researchers' extent of Python expertise was limited to disjointed Jupyter notebooks - while I had a running package with extensive documentation.
I got 3 job offers just based on my package itself - while very few non-academic jobs were interested in my publications. The fact that I had a few first-authors in reputed journals was enough; nobody was interested in their contents.
At my workplace we assume if you're a good Matlab programmer, you can teach yourself Python in a jiffy. What's a bit tricky is that every resume mentions Matlab and Python. Having a public repo is a useful way to show what you can do.
This is the key. Everyone claims Python/MATLAB proficiency, yet there is honestly a vast competency spectrum. A polished public repo, in my experience, will really help you stand out.
Industry is still moving from excel / spreadsheet modeling to tableau and python and alteryx in the workplace. Knowing python today is knowing excel 20 years ago.
Most companies aren’t ready to do real data science, and the transition is as much an organizational problem as it is a technical one.
What makes difference is the autonomy in the job. Are you hired to work as a cog, or do you have power to influence what you do? In the latter case, data science skills multiply the amount of impact your work has (assuming you have ideas).
It's the difference between, "I explored these 5 scenarios last night, none of them is pans out" vs. "This idea seems interesting. We need three weeks, and small team to explore this idea that might have potential."
My impression is that most companies don't know how to appreciate and take advantage of these data science skills, so it's not so much that the skills aren't as valuable as academics think, it's more that they're not utilized in an effective manner in real jobs.
Honestly it sounds a bit desperate. I'm not dissing the idea of getting a PhD, I think it's a big achievement. But this sort of argument is the kind of thing I heard when I finishing my undergrad, and it wasn't convincing then either.
Grant Applications:
Very specific academic "skill" which is intimately tied to the bureaucracy of the university system of your particular country or region. It also has a lot to do with things that aren't even in the application: who you know and what they think of you. Whatever org you join later will have some other way for funding to come down, and there are so many varieties of org that you'll just have to learn on that job.
Analyzing Data:
That's definitely useful, but there's general and specific. There are lots of people who can't do it, and by doing a PhD you definitely show people that you are capable. But, and it's a huge but, you will be spending a heck of a lot of time poring over the evidence on some very specific area of science, with its own conventions and paradigms. When you go and do something else, that won't be relevant. People will of course see that you are generally intelligent, and that's great, but you'll have to learn whole new fields at your next gig. Your issue then is that it's possible to show general intelligence and ability to learn deep material without doing the whole PhD. In fact plenty of people will look at a Bachelor's and conclude the person is smart enough to learn how derivatives work, or how the advertising business works, and so on.
Presentations:
There's simply no less efficient way to learn presentation skills than to do a hard science PhD. It doesn't make any sense, a PhD takes a lot of work and how much of it is actually doing presentation specifically? Yes, I know they do presentations as part of it. But the kind of presentation skill you want is what your typical smooth talker is good at: confidently being able to talk about something that everyone can relate to. Think Obama or one of those Apple-guy-on-stage things. Are you going to learn that on your PhD? From what I hear, science fields are very very specific and although you are presenting, there's a fair bit of focus on the substance, plus you are presenting to people who are also insiders. If you want to just be less nervous and uhm-and-aw less, there are better ways to practice.
There are of course good reasons to get a PhD, such as a desire to learn that specific field. But the cost/benefit of doing it in order to learn generally applicable skills seems off to me.
I generally agree with your thesis that the case here is desperate. However,
> a PhD takes a lot of work and how much of it is actually doing presentation specifically? Yes, I know they do presentations as part of it. But the kind of presentation skill you want is what your typical smooth talker is good at
Step zero of "smooth talk" style rhetoric is to start with the audience and work backwards to the speech.
Doing/understanding something complex and difficult, and then digesting it down to something that a diverse (if specialized) audience can understand, is quite different from what "typical smooth talkers" do. In fact, it's literally reversed. Your message -- the unvarnished complicated truth -- is fixed.
Giving an eloquent speech to a friendly general audience on an easy topic is a good skill, particularly in sales/fundraising/etc. But it's quite different from the communication skills one learns during a phd.
There are many jobs in which the phd-style communication skills are more important than sales-style communication skills. Generally positions where there's a team of folks working on a "hard" and rigid truth that doesn't care about you. E.g., leading/helping/consulting with a team of technologists to understand and solve a hard problem with a physical device or piece of software. "Uhms" and "ahs" don't matter, eye contact is optional, but cleanly communicating about complex ideas with other experts is necessary.
I think you wrong to separate these things. I have sat through a pretty wide variety of presentations, from non-technical HR fluff no one cares about to deep technical designs involving correctness proofs etc. The main difference between the two is how many people are familiar enough with the topic to be able to get any value from the presentation at all. As far as the skills needed by the speaker to present the content effectively, it's almost exactly the same. Yes you need to understand the material yourself, but if you are boring and articulate poorly and don't understand how to relate the audience, it doesn't matter if you're presenting rocket designs or sales fluff. No one is going to remember it 20 minutes later.
Very specific academic "skill" which is intimately tied to the bureaucracy of the university system of your particular country or region. It also has a lot to do with things that aren't even in the application: who you know and what ? they think of you. Whatever org you join later will have some other way for funding to come down, and there are so many varieties of org that you'll just have to learn on that job."
I think the country/region part you mention is most relevant here, but I disagree with the overall assessment of grant writing being irrelevant. If you work in a start-up or the R&D division of a company, there's still grant money to be had. This lets companies pursue high risk projects while not jeopardizing overall revenue.
Well certainly a large R&D division is similar to a university, but if you go that way, of course it makes sense to have a PhD. The question is whether the skill is applicable to things that aren't basically the same thing.
For start-ups, getting funding does not look like a university grant application, surely? At least I've heard many many versions of what people did, and none of them sounded much like what my PhD friends did.
> Grant Applications: Very specific academic "skill" which is intimately tied to the bureaucracy of the university system of your particular country or region. It also has a lot to do with things that aren't even in the application: who you know and what they think of you.
Honest question: how many grant applications have you written, and to which funding agencies? Your experience does not resemble mine at all.
This is easy to mock -- and others already are -- but there is some element of truth. I've always said that the PhD gave me two valuable things:
1) the demonstrated ability to complete a large, self-directed project from beginning to end, starting with ambiguous, poorly defined goals.
2) a herculean tolerance for bullshit.
Both can be earned elsewhere, but the PhD was definitely a fast path. The first one is still quite rare, in my experience, and I'd call it a "hard skill".
I don't know if anyone can call getting a PhD a "fast path" to these things. It's just a "virtually guaranteed" path.
For example, at Goog most PhDs are hired at L5 "senior" positions. BSc (and many MSc) are typically hired at L3. Assuming they are the kind of person who could have done a PhD, they will get to L5 within 4-5 years, fewer for some, with a job ladder (read "promotion criteria") that reads almost exactly like point 1) and a committee of 10 tired, skeptical people they've never met before deciding whether they've demonstrated it for "sufficiently long" to warrant promotion. All the while, the BSc is making more than your advisor and gaining experience that is 100% relevant to their career.
Not to say that PhD vs "a few years in industry" are equivalent in all respects, of course, but I would not at all say that the PhD is the fast path towards the skills you're calling out here.
Yep. Professors are not only able to get an answer, but able to define the problem. It varies by region and domain for when that is meant to be true, so you need to understand the system a bit better.
Ex: US CS PhDs from the Top 30 are generally like this and a quick CV scan and conv can help verify they achieved that. Conversely, many EU CS PhD stage programs are shorter and for more prepackaged concepts, so the same level of experience does happen.. but later.
Ex: US bio fields often are 'one big paper' for a PhD, so you can be assured of grit and uncertainty tolerance. But for the same reason, you don't know independence until the post doc.
I think this article makes sense and makes a strong case for valuing a PhD if you have already completed one and don't plan to enter academia. You haven't wasted your time. But it doesn't make a strong case for choosing to pursue higher education beyond a proper, legit masters degree.
I know MS degrees have been somewhat cheapened, as they are a quick and easy degree that qualifies for the much higher federal loan limits. A university can enroll lots of MS students, supervise them poorly, and profit. But there are still MS programs with a lot of integrity, where you can learn (provided you seek out the opportunity) to write a grant, publish, and do intense data analysis. In STEM, many masters programs also tend to offer RA and/or TAships, which means you can graduate with minimal debt, and you'll have the opportunity to present at mini-conferences or seminars.
You won't be as good at this as you would if you pursued a PhD, but if you get a job and keep working on it, you will be better off in 4-5 years. That said - if you want to go the academic route, a PhD is so close to essential that I'm happy to give a half hand wave and call it essential.
Those were the only `hard` skills they could come up with? This reads like someone who is trying really hard to validate their life decision to spend many, many years in higher education.
I mean:
> we still say (and write) things such as ‘heuristics’, ‘confirmation bias’ or ‘family-wise error rate’
If you came out of a PhD and think these are challenging concepts to pick up, or that they somehow make you more valuable than your average technical employee, well, I don't know what to tell you.
I generally avoid hiring PhDs onto my teams unless the problem I'm faced with is PRECISELY what they researched. 9 times out of 10 a highly motivated generalist is far more valuable than a PhD.
Do you have examples of how hiring PhDs can go wrong? I have a similar feeling that generalists are usually better unless the problem is exactly what they studied. But I don’t have number nor stories to back this claim
> Do you have examples of how hiring PhDs can go wrong?
Not so much wrong, but there's a curve IMO. This is biased and based on my limited experience.
Writing code or working on production systems sometimes they have trouble working the code bases. They can code, but don't necessarily have developed yet the skills to dig into code. I think everyone has this issue starting. There's also a big difference in code from academia to industry.
Same with problems. We need a solution and a usable one. It doesn't have to be THE theoretical best thing always. It has to work. For example, you can't just not consider constants in complexities in production. Those constants take time.
I don't necessarily agree with the generalists part. I think we all specialize over time and have some of the same issues changing language or even frameworks. We just don't have or know the best way of doing things within it yet.
‘Hard’ skills from our PhDs remain relevant beyond academia
Experience in grant-writing, data analysis and presentation will serve you well, say Samantha Baggott and Jonathan McGuire.
You can say that about anything. My version (which I firmly believe):
‘Hard’ skills from McDonald's remain relevant beyond foodservice
Experience in showing up on time, being trained and prepared, caring about customers, and getting done what must be done will serve you well, says edw519
Edit: Oops, I just noticed a similar (and much more insightful) McDonald's reference in this thread from temporalparts. :-)
Indeed, and this is an outgrowth of the fact that PhD education is relatively unstandardized. If you let people chart their own course, a few will figure out how to slide through without doing anything. Also, it's quite possible for someone to be capable of doing the research and producing the dissertation, but with such abysmal personal skills that they render themselves unemployable.
On the other hand, HN is where I learned that employers think so highly of CS graduates that they find it necessary to give them an exam to find out if they actually learned to program. So I don't think it's unique to PhDs.
A huge proportion of them do fail this exam, though, and the pass rate is not higher for PhDs.
The problem is that most places are not hiring computer scientists. They are hiring programmers. You wouldn't hire a team of structural engineers to actually build your house. You'd hire a construction contractor. If you ask a bunch of structural engineers to install a window, you're not going to be very impressed with the results.
Having both a STEM PhD and been software engineering for decades, I'd say article covers the operational aspect of the PhD, i.e. project management skills. The other half of the PhD is knowing what is a significant chuck of doable research. Newbie PhD students often err in selecting topics too small, too large to do in 3 years, or already have been done. This would then correspond in software engineering to a program managers role in selecting features for the next successful release.
I have a buddy who is a lab tech at a university. He is under-paid and under appreciated.
He has a masters in physics. Really smart guy!
I keep telling him: The math background he has would be a huge asset in CS. He should change careers!
He is very skeptical. He feels like he is too old to start( 35 ).
At the end of the day, if he doesn't want to change that is his business, and I don't want to badger him. I just feel like he would be great as a ML/Data engineer.
> Experience in grant-writing, data analysis and presentation will serve you well, say Samantha Baggott and Jonathan McGuire.
These are in my opinion also soft skills from a PhD program (not 'hard' skills). The hard skills are the scientifically deep results that you learn and/or investigate as part of your research.
My wife has a PhD and I've always been impressed with her skillset. Although she's not quite ready to join "the industry", there's truly a massive gulf between the quality of research, writing, organization and analysis she can produce compared to what I've encountered in the workforce.
One "useful" skill PhDs seem to learn is how to manage a busy non 9-5 schedule while earning grad-school wages. This is a non-trivial and probably useful skill to have for many people, but if I were to earn an PhD and quit academia I sure hope that I would have a 9-5 job with high pay...
Most of these skills are not so useful as an individual contributor, but rather for someone in leadership/c-level. But taking those roles as a fresh graduate is very hard. Most established companies would much rather hire someone with actual experience in those roles
The thing articles like this miss is the problem is not will there be some set of skills gained on a given path, but what skills are gained (and at what opportunity cost) versus alternatives.
(1) Financial Math. Some of the comments here have to do with mathematical finance, e.g., the Black-Scholes model.
Mathematical finance done with theorems and proofs has some strong prerequisites of measure theory with sigma algebras, Markov processes, Brownian motion, the Radon-Nikodym theorem, conditional expectation, stochastic integration, e.g.,
Ioannis Karatzas, Steven E.
Shreve,
Brownian Motion and Stochastic
Calculus,
Robert M. Blumenthal and
Ronald K. Getoor,
Markov Processes and Potential
Theory.
Can also want parabolic partial differential equations and
stochastic dynamic programming.
In my experience, with a pure/applied math Ph.D. and as a prof in an MBA program, nearly no MBA students or profs have those prerequisites. And with my experience in finance, nearly no one in finance has those prerequisites, either.
The people in MBA programs, as students or professors, may have gotten such a background from Cinlar, Shreve, or one of a few more.
Bluntly, not many people, a tiny fraction even in pure math, and a good background in pure math is a prerequisite, actually study stochastic processes and stochastic optimal control with a measure theory prerequisite.
Anyone X with anything like that math background might want to be careful or that background will scare nearly all potential employers, in finance or anything else, and result in X being ostracized and fired or not hired.
(2) Math in MBA Programs. When I was a prof in an MBA program, all the math used was undergraduate level, e.g., with hardly even a calculus prerequisite.
(3) Advanced STEM Educations for Non-Academic Jobs. It appears that an advanced education, e.g., in computer science, can help in getting a non-academic job, but the real help is just passing a filter, having proved oneself, and not aiding in doing the job.
This situation should not be a surprise: Having exceptionally good qualifications will make a person X exceptional and rare, that is, one out of 100+. Then the other 99 can be afraid and resentful and gang up on X, sabotage X with gossip, keep X out of the loop, etc.
(4) Jobs and People with Exceptional Qualifications. It is common for someone X with exceptional qualifications to attempt to use those qualifications to get and do well in a job as an employee. That would mean that person X has a supervisor Y who created the job. But as we can anticipate, it is rare for a supervisor Y to have the exceptional qualifications or to create a job that needs such qualifications. Moreover Y could be uncomfortable, feel threatened by, a subordinate X having and using qualifications Y did not have. In simple terms, commonly Y hires X to apply routine muscle to work conceived by Y. To borrow from a James Bond movie, X is not hired to think and, instead, is hired to do what they are told.
(5) Potential of Exceptional Qualifications. For a person with exceptional qualifications that might be powerful and valuable for a career, don't look for a corresponding job created by someone else and, instead, create the corresponding job for yourself. E.g., do a startup.
(6) Financial Success. If person X works as an employee, then the supervisor of X knows to the last penny just how much money X is making and usually will do their best to keep down the amount. However, if person X owns a business, then for X the customers replace the supervisor, and it's common for the customers not to know even roughly how much money X is making.
So, net, for X with exceptional qualifications that seem powerful and valuable, start and run a successful business and make use of those qualifications. There X can be making a nice fortune, with no supervisor at all and with customers who neither know nor much care how much X is making.
(7) Academic Research Jobs. In academic research, the prof is essentially begging a little money from sources with big money, working hard, and then begging journals to let them give away the results of the work for free. So, we might anticipate that such jobs won't pay very well. I concluded that broadly academic jobs were, for someone who wanted to support a family with financial security, financially irresponsible.
Some of these points are correct, strictly speaking, but the analysis is pretty one-sided, imo. This is very "Anyone who's not an idiot should just found a start-up because 'regular jobs' are full of immature monkeys that can't appreciate you." That hasn't been my experience at all and it sounds more like 'being your own boss' is a way to avoid having any people skills or being able to work as part of a team. There are a ton of great reasons not to found a start-up, even if you could. For example, if you prefer living life instead of working, a start-up is probably not a great choice. If you need stability more than a lotto ticket, a start-up is probably not a great choice. If you want to just work in your field and not spend your time worrying about fundraising and
finance and marketing and hiring and HR and revenue cycles and payroll and .... , then founding a start-up is probably not a great choice. The list goes on...
> This is very "Anyone who's not an idiot should just found a start-up because 'regular jobs' are full of immature monkeys that can't appreciate you." That hasn't been my experience at all and it sounds more like 'being your own boss' is a way to avoid having any people skills or being able to work as part of a team.
Naw, not really: My people skills are not nearly the best, but I have learned a little, have quit being afraid of having really bad skills, and have noticed that a lot of people do really well with skills clearly a lot worse than mine! Some of those people are really obnoxious!
For your "not appreciated", right! That's basic: There is competition for promotions, recognition, etc. Soooo, nearly any time person X does something good, lots of people around them in the organization chart can feel like they are about to lose in the competition and then throw person X out of communications with meetings and memos, engage in destructive gossip, form closed cliques, etc.
This matter of "not appreciated" is not just my imagination or experience but is well recognized in studies of organizations, e.g., the academic subject Organizational Behavior, and what I am describing are cases of what is called goal subordination -- person Y pursues goals of their own advancement in ways that are against the good of the organization, i.e., person Y subordinates the good of the organization for their own good.
Goal subordination is serious, common, pervasive, nearly ubiquitous stuff as anyone with good people skills needs to know.
If person X has some good ideas and is ready, willing, able, and eager to work hard to get some financial security, then, sure, consider founding a business.
In particular, in nearly any US city, if look around in any neighborhood of nice, single-family, detached houses, then will find a large fraction, maybe a majority, of the people making the money owning and running their own business. That was true in the neighborhood I grew up in.
Next door, one brother in a family building a nice business in rebuilt auto parts. Two doors away, a guy with his own scrap iron business. A few more doors down, a guy who went to farmers and contracted for their cotton crop and sold the cotton to textile companies, Johnson and Johnson for medical bandages, etc. Across the street, the guy had been a good beer salesman for a beer company that went out of business but soon was the dominant beer distributor for a big chunk of the state. A little further down, a guy running a trade magazine for the trucking industry. One street over, the main wholesale liquor distributor in the city. Two doors up, in a family that ran a very busy tire shop. Where I was asked to tutor a kid in math, the father ran a business building and putting tanks, for milk, gasoline, water, whatever, on the backs of truck chassis, etc.; clearly the father wanted his son to be able to do the math for the engineering of the tanks! Their house was immaculate, like a picture out of Town and Country -- no dust particle would dare settle! See a pattern here? Net, in the US, owning a business is not rare and not nearly always unreasonably challenging. And, now, a business able to exploit applied math, possibly original, computing, and the Internet has some powerful, valuable advantages over the businesses of those other people living in nice houses and driving late model luxury cars.
Here are some war stories:
(1) Saving FedEx. Twice I saved FedEx from going out of business. The first time, in a mad rush, from my home in Maryland (of course, the FedEx HQ was in Memphis) used the best time-sharing, on an IBM VM/CMS system, to write software to schedule the fleet. The BoD was VERY concerned about the challenge of fleet scheduling, and crucial funding was waiting on a solution. Six weeks later I had ended my teaching of computer science at Georgetown U. and had the program running. One evening a FedEx SVP and I developed a schedule for the whole planned fleet and printed it out. Then next day copies were made and distributed. Our two representatives of BoD member General Dynamics went over the schedule in detail and pronounced
"It's a little tight in a few places but it's flyable.".
The funding was enabled. The comment of the FedEx founder F. Smith was that my work "solved the most important problem facing FedEx." For this, I received a LOT of push-back. The guy who had hired me tried to get me fired. We nearly came to blows. Since I had just solved "the most important problem", these blows were about competition, not my people skills. Uh, the blows were about skills, about my skills with VM/CMS, PL/I, the law of cosines for spherical triangles, some vector algebra, etc. A guy I'd never met and recently from some aerospace company Y said to me for no reason
"We have a lot of people like you at Y".
Now who is low on "people skills"?
(2) Revenue Projections. Later the two General Dynamics guys wanted some revenue projections. No one was really assigned to get the results. People had hopes, intentions, etc. but nothing numerical.
So, I got involved: I started with what we knew, current revenue and revenue for the full planned business with the full fleet. Soooo, the projections were essentially an interpolation between those two. Soooo, how might the growth go, that is, what would be driving the growth? Well, there would be current customers and target customers not yet customers. So, the rate of growth might be proportional to both of these and, thus, their product. So, let y(t) be the daily revenue at time t and let b be the daily revenue for the full planned business. Then from some calculus the rate of growth would be
y'(t) = d/dt y(t)
and for some constant of proportionality k we would have the first order, linear, ordinary differential equation initial value problem
y'(t) = k y(t) ( b - y(t) )
I found the closed form solution. The SVP Planning and I picked a value of k for a reasonable projection, and I drew the graph of the solution.
That was on a Friday. The next day the SVP was traveling, and I was in my office. There was a BoD meeting starting at 8 AM. The first order of business was the graph of projections, and the General Dynamics guys wanted to know how the graph was developed. By 11 AM, no one had an answer. I had NOT been invited to the BoD meeting -- that was due to my social skills, right? I'd guess goal subordination, by everyone from FedEx at the meeting. The General Dynamics guys lost patience with the FedEx C-level, returned to their rented rooms, packed, got plane tickets back to Texas, and as a last chance, returned to the meeting. Another SVP, I'd gotten along well with (must have been his skills and not mine, right!), guessed that I'd done the graph, called me at my office, and asked me to come to the meeting, then scattered. As I arrived, the two General Dynamics guys were in the hall, unhappy, with their bags packed, about to leave for Texas (that would have been the end of FedEx). I reproduced several points on the curve, didn't review the differential equation, and the General Dynamics guys stayed and FedEx was saved. A guess is that the General Dynamics guys also guessed I'd done the projections and, really, were torqued at the FedEx C-level for keeping me out of the BoD meeting because of uniform across the C-level goal subordination.
I never got a raise, promotion, thank you, hand shake, or even a smile for saving FedEx the second time. My people skills, right? Or goal subordination?
I can go on this way, lots of cases, organizations, projects: Nearly always when I did such good work, I got attacked, usually seriously. You suggest my poor people skills. I suggest goal subordination.
In the past, I worked really hard and often did some really good work. So, instead of putting up with being attacked by others, for doing good work, as in goal subordination, I should run my own business.
I know if you have PhD you (probably) worked really, really, really hard, were able to focus on a single problem for years, can communicate, analyze and so forth. I have enormous respect for PhDs, and believe they can deliver enormous value. But they are not the same kind of hire as an MBA.
What is the difference between a PhD and an MBA? I think to me it captures the single most important attribute that you cannot pick up directly: what people care about.
The thing about an MBA is, they really enjoy this "business stuff." They like thinking about markets, the customer, costs, finance, how to cross-sell, how to avoid dilution and the like. Its not just that the skills are different (though they are), it is that the interests are different.
Source: comp sci PhD student who switched to an MBA.