Don't want to guess at numbers but my gut instinct & experience tells me that very few physicists end up doing physics and most end up in finance, software, and hardware.
803 responses, 297 in potentially full-time positions.
62% (of all physics phds employed at potentially full-time positions) in non-physics field, break down by industry:
Engineering - 20%
Computer software - 14%
Business or Finance - 11%
Other sciences - 8%
Education - 2%
Medical services - 2%
Other - 5%
The computer industry is notoriously unstable with jobs often quite short term in practice, with layoffs common.
Just to be clear, the referenced AIP report starts with the phrase:
Positions accepted by PhD degree recipients following receipt of their
degrees fall into three categories: postdoctoral fellowships, potentially
permanent positions and other temporary positions.
The figure on page two of the AIP report specifically lists a "private sector" block under "potentially permanent".
Tabel 1 on page 3 lists 70 percent of potentially permanent positions as "private sector." Again this is very misleading as private sector positions are potentially permanent only is the sense that there is a small chance that you might continue at the same employer until retirement; there is nothing like a guarantee.
To me "potentially permanent" seems like a serviceable phrase to encompass positions in academia and in industry that do not have an end data built in to them. It contrasts them to "temporary" positions like postdocs in academia or contractors in industry where you know going into the job that your employment ends on a certain date.
The primary audience for a report on outcomes for Ph.D.'s in Physics one year after graduation, getting their Ph.D. degree, is students (and parents of students) evaluating whether to pursue a Ph.D.. Obviously, they should investigate longer term outcomes, but young people often don't, focusing the next step in their life/career.
Potentially permanent position sounds like a general version of "tenure track research job" as another commenter correctly noted below.
Students with little or no work experience often do not have an accurate impression of salaries, working conditions, and other aspects of the work world (or academia), specifically in the private sector computer industry where most Ph.D's in Physics currently end up.
News coverage of computer companies like Google emphasizes many far out research-like projects such as the Google self-driving car, the AlphaGo deep learning project, and so forth. These sounds like academic research, so why wouldn't these companies have comparable positions to tenure track research jobs? Indeed, in very rare cases, they may have such positions.
However, the vast majority of industry jobs are at will full time jobs without a specified end date. Especially in the computer industry, they are quite insecure and often short-lived, very different from what potentially permanent position implies. They are not analogous to tenure track jobs.
My job is described as permanent full time, because my contract is ongoing, and I work 40 hours a week; when I was a student, I had a permanent part time job.
This is in contrast to a casual job, which in New Zealand is similar to at-will (but with strong restrictions on what roles can be casual); or a fixed term contract.
In any case, with the closure or heavy cutbacks at corporate research labs, AIP should update its language to reflect current realities.
For a few years, he started checking the median age of physics professors, and noticed it went up every year instead of down. Tenured profs decided to stay in their seats instead of retiring, and there were no new positions being created for young professors. That was his indication to get out (this was during the mid 90s).
The real crux of the problem is that modern PhD programs were originally designed post WW2, in the era of the GI bill and fast expansion of college education where there was high demand for new professors. That's no longer the case, but we still train PhDs as if they would get a job even though most won't.
Still, it seems low IMHO. Unless you start counting at the point where the person gains full tenure rather than from the point where they become some kind of "group leader", "assistant professor", or such "tenure track" type position?
Why? I know an academic who has done exactly that. Nearing retirement now.
Does physics academia really not have another position you can promote someone to, if you want to keep paying them to do research for you, but they've "timed out" of the postdoc role? I had assumed this kind of research-scientist role wasn't CS-specific, since it's existed for a while, and at big universities, CS hasn't traditionally had the clout to invent completely new job titles, so hiring is usually mapped onto stuff that exists in some form or another at the institutional level.
Eventually you need the money for retirement...
Unfortunately, becoming a developer is a short-term optimal choice that you pay for in the long term.
The 40-year-old physicist has tenure and can work on whatever he wants. As far as his peers are concerned, he's mid-career. The 40-year-old programmer is considering plastic surgery so he can still get hired in the Valley.
The people who become developers find out that they don't have permission to get a day older than 35 unless they can make it into management. And if you were going to be a manager anyway, you might as well have gotten an MBA in your mid-20s, and then you'd be making far more than the developer would even dream of.
Coding was always way more fun than doing 4th year/grad school level physics for me too.
Programming is a lot of fun, but most software jobs aren't programming intensive. The coding is trivial and a high school student could do most of the work. The hard part is dealing with tickets, PMs, and unnecessary meetings.
For me it was also that nothing truly new was being discovered in my area of physics, so it all felt a bit boring.
Plus the politics that comes when post docs have to fight and plot over rare tenured positions when they appear.
I really regret going to grad school for physics, I gave up a lot of earning potential, time to develop new skills and I've forgotten a lot of my CS background. Most methods in my area have already been developed and the pressure to publish (pushy advisor, several projects not producing the intended results, and proposal writing) has always been hampering away at my free time to think. I really dislike this system. At the same time, I'm sure exists in industry too (never had a real job). I'm not going lie, I feel a bit trapped, but I'm still a fairly positive person.
In the sciences, grad school is sold as a necessity but I'm starting to think this is a way of fueling the ivory tower.
I realize that this is a bit incoherent/ranty, but whatever. :)
I know a lot of graduate students and early professors who have a fear of private industry. I think they all know they can make more money outside of academia, but they don't realize they can work fewer hours and not have worry about a bad tenure committee screwing up their careers. A bad boss can make my work life suck, but he'll need to work hard to completely nuke my career.
Anyway, I understand how you feel. Despite my lackluster GPA, a couple professors want me to go on for a masters in math or statistics. You should see their faces contort when I tell them I don't want to go to graduate school or I don't want a career in academia.
You spent your life in academia, surrounded by people who spent their lives in academia, who have surrounded themselves with people who spent their lives in academia. Get some fresh air, fill your bank account, and then decide.
Some people think of this as 2 years for a Masters and 2 more for PhD, but in practice most hard science programs don't really offer a true Masters program. When I got my doctorate, a bunch of us figured out we could apply to get our MS diploma after we finished our coursework -- hardly anyone in the PhD program had ever bothered before. We all got them and put them on our desks as a kind of inside joke.
Although searching Google - I can find references to the usage you're giving.
But, I fully agree that a Masters in a hard science, especially from a top research school, is almost always a consolation prize for those who drop out. My program would not accept students into the program for a Masters, it was PhD only, unless you dropped out.
My impression from talking to a lot of people, is that the European schools simply manage the PhD process to make sure it doesn't drag out too long. It's expected to be shorter, so it's shorter.
The big difference was that the PhD in the UK launched you straight into research on day 1. No teaching, no courses. He spent the first couple of years in courses, and a lot of time throughout the degree in teaching. In some ways I envied his coursework, since I was basically on my own from the start (his grad courses sounded very thorough and interesting). However I do not envy the 5 extra years it took him.
That was all a couple of decades ago now, so things may have changed.
Only 3% of surveyed physics PhDs took 4 years, nobody took less and 6-7 years was average.
I was, relatively, fast at it as well. M.S. was 2 years. Ph.D. was 4 years of research, 3 years of writing. Though I spent the last 2 years of writing ABD working full time for a "startup".
I really enjoyed my time in grad school. I was a computational physicist, doing molecular dynamics. The reason I went into academia was lack of professor positions . The writing was on the wall. So I decided to go the industrial route.
With the impending closure of my latest venture, I am looking around at options, and am actively interviewing. As it turns out, it looks like my thesis advisor is retiring at my alma mater, and they have an open position.
I am thinking deeply as to whether or not I really want to do physics. The pay is terrible, and I am marked as an "undesirable" as I had left academia.
I'll keep looking for my next place. This said, the Ph.D. was a good thing to do. I learned a tremendous amount, not just about a narrow area, but how to compute, measure, and reason about things that are complex.
Seems like this is in demand now. Makes me happy.
One thing - if you decide to move to industry, make sure the first years you can learn from someone more talented than you - even if informally. This is critical, software is a craft.
I spent 7 years in grad school, then quit without a PhD. I don't regret it for a minute. I managed to spend 7 years studying topics that interest me, that I will not get to study had I gone straight into industry. Sure, while working you may have some spare time to study some math/physics. But nowhere near at the level you could/did in grad school. There's really no comparison.
If you're not happy in grad school, perhaps:
1. You don't like your research topic.
2. You don't like your advisor.
As an example, I did not have a very pushy advisor, so I learned what I did in a very relaxed manner. Bad idea if I were to end up in academia as a career, but it worked out great for me.
It does suck not finishing the PhD, but once I was in industry, I saw the jobs that most PhDs in my research topic would have ended up with - and most of those jobs are horrible. I worked with them for 4 years, and moved into programming. While the satisfaction of solving really challenging problems is no longer there, the programming job overall really is much better: More autonomy, more creativity, better work schedule, etc.
But bottom line: Grad school is for learning the stuff you are passionate about. If you're not doing that, either change your topic/advisor, or leave.
Sounds like you have plenty of "real job" experience ;-)
"Study physics now. Because later when you have a job, you might get paid to learn using some software tool or a new programming language. But no company will ever send you into a quantum mechanics lecture. "
He was right.
Both of my parents started their careers as scientists in industry after getting graduate degrees in the 50s. They saw people with science background being able to go into practically anything, including programming, business, entrepreneurship, and so forth.
I learned programming in high school, and fell in love with it, but I had an internship at a computer facility, and also looked ahead at the typical CS curriculum. It all seemed terribly boring. So I majored in math and physics. Oddly enough, the people who were doing things with computers, that interested me, were in the physics department. I developed the ability to design computerized electronics for measurement and control systems -- which became my career. This was even before "embedded systems" was widely taught in EE departments.
But realistically, a large portion of the software industry today does not require people with a science background. What I don't know is if I'd still find it boring.
Perhaps my parents' attitude was along the line of "you can do anything with a liberal arts education," but with the stipulation that the liberal arts include math and science.
I bet D-Wave has some interesting quantum mechanics lectures!
A high profile example is Emanuel Derman, author of My Life as a Quant (2004) and later books, who worked at Bell Labs from 1980 to 1985 before moving on to Wall Street. He mentions quite a number of other physicists at Bell Labs at the same time.
Most physicists end up in some sort of software development. The high profile "quant" jobs are actually rather rare and hard to get. The Wall Street firms are typically going after very strong physicists, especially theoretical physicists like Derman.
Nathan Myhrvold of Microsoft and Intellectual Ventures fame (or infamy) has a Ph.D. in theoretical physics from Princeton. Did not go to Wall Street. :-)
The Large Hadron Collider (LHC) produced a huge surplus of experimental particle physics (high energy physics) Ph.D.'s with no jobs in the field. Experimental particle physics involves large amounts of software development for data acquisition, instrument monitoring and control, and data analysis, mostly in C and C++, although there is still some "legacy" FORTRAN software. The heyday of FORTRAN in physics was a long time ago.
Although there have been attempts to use neural networks and other machine learning methods in particle physics, the workhorse of data analysis in the field is Ronald Fisher's maximum likelihood estimation and classification -- primarily estimation of parameters such as the mass and width of the Higgs Boson. The discovery of the Higgs was a maximum likelihood analysis.
Although it is undoubtedly possible to map maximum likelihood onto neural networks, in practice they are different. Neural networks are an attempt to simulate the low level structure of the neurons in the brain and solve problems by brute force fitting of data to models with huge numbers of adjustable parameters. In contrast, maximum likelihood involves attempts to understand the phenomenon under study and model it as a small number of functions corresponding to higher level concepts such as the Higgs Boson. A neural net could exactly fit the Higgs Boson peak yet never produce or confirm a physical model of what causes the peak.
Ahh, how I wish that were true...
Degree in physics, working software for 20 years and still feel like an impostor ... not that I remember any physics at this stage
We had a long discussion during which I told him that I'd always wanted to program computers, but didn't think it made sense to get a degree doing something I could learn in my bedroom. Physics, on the other hand, was fascinating, and could only be truly learnt and appreciated from people who'd devoted their lives to it.
20 years later, I think those years of training in physics have made me a better developer, and taught me to better see patterns in data. To prefer "good enough for the task at hand and elegant enough to be proud of" over "perfect" (which is what many of my math friends ended up seeking).
On to your question: It's hard to pin down the exact root cause, but it seems (in my case at least) to stem from being an outsider -- a y in a sea of x's. When everyone else doing your job has a specific attribute that you don't, you can start to wonder whether you really belong in that role. The literal feeling is, "One of these days, I'll say the wrong acronym in stand-up, the rest of the team is going to figure out that I'm just winging this Agile Scrum thing, and that will be that."
That can mean being the only physicist in a company of CS grads, or the only woman on an all-male engineering team.
Being in a job where you're always learning, received no formal training, and where the expectations are fluid can amplify the feeling, since there's no yardstick against which to compare your performance. If you're a fast learner, it can be hard to believe you've gotten as good at something as people who had years of training/have been doing it for years.
Here's a group of people. Compared to each one of them, I know less about something. It's really easy to go from that to feeling like I know less about what I'm supposed to be doing than everyone else.
I don't really experience this, I suspect because of arrogance. I'm not sure that's really an improvement, though...
> I can't understand something so it must not be real
Maybe that's a problem with you more than reality.
I never said that, don't twist my words. I said it seems bogus because I've never experienced it and I don't understand it. I never said it isn't real - it probably is; I just have a hard time relating to it and accepting it.
While true as far as our current mathematical models are concerned, there are good reasons to expect that deep down the world's fabric is discrete. (Whether that counts as being "digital" is probably just a matter of language.)
There are no reasons to expect one way or the other. As far as we can measure, spacetime is discrete. However we can't probe very far in the grand scheme of things. Without experimental evidence, and without a particularly strong argument one way or the other, it's useless to be speculating.
Did you mean to say "continuous"?
Quoting Wikipedia , "physical spacetime is expected to be quantum because physical coordinates are already slightly noncommutative."
... until 20 years into my career, when I got hired by that medical instruments place that was doing 3D reconstructions of patient anatomy from a series of 2D X-rays. Then it was 3D coordinate transforms, Fourier transforms, some other calculus, plus radiation physics.
A lot of stuff you learn isn't relevant to your career - until it is.
I've done a fair amount of hiring and I know that a candidate with a physics or mathematics degree who also had a programming background that was commensurate with the position would look a lot better to me than one with a compsci degree, all else being equal.
That should tell you what college degrees are worth these days, though. The candidate primarily needs commensurate background. Degrees are effectively neutral -- too widespread to penalize people for taking that route, but really not worth much extra, unless in the course of your studies you accidentally did something really interesting, but then you probably could've done interesting things without the educational institution.
Was thinking of pursuing masters/PhD in economics but have heard/read a lot along the same lines of an intellectual pursuit with diminishing returns both in terms of salary and real world impact.
The most useful thing I've done on Twitter is follow a bunch of economists and professors who are constantly debating current econ thinking/papers. There are a few I really respect, but for the most part both the academic environment for PhD level economists seems kind of toxic to me as an outsider.
Been thinking a lot recently about how my generation (I'm 24) is going to have to be a lot more practical due to economic realities...
At one point, I mentioned to the professor that I was concerned that the model he had presented was overfitting, and he had no understanding of what the term meant.
I think that economics studies fascinating problems (how do people make choices? What are the optimal choices for policy makers to make?) but economists approach the problem completely wrong.
Dropout, for instance, or k-fold cross validation.
We are 20 years out of NAFTA and free trade with China. I see a lot of journeymen/trades that do well in union gigs, police/fire/teachers and blue collar professionals... Those who aren't doing well are people outside of this (service/retail) and illegal immigrants. I guess manufacturing was a union gig that one could get into... On a side note.. why does a Tahoe cost 70k... I mean I know people pay that but seriously... unreal. A Tesla too.. crazy expensive.
He starts out strong and gives a good overview of the "schools" that have been laid to rest, but concludes on some kind of utopian, happy note that I don't believe he has any evidence to support. But then again he spends most of his time in conversation with leading economists (I think) so he should have a pretty good idea of what's up.
Paul Romer has a good summary of the state of macro in his 9/16 paper "The Trouble With Macroeconomics". Here's the abstract:
For more than three decades, macroeconomics has gone backwards. The
treatment of identification now is no more credible than in the early 1970s
but escapes challenge because it is so much more opaque. Macroeconomic
theorists dismiss mere facts by feigning an obtuse ignorance about such simple
assertions as "tight monetary policy can cause a recession." Their models
attribute fluctuations in aggregate variables to imaginary causal forces that
are not influenced by the action that any person takes. A parallel with string
theory from physics hints at a general failure mode of science that is triggered
when respect for highly regarded leaders evolves into a deference to authority
that displaces objective fact from its position as the ultimate determinant of
It's a good one...https://paulromer.net/wp-content/uploads/2016/09/WP-Trouble....
Personally, just from reading books and papers it seems like there are very few economic clans left, just economic celebrities (Krugman, Varoufakis, etc.) and the profession's credibility is suffering due to the Fed (and other leading central banks) inability to make anything happen with an empty clip monetary policy wise.
Tim Duy: https://twitter.com/TimDuy
Claudia Sahm: https://twitter.com/Claudia_Sahm
Stephen Williamson: https://twitter.com/1954swilliamson
David Andolfatto: https://twitter.com/dandolfa
Tony Yates: https://twitter.com/t0nyyates
Neel Kashkari: https://twitter.com/neelkashkari
Noah Smith: https://twitter.com/Noahpinion
Tyler Cowen: https://twitter.com/tylercowen
Justin Wolfers: https://twitter.com/JustinWolfers
Paul Romer: https://twitter.com/paulmromer
Jan Zilinsky: https://twitter.com/janzilinsky
I actually have often found it awkward to fit into the usual computer science bins that companies organize projects into. In addition to a hefty amount of data analysis and software engineering, my PhD has required knowledge from a wide variety of different types of engineering, including optics, microwave engineering, and semiconductor fabrication. Combining everything into a 1-page resume, I've found it's not super obvious where my placement would be within most companies. I am currently in the team-matching phase for a google internship, for example, but haven't heard anything.
For those that are getting themselves in the door, do you have any tips?
The best thing getting my food in the door in silicon valley was to move there; quite a few opportunities for qualified candidates afterwards. Even research labs at Stanford or Cal still value hiring the broad experiences you describe and could be a foot in the door.
I have no tips, but a guide to dealing with recruiters (who are typically pattern matching for the framework du jour) for developers from a (hard) science background would be really cool!
Yes, Google does interesting science research as well as hiring physicists, though we have plenty of those, too. I'm one of four on my team.
I know many physicists who became hardware engineers and even software engineers when they needed it. I know no software/hardware engineer who became physicist.
An iOS/Android app for Three Body simulations (Three Body) and a Unity gravity simulator (Gravity Engine).
I would have loved to make a living wage in physics but the post-doc path is badly paid.
Even nanomanipulators are DIY: http://www.popsci.com/diy/article/2011-07/you-built-what-sca... or http://hackaday.com/2015/01/13/cheap-diy-microscope-sees-ind...
(ESM can be used to manipulate atoms)
Yet I haven't had time or courage. Or badassness.
For many, it's because they never needed it.
Sure, due to the premise of the article; no pay in it.
John Carmack was wealthy enough to afford spaceship company.
This is closest one get to physicist. And yet - no quantum or particle physics, theoretical or experimental.
I much rather hire people that I know can think in both in terms of fundamental logical principles, and who understand the scientific method, which comes up more than one expects in business. For example, I've seen physicists run much better marketing analytics than "growth hackers".
And who cares if someone can cross-off a list of generic programming language/frameworks? If you need that specific of a cog in your machine, the position probably isn't that innovative and you're better off outsourcing. Or hire someone smart and eager, pay them decently, and they'll learn what they need, if they're so inclined.
If you have the top 2-3 things they'd need (e.g. CS algorithms knowledge + some scripting language + some database experience), and are otherwise a good applicant, go ahead and apply with the mindset being that you'd have no problem picking up other frameworks X,Y,Z. Of course, spend an evening to take a look at them and be confident enough to make that the case.
I'm studying physics because I want to help shed light on the mysteries of reality.
It might be hard to get a job in physics and achieve that, but it's a hell of a lot harder to get in a job in software development and achieve that.
Despite this, all I'm reading is optimism about how well-paying and interesting software development is for physicists. So what? If those were my primary concerns, I wouldn't have studied physics in the first place!
I get the hidden impression that the meaningfulness of science -- pursuing the truth, the nature of the universe -- is being swept under the rug because it's no longer paying the bills. That's a goddamn tragedy, not the cause for celebration this article is making it out to be.
Von Neumann, Feynman, Metropolis, Vapnik, Babagge, Lovelace, Turing, Church, Hopper, Brooks...
(Was within a hair of a physics minor before I studied C.S., but might have wiped out the advantage by switching to Mathematics?)
AIP does many nice studies of where physics PhD's end up.
One of the hardest challenge as a Physics PhD in industry is finding interesting problems. See fig 2.2 for reporting on that: https://www.aip.org/sites/default/files/statistics/phd-plus-...
Does this mean that only physics/math/stats majors are getting into this field? At my current place of employment this is the case for the data scientists; all have an academic background in one of these three fields, all have a Masters or PhD. I assumed this was because the head data scientist had a long history with academia and thus had a predilection for academics.
From my experience the data science concepts I have encountered thus-far seem pretty straightforward and I suspect that people are trying to make these concepts seem more difficult/arcane than they are by using notations/concepts only learned in academia. Am I wrong here? Is it actually a field which requires a PhD in Math to understand? My exposure has only been with toy examples (logistic regression, simple perceptrons, similarity algorithms, etc). How easy is it to get into the field without a heavy academic background?
edit: this is just a tongue in cheek comment that physicists will rule silicon valley with about the same chance of success matlab will be used for enterprise dev.
As far as the machine learning market goes, 90% of the projects require software engineering skills, the last 10% requires being able to go underneath the covers of linear algebra libraries, etc.
I just think the whole physics>cs degree for machine learning argument is not totally persuasive given my experience.
Now for the past 5 years, I am highly appreciative of such people because each of them bring knowledge and interesting viewpoints from their respective academic/training backgrounds which serves to enrich the IT world. Thank you guys :-)
So this article comes all the way to this. Is there data that hiring of physicists speeded up in the recent years? And why roles they are taken? Presumably not all in machine learning, if I have to guess.
Until then, this is another effort to turn an otherwise good interview into an unsubstantiated editorial aiming for nothing but hype.
I got my undergrad in physics, but spent all of my free time (and a lot of what probably should have been study-time) teaching myself to code. It had a pretty brutal impact on my GPA, but I've no regrets.
This eventually drove me to take a year off and then resume as a Physics major. I'm almost done now with my BS and am so happy I made this choice.
Despite switching to a Physics major, I'll probably end up working in a programming position eventually. It's a far easier career to get into.
(I'm a CS PhD and I know many physics PhDs as well.)
Also, physicists working for software is not only a silicon-valley phenomenon .
Turns out he got a PhD in theoretical physics and did some graduate work at Fermilab.
Holy crap...this run of the mill Hadoop administrator is some kind of physics genius.
I asked him why....he said that his true love, particle physics, just didn't pay the bills like any random IT related job does.
Which is a damn shame.
PS: While we were talking, a dude one row over popped up and said he had a PhD in Chemistry or something.
I, for one, welcome our new physicist overlords!
Not a physicist and not that there's a lack of open jobs, but I was also way more interested in programming.
I started my own project in 2nd year of university and it changed everything to me. I've been able to apply my knowledge during my years of university and also out of it with my side-projects.
All my best developers over the past 25 years I've managed were either physicists or mathematicians.
I would do open ended research if there were any jobs going, but there generally aren't. And I'd be just as happy in computer science as physics.
Physics as a career doesn't pay terribly well, but as a degree offers a very wide breadth in coursework and skills. Grad students in particular usually have experience in HPC, for instance.
And soon we will rule the world, muhahaha...!
"Coders" - sure, but not software engineers.
Incidentally, the skill of "coding" (i.e. using a programming language to automate tasks on a computer) should, in my opinion, be taught in school at an early age along with the multiplication table.