Leaving Academia (2013)
jakobegger.at
jakobegger.at
I was in academia. Not just a grad student, but achieved the phd and did a post doc. I don't regret any of it. It taught me many invaluable lessons, all of which have been useful in industry starting a company.
My impression of industry so far: they don't get progress. They don't get innovation. They have problems, but they they don't understand them and are unable to commit resources to the procedures that are needed to solve them.
Only the very biggest of tech companies like Google, Palintir, Apple, Facebook, are able to tackle the really hard problem. These are the problems that you can not make an estimate for in terms of completion. The majority of industry will not even entertain this class of problem.
Academia, too, has a lot of problems. I left because of the publish or perish mentality. I liked to work on hard problems, and had some successes, but overall it is the numbers that matter not the difficulty. There are an equal number of politics in academia as in industry, though they are different in academia.
In industry, politics often involves a lot of incompetent people in high places following paper work procedures that do not support their supposed cause, while in academia it is more a lot of infighting over government grants and insultingly simple metrics like the number of papers published.
In short, the author of this post is full of ignorance, and I am only slightly less full of ignorance in that I haven't achieved tenure in academia, and only have several years in industry helping to run a company. None the less, what I learned in academia in terms of the science of computing, it was invaluable.
"Only the very biggest of tech companies like Google, Palintir, Apple, Facebook, are able to tackle the really hard problem."
Wrong and wrong. First, all of those companies made concrete progress on hard problems when their size was less than 10 people. Woz pretty much singlehandedly designed the first PC for normal people. Zuckerberg built a pile of code that would become one of the largest social networks.
Let's talk about the "really hard problem". Let's talk about John Carmack, friend.
Made the first realtime 3D fps games, the networked dialup realtime 3D games, the first networked shooters over the internet, the first true commercial 3D engine, and is basically responsible for the fact we have GPUs in consumer desktops now. Dude only had two (2) semesters of college under his belt, and he runs rings around the academic CS wanks, because he fucking ships code that works. No cutesy benchmarks. No government funding. No papers, barring perhaps some stuff in Graphics Programming Black Book, none of that. All of that stuff running on consumer hardware without graphics accelerators.
There are plenty of hard problems being tackled by small teams, even on shoestring budgets. To say otherwise is magical thinking.
~
If you want to publish a paper, you don't have to actually be in academia. If you want to do research, you don't have to be in grad school. In fact, and hopefully more people will get in on this, you will probably do better work in a healthier environment.
What does academia get you? It seems in the vast majority of times, it gets you...
...an "advisor" who is more concerned with finding funding and keeping lab space than making progress.
...an "advisor" whose last original contribution to the field was, almost by definition, probably a decade ago.
...a chance to talk with world experts in your field--the same chance available to any schmuck with an email account.
...a chance to submit papers whose only circulation will be behind paywalls.
...a chance to be woefully underpaid for the above, because they know you are stupid and naive enough to take it.
Fuck that. The author was smart to get out while they were only slightly behind.
In one project I worked on, I developed a new algorithm. Or so I thought. Then I went to the chemistry library about an hour away to read the papers they have on the shelves, available at no cost. Turns out the basic approach was done in the 1970s! Luckily, further reading showed how what I did really was something new.
In another project, I managed to dig up another paper from the 1970s which solved a problem I had been working on for a couple of years.
I am self-employed, and don't have the $30/paper just to find out that it's worthless for what I'm doing.
It also gives you cheaper rates for conferences. As a "commercial" attendee, I pay a lot more than academics, even though I make less than most academics.
In many cases it also gives you physical access to those world experts, who may visit your institution because of your advisor.
I think that sums up why so many people hate being in academia. They go into it for the wrong reasons. I've seen many people make this mistake.
If you go into contemporary academia with the expectation that you will not get a TT job, or a TT job in a desirable place to live, at the end, then perhaps you are doing it right.
Right now there are simply far more PhDs than TT jobs. That means a large number of PhDs cannot get the jobs for which they are being trained. People entering should know that. Knowing that is not equivalent to pursuing academia for "fame and fortune."
I wrote about this in a humanities context here: http://jakeseliger.com/2012/05/22/what-you-should-know-befor...
Amen. It's kind of amazing what academics put up with.
Take relocation. I'm 31 and while I'm at the top in terms of talent, I've made some mistakes and I'm probably only upper-middle in terms of tech career success, and I would simply not take a job, with a cross-country move, that didn't offer a full relocation package. Yet academics who are at much higher quantiles of success in their industry (just to have a TT job is 85th or higher, these days) are happy just to have a salary, because their low self-confidence ("impostor syndrome") is used against them to the point where they think it's egregious just to ask for basically decent treatment.
A lot of the PhD's in pure sciences in the US are from foreign countries, especially India and China. Actually, foreign students dominate in most engineering/natural science departments of US universities. So, I guess most American kids do realize that its not a valuable decision to pursue a PhD.
> If I'm not going to grad school because of job opportunities, what other reasons are there? Pursuing my interests? As a PhD student you get to choose some interesting problem and work on that, right? That's what I naively imagined before I started at the institute. Some day at lunch I told my supervisor about this idea I had. We could take his fluid simulation method from computer graphics and apply it to a problem related to molten polymers. There was this experiment by a group of applied physicists that would fit nicely. He asked me how many people would be interested in the problem. Maybe a handful, I said. And then I realized that there was no way I could work on that problem. Success in academia is measured in the number of citations your paper receives. What point is there in writing a paper that is only interesting to such a small audience? To be successful you need to target a large audience, and not just pursue whatever obscure problem takes your fancy.
If you aren't paid well, and you're not well respected, and there is no great job waiting for you at the end of the tunnel, and you can't even work on what you're most interested in, I can't really blame anybody for leaving.
I hope more kids in College and high school realize that a) working physics professionally is very different from what is portrayed in popular science and b) you really have to spend a lot of time thinking about the future in these kind of things.
That said, I still love popular science: interstellar made me want to study Relativity again :).
My impression of industry so far: they don't get progress. They don't get innovation. They have problems, but they they don't understand them and are unable to commit resources to the procedures that are needed to solve them.
Academia, too, has a lot of problems. I left because of the publish or perish mentality. I liked to work on hard problems, and had some successes, but overall it is the numbers that matter not the difficulty. There are an equal number of politics in academia as in industry, though they are different in academia.
In industry, politics often involves a lot of incompetent people in high places following paper work procedures that do not support their supposed cause, while in academia it is more a lot of infighting over government grants and insultingly simple metrics like the number of papers published.
Academia and industry both have the same root problem and it's this: Those Who Can, Do; Those Who Can't, Evaluate. We see them as different cultures with separate sets of values, but the fact is that people who are good at office politics tend not to be good at anything else, so Decisions get made by incompetents. I don't think there's a scalable organizational antidote, because people who are politically adept and competent are incredibly rare.
I think that the power players in academia are even more egregiously incompetent. They often lack basic social and leadership skills. Academia has been dying for 35+ years because its leadership is completely inept at fighting for its people against the cost-cutting mediocrities. Say what you will about MBAs, but I'd rather have one fighting against HR for my budget than an ex-academic. However, the professors are high-IQ incompetents-- though often not that smart, even still-- which makes the cerebral narcissist in all of us a little bit more sympathetic to them.
The worst thing about the "cost-cutting mediocrities" is the cost-cutting doesn't extend to themselves:
"Forty years ago, America’s colleges employed more professors than administrators. The efforts of 446,830 professors were supported by 268,952 administrators and staffers. Over the past four decades, though, the number of full-time professors or “full-time equivalents”—that is, slots filled by two or more part-time faculty members whose combined hours equal those of a full-timer—increased slightly more than 50 percent. That percentage is comparable to the growth in student enrollments during the same time period. But the number of administrators and administrative staffers employed by those schools increased by an astonishing 85 percent and 240 percent, respectively."[1]
[1] http://www.washingtonmonthly.com/magazine/septemberoctober_2...
It seems to me that we should stop standing on each side of the fence and looking down on each other, and start working together. We might actually achieve something great it if we do that.
Regards the OP, if it was the right time for him/her to leave, that's great. Nobody should continue on a path that they do not find fulfilling. It says nothing about academia, or industry though. It only reveals the feelings of the author.
The true travesty here, is that he/she will find it extremely difficult to pick up that work again and find a place in a university if they change their mind in the future. Our system in the UK is broken, since you only get a single shot at academia -- unless of course you are independently wealthy.
Perhaps this is the cause of the divide. You are either in, or you are out -- and once you make your choice there is no going back.
I wanted to say at first that if he was in it for the sake of being a successful, high-impact researcher, he was in science for the wrong reasons. But really, unless you are one of those high-impact people, you aren't getting tenure. I don't feel like tenure is asking for much if you're a hard worker who's produced some novel results. I get that science is supposed to be about passion, but you can't support a family on passion alone.
> A career in academia would just require even more proving my worthiness...I'd constantly be evaluated. That's not what I was looking for.
He's in for a rude awakening. You're only as good as your last envelope.
Also, of course you are evaluated in anything in life, from the desk to the friendly get-together. The issue with academia has always been its ROI. You invest a lot into the academic life, your time, your youth, your energy, your mental state, all to graduate close to 30 make a salary some of the people you tutor make fresh out of college? All while working 50-60 hour weeks...it isn't quite the life.
Perhaps being in a start-up requires such dedication too, but that can be seen a risk that can yield a great reward. For academe, there is no reward, save the "academic points."
This [1] says it well. At least if you work hard and stress and have no life in the "real world", you'd have the chance of monetary compensation. Really, in academia, unless you are lucky enough to become a tenure prof, all you have is your pride that you understand something no else cares to. That, and the occasional free pizza and beer.
[1] http://100rsns.blogspot.com/2010/10/25-academe-is-built-on-p...
I disagree. First of all, there is a rather high place in the societal ladder. A university professor is still somebody who induces awe, plus the job has this "uncorrupted by money" vibe which you cannot get as a manager/entrepreneur. Furthermore, you have an extremely high position in the local society, i.e. at the university.
But I think the real reward is not vain or materialistic -- the reward is doing something you adore, and something you find a lot of meaning in. Both teaching and research are extremely fulfilling for me (as a PhD student). Personally, I can think of other jobs in compsci/programming that would be fulfilling, but they are likely as unavailable as the university position.
And no, the jobs you are thinking of are certainly not just as unavailable. For fuck's sake man, the number of university positions in the US for my field, physics, is in the single thousands...universities themselves have 3000 jobs in total[0]. That's total number...and this is with 1700 PhD's awarded per year[1]. It looks a little better for you in compsci [2],[3], so may be you have a right to be more optimistic than me.
Perhaps an important thing here is the area of study. Physics, IMO, is pretty bleak, especially for theorists.
[0] http://www.bls.gov/oes/current/oes192012.htm
[1] http://www.aip.org/sites/default/files/statistics/rosters/ph...
[2] http://www.bls.gov/oes/current/naics4_611300.htm#15-0000
[3] http://cra.org/uploads/documents/resources/crndocs/2013-Taul...
It really depends on what you can see yourself doing. My closest compsci friends are all employed by either big companies working on their backend or by "young hip" companies doing web solutions for the local tobacco companies and the like.
Anything that I would like to do would have to (subjectively) help humanity as much as research/teaching; and getting jobs like these either requires a whole lot of dedication (leaving your family behind, moving to Asia/Africa) or a lot of connections (governmental public good jobs are rare for compsci people).
This really depends on where you are looking for tenure. If you're at, say, a top-30 research-focused institution, then yes, that's what they angle for: big grants and "high-impact research". But if you enjoy and/or are good at teaching (obviously not all researchers do or are), there are a lot of places where getting tenure is more a matter of doing some interesting research, combined with teaching at least fairly well, and being perceived as an asset to your department/university's community (ideally excelling in one of those three areas, but not necessarily all three). To pick some examples: tenure at, say, UT-Dallas, Cal Poly, or Occidental College is totally different from tenure at Stanford, UIUC, or MIT.
However I understand from his perspective which is from someone who wants stay on research, go from PhD to post-doc to to fellow to principal investigator to whatever. Competition is intense and everyone along the way is selected for people who are so interested in their work that they might be very unbalanced in other area's.
It's all about trade-off's. I personally would rather trade for working on things that I find interesting and technical than A/B testing. But I can understand someone who is tired of academia and wants to live in the "real world" for a change. I can also understand people who are very happy doing things that they think are very engaging/fun/important in consumer web and respect that.
The whole grant and annual weighing scales approach nowadays strongly dis-incentivizes risk-taking. Without risk-taking with great big brass genitalia, research results are pretty pedestrian.
There's tons of low-hanging fruit that could be worthy of some serious papers / prestige if the work was done for them.
1. Biology : Repetitive Strain Injury calculations via Neural Load
2. Computational Photography : Low Cost, Quick Photogrammetry Calculations via GPS/9DoF metadata
3. Biometric Security : MS Kinect Biometric Security via RGBD and 3D Sound
"As a PhD student you get to choose some interesting problem and work on that, right? That's what I naively imagined before I started at the institute. Some day at lunch I told my supervisor about this idea I had. We could take his fluid simulation method from computer graphics and apply it to a problem related to molten polymers. There was this experiment by a group of applied physicists that would fit nicely. He asked me how many people would be interested in the problem. Maybe a handful, I said. And then I realized that there was no way I could work on that problem. Success in academia is measured in the number of citations your paper receives. What point is there in writing a paper that is only interesting to such a small audience? To be successful you need to target a large audience, and not just pursue whatever obscure problem takes your fancy."
As a society, don't we want our young academics to be able to pursue their intellectual passions, no matter how obscure? A lot of original research is produced in areas that others have overlooked or considered unimportant. Do we really want a "citation market" to determine whether researchers can pursue their passions?
For a start, PhD students don't know how to do research. That's one reason one does a PhD: to learn how to do research. That means listening to your supervisor and working on problems they suggest, at least in the early part of your PhD. I have seen many PhD students, including myself, arrive with half-baked ideas that go nowhere. Learning to formulate a useful research problem is a part of what you are taught.
The second reason is that resources are finite. You don't get to piss around with other people's money anymore than other adults do.
Another way to look at it is that failure is a part of learning, and chasing those half-baked ideas can build experience (by finding out what doesn't work).
I'm still amazed to consider that the college I attended, had I not received scholarships, would have cost me $50,000/year to attend in total. Yet I can't actually think of a single faculty member in the department I majored in who attained tenure while I was there. I do remember somebody joining with a Ph.D. from Harvard, having studied under a renowned mentor, for a non-tenure-track and low-paying "lecturer" post. This greatly talented person, naturally, is now gone, presumably at some other transient and low-paying post elsewhere. I just looked up one of my professors, who was a brilliant topologist and knot theorist doing a postdoc, and was surprised to find she's abandoned academia and is working as an analyst in oil & gas now.
https://web.archive.org/web/20070221033032/http://news.ycomb...
The writer did join academia, just not for very long.
If I were to get a job at Google, work there for a week, and quit, I think it would just fine to write an essay about "Leaving Google".
How long do you think you need to be working at X before you can write that you are leaving X? Or rather, when do you think you actually join academia? Is a 6th year graduate student, pre-defense, still not yet part of academia?
But then why do you imply, in the conclusion, that leaving academia is like stopping educating yourself? If what you were doing was mainly educating yourself (which was not the case), then you would have been exactly "just a student".
I do not think the industry will accept you without you regularly proving your worthiness. Of course, a PhD is something special so I understand anybody would not want that, but not for the reasons given in the article. Same for crazy ideas and being educated. I believe someone could have given roughly the same reasons for "Leaving X" because she's tired of doing the things that X requires. At least we learn about what doing a PhD entitles you to do, but as others have pointed out, this is quite a rude awakening.
Exactly. I left a physics PhD program after my first year because I couldn't stand the environment anymore, and went into software development for an established company. I'm still given "homework" and "exams", still judged on merit of my work, still proving my worthiness of my paycheck everyday. Luckily I had to work throughout undergrad and grad school to afford it, so I already had that knowledge going into my career, but it really sounds like the author of this article did not.
Bah humbug.
As such, grad students don't really "drop out", just like employees don't really "drop out" of a job when they quit, nor unpaid elected officials don't "drop out" of office when they resign, nor volunteers drop out when they stop doing volunteer work.
Personally I think there was a mistake in the logic. Success as an academic may by related to the number of citations one gets, but it's not a factor which affects one's graduate career. Only the handful of people in your graduation committee needs to be convinced about the work, and once you have the PhD there's no continue on that topic, or stay in academia.
I would by quit interested in your next steps and way down the road.
Go work for Autodesk, SolidWorks, Flux.io, PlanGrid, or one of the other companies making CAD software. WE NEED YOU GUYS.
Substitute "writing a paper" with "building a product" and you're well on your way to being an entrepreneur.
One (might) be influential, or it may be relegated to the annals of some obscure journal, never to read or cared about.
Still, there might be a pay off for one (monetary). The other buys your "academia points". The pay-off is a little more suspect on the academic's side, I think.
Which is why the quality of papers (among other reasons) has gone downhill. I knew a particularly brilliant scientist who would do basic work at a reduced rate in a corporate setting as long as the people actually writing the paper included him in the author list. He shows up on over 1000 papers, many of which are of dubious scientific merit, but because of those kinds of numbers, more and more people flock to him for similar work, without realizing it.
Another reason for leaving that I didn't mention in the text was the birth of my first son. I saw how colleagues had little time to spend with their newborns. Now that I'm on my own I can choose myself how much time to spend with my kids.
Is it bringing enough money for you to pay half of your family's bills?
Forgive me for being so direct, but "independent software developer" is a very broad label, and it contains both the successful and the struggling programmers.
Not that being dependent on a family is a bad thing; the only downside is that it is not something that everyone can do.
I'll share numbers too: My revenue this year is around 50000€. (mostly sofware sales + some consulting) My take-home salary (after expenses, social security, taxes) will probably be around 30000€, which is significantly above the median income in Austria. It's also more than what I'd make as a junior software developer.
I made database apps because I saw an opportunity. I had been thinking about doing something more related to my area of study (physics & computer graphics). However, all I came up with were some vague ideas for games, and games seemed too risky for me.
I do miss some parts of my research (especially maths), but for the most part I'm happy with what I do now.
I like both. Industry is more pleasing because you work on products that are directly used and you are judged (at least in a small company) on tangible results. The downside is that work is mostly driven by the most urgent needs or cash flow, so there is far less time to explore certain things deeper.
In academia on the other times, you are judged on getting grants (which is definitely process that is to a large extend out of your hands, and tends to favor people who were already successful in getting grants). On the other hand, you have far more time to explore, try, fail, and try again (as long as you are not a professor).
One should also realize that graduate studies highly differ per country. E.g. I did my PhD in The Netherlands, and there are no tests, or anything alike. I basically got funding for four years to do research, a small amount of teaching and to write my thesis. There was only one serious sanity check after the first year to filter out PhD candidates who were very unlikely to complete a thesis in four years.
I think this is a point that deserves more emphasis. It's easy to forget the degree to which US education is uncommonly expensive, among other things that skew its ROI.
Makes me wonder if I could afford a CS education abroad (being self-taught), without taking loans.
So, income-wise it also gives you an advantage of four years.
But I do know why I'm there. I'm there to continue to learn new things, to interact with others who are like minded, and to work on problems that I may not get to explore in a paid position for a private company who seeks profits above all.
I could probably write an entire book on what is wrong with academia (the short answer: a lot). But at the same time, I have a freedom to explore that I wouldn't have at even the most liberal companies (unless it was as a pure researcher, which quite frankly are rare positions, especially for a non-PhD).
All this being said, it seems obvious to me that the author made the right decision in his case, however, I think he misunderstood the reasons for going into academia. That might sound naive from my point of view (and to some degree, it absolutely is, but I won't get into that now.), but I also think that it was naive in the author's case to think that the only reason to go into academia is to further his career prospects.
Jumping through all the hoops of academia sucks. I was supposed to graduate in two weeks. Let's just say I won't be. It was quite the punch to the gut when I finally came to realize that I was going to miss the deadline to turn in my thesis. I made the comment to my advisor just two days ago: "I hate this system, as if officially turning in what we've talked about for the past two years means that I will have learned any more than I already do, but if I don't do it, I don't have a piece of paper that certifies me as somewhat knowledgeable in a specific field." Which is pretty normal for almost anyone to note about this academic system. But I know that for many jobs, I need that piece of paper. So it goes. But I didn't go back to school for a job, I went for an education, which I have received.
Academia isn't for everyone of course, but if you go into it just for a better career, I feel many people will be sorely disappointed. If you go into it trying to see how much you can learn and realize that you get out of it, what you put in, you'll still have a lot to complain about, but hopefully it won't be a waste of time.