In Defense Of The PhD | Hacker News Reader

In Defense Of The PhD

soundandcomplete.com

60 pointsarespredator43 comments

I'll defend the PhD for different reasons.

It educates you further. You acquire knowledge you don't get elsewhere. You get a range of skills and learning opportunities you don't get elsewhere - some of which are very useful when trying to discover new human knowledge. Research labs in universities do - and have traditionally - discovered much new human knowledge; they are good at it.

Using earnings as the sole measure of the worth of a PhD - while maybe what you'd expect The Economist to do - is a little simplistic.

If getting an undergrad CS degree didn't lead to higher lifetime earnings than, say, doing data entry, would you do data entry?

I wouldn't. I put a premium on acquiring knowledge, and doing work I consider interesting.

Research is very interesting work. Having a research degree can lead to more interesting jobs. Like maybe working on the algorithm design team, as opposed to the web development team. A lot of corporations do have labs doing research activities, where people with PhDs work. This should be factored in to any analysis.

Of course, if you view a PhD purely as training for academic positions, then clearly something is broken, because there are very few academic positions.

But if you see the PhD as an educational experience which broadens knowledge, and teaches the skills necessary to break new ground in hard technical problems, that's a different story.

My dad has worked at a grocery store for almost 33 years, and he's 50 years old. He always made good money (especially for not having a degree), and did a good job, but I never felt like he actually liked his job, especially with all the crap he has to go through from both employees and bosses. He's been salary for a long time, and has worked probably an average of 60+ hours a week for the past 12 years.

This always had a profound effect on me. A long time ago I decided that I would always strive to have a job that I truly liked, something that payed enough but didn't suck the life out of me. I'm in a job that I love now with only a Bachelors, but I kind of got stupid lucky with it. A PhD is really the best ways to secure a job that you'll likely love. You may get paid lower than you'd expect for someone with your skills, but money isn't everything.

I don't know anyone who has gone to school for a PhD just to get a better paying job when they get out.

"I don't know anyone who has gone to school for a PhD just to get a better paying job when they get out."

Now you know of one.

What are you going to school for?

PhDs open up a lot of jobs that you wouldn't be able to get in with only a Masters, but they don't typically pay that much more (most my background is with Physicists)

I do know lots of people who go for a Masters in order to make more money.

"What are you going to school for?"

I got a Ph.D. in engineering; the program was in essentially the mathematics of operations research. So, the emphasis was on optimization and stochastic processes. The work was fully careful mathematically, especially in measure theory. My dissertation research was on stochastic optimal control (yup, with careful attention even to 'measurable selection').

"PhDs open up a lot of jobs that you wouldn't be able to get in with only a Masters, but they don't typically pay that much more (most my background is with Physicists)"

Eventually I concluded that for a 'job', a Ph.D. was from nearly useless down to less good than a serious felony conviction. The main reason is that for a "job", the people hiring are still almost entirely stuck in the 'job hierarchy' model of 120 years ago on the factory floor where the supervisor was supposed to be older and know more than the subordinate, and the subordinate was supposed to be there mostly to apply labor to the ideas of the supervisor. So, since there are nearly no supervisors with Ph.D. degrees, the situation is as I described.

I was misled: My early career was around DC on mostly military problems. There a technical Ph.D. would have helped a lot. During that work, at times I wanted to know more about optimization and stochastic processes, so I selected such a Ph.D. program.

But then I drifted away from such DC-DoD work. Dumb move.

I never had even for a milli, micro, nano second any desire for an academic job.

But in the US now, technical "jobs" are essentially a joke. So, major US employers have often been nailed due to lack of ability to change. So, 'globalization' nailed GE and Detroit, both of which at one time tried to do highly technical work. IBM and AT&T were nailed by changes in microelectronics. Of these four, in my career at times I interviewed at all of them and worked for two of them -- NOT a good situation.

Broadly, the chances that a person can, in their 20s, get a technical 'job' in a company and then hold that job until retirement are next to zilch.

So, somewhere between the 20s and retirement, the technical person will be out of work. Then they can send a few thousand resume copies, get help from all their friends, etc. and still be out of work.

They will discover the crucial importance of geographical barrier to entry and conclude that an electrician's license is a MUCH better foundation for a good career than a Ph.D. in electrical engineering. The Ph.D. of 35 who gets fired from a technical job in some big company can conclude that, literally, in his neighborhood, while he still had his house, the guy cutting grass who started 15 years earlier and now with, maybe 10 employees actually has a much better career. E.g., the grass guy has customers from the broad economy of people who have grass; the grass will keep growing; grass mowing just cannot be 'outsourced' and has a terrific geographic barrier to entry; do well in a radius of 50 miles and can do well. Just mowing grass. Shrubbery, etc.? Maybe still better. Landscape architecture? Maybe still better. Also own a nursery and be 'vertically integrated'? Maybe still better.

I mowed grass, etc. as a teenager and had more happy customers than I could serve. I actually had an opportunity, basically better than a Ph.D., but didn't know enough about the economy to see it. I was also good at repairing cars, e.g., clutches, radiators. That was another good opportunity I didn't appreciate.

Net, in the US, a technical 'job' is a fool's errand, and anyone needs something MUCH better.

Broadly there are good examples of what to do on Main Street in each village, town, and city all across the US.

So, net, one of the main, good career directions is just to be a sole proprietor. So, can be an electrician, plumber, landscape architect or run a driveway paving company, a roofing company, a dry cleaning service, a few pizza carryout places, etc. Just look on Main Street.

Basically trying to have a career from a 'job' just sucks and is usually less good than just being a Main Street sole proprietor. Being a successful Main Street sole proprietor is not always easy, it's just that the odds are that it's a much better direction than being a technical 'employee'.

Note: The employee just can essentially never have a better job than the one provided by their supervisor. Unless the supervisor is doing fantastically well, the employee is on a long walk on a short pier. E.g., look at a TV biography (go to Clicker and follow biographies) of Carnegie: Early on he was an assistant to a big shot in the Pennsylvania Railroad when rails were likely having better growth than the Internet now. So, he had a good start. He accumulated some cash, made an investment in a competitor of the Pullman railroad sleeping car company, did well enough to buy a long yacht, and then just kept making investments, especially in steel. He finally sold the steel to Morgan who took it public and cleaned up big time. The key was for a while he had a really GOOD 'job', but these are rare.

Assuming want to be a sole proprietor, besides a geographical barrier to entry, so common on Main Street, with a Ph.D. and some targeted research can also get an advantage and a technological (intellectual property, 'secret sauce') barrier to entry. So, do that. So, such applied math research "Isn't just for peer-reviewed journals anymore.".

So, start a Web 2.0 company, derive some applied math, use it as the crucial, core, intellectual property 'secret sauce' to give results a few hundred million people will like a lot and can't get elsewhere, use info on the users to do ad targeting, and get revenue.

Notice the Canadian romantic matchmaking Web 2.0 company Plenty of Fish: One guy, two old Dell servers, ads just via Google, and $10 million a year in revenue. Nice sole proprietorship, maybe worth $1 billion. Nice pay for writing a little routine software.

An advantage is, while some pizza guy is tossing these flexible disks in the air one at a time, a Web 2.0 guy can have some $500 servers sitting there pumping out dozens of Web pages and ads per second.

Uh,

3600 * 24 * 365 = 31,536,000

A few dozen ads per second from each of a dozen or so servers, with a little targeting, and multiply it out.

If the Web site is really popular, then get a nice house, car, yacht, and IRA brokerage account in a big hurry. Else, use the same Web site and software 'skills' to do another Web 2.0 site.

Here the Ph.D. can be crucial, a huge advantage.

But, net, in the US now, a technical 'job' is a really dumb career direction. Instead, it is just CRUCIAL to OWN the enterprise that are going to count on to keep paying the bills for several decades.

Employers: You're screwed; no one with any good sense will work for you without one heck of a salary, golden parachute, or founder's stock.

Technical guy: Learn programming, data base, system and network management and administration, Web site development, etc. and go for it.

From all I can see, for likely all the rest of this century, the opportunity is from (1) processors with many cores for less than $200 per processor, (2) optical fibers for 1 Gbps or so to/from each IP address, (3) from (1) and (2), wide, deep, rapidly flowing oceans of data, (4) software to process that data and deliver valuable results, and (5) some original math to make the processing more powerful. So, it's 'automation', third millennium style.

Automate what? As much as possible of all the work there is to be done. Why? Because we already know what people want in the famous one word answer, "More". So, use (1)-(5) to deliver more for less cost -- that's called 'productivity' and the key to economic progress.

  >> Of course, if you view a PhD purely as training 
  >> for academic positions, then clearly something is broken,
  >> because there are very few academic positions.

It always surprises me when people think this. It stands to reason that in a non-expanding market that each professor can on average train only one replacement in his/her entire career.

As it happens, I'm defending my dissertation tomorrow (neuroscience, U.S.) and this post pretty well sums up my sentiment.

I particularly agree with the ending:

    Oh you’re saying I’m a dreamer, and that simply never 
    happens? Well what about those thousands of internet 
    start-up companies? They waste their time as well, trying 
    to become another Facebook or another Google. Yet they 
    still do it, because it’s their dream to pursue.

    And so is academic career ours.

I think there are two modes one can do research in: one is a conservative, do-what-you're-told kind of incrementalism; another is higher risk, swing-for-the-fences, and entrepreneurial. I think there is a similarity between doing research in the latter mode and a technology startup, which is why I got sucked into Hacker News when it was still called Startup News. Nearly everybody on it was a maker, swinging for the fences. I felt a lot of affinity for the attitude if not the methods.

Anyway, enough HN procrastination, I need to practice my talk!

Good luck!

Break a leg! I defended mine last month, and let me tell you- the sense of relief afterwards is amazing.

Thanks very much for the well wishes you guys. I am very much looking forward to that sense of relief. In the coming week or two I will probably post a 'Show HN' of my project (wiring diagrams of neuronal connectivity).

Best of luck!

I attended my first dissertation recently and it has energized me to finish my master and to suss out a nice PhD for myself. To be able to discuss my ideas for a couple of hours: yes please!

Have a ball tomorrow!

Best wishes!

I'm happy I'm currently in the middle of my PhD, but this article makes a bunch of errors. It generally strikes me as the naive impression that undergrad have of grad school until they spend some time talking to post-doc and professors. I'll just identify a couple of obvious ones.

1. "We all read the PhD Comics and we all hear about how many hours of coursework or admin-duties a typical US grad student has. I don’t know how does it look like in other countries, but in Norway, Denmark, Holland and Belgium this is definitely not the case. "

The Economist article doesn't say anything about there being too much non-research work. It says there are very long hours, and part of it is non-research. The OP has simply misinterpreted the Economist article. I don't know any grad students in the US who spend more than 25% time on teaching, and the OP lists this maximum as some sort of great aspect of the Nordic system.

2. "Apart from all that, a little bit of teaching looks good in your CV, especially if you want to apply for post-doc or other academic positions after finishing a PhD."

For most grad students, than this is just plain wrong. No one cares how good of a teacher you are (especially not for post-doc positions) unless your end goal is a teaching position at a small liberal arts college. But most most grad student's end goals are either research positions or industry.

I don't know where you're getting 2 from. To throw one anecdote out there, my advisor, who has experience with hiring committees, has told me that for academic positions, they're looking for a wide range of things - including demonstrated teaching ability. And for professors (even at a research facility), poor student reviews of classes you've taught can get in the way of being awarded tenure.

> I don't know where you're getting 2 from.

I'm in my 4th year of grad school, and I tend to pick the brains of professors and post-docs a lot. I have never once heard them (a) recommend that I get more TA experience than is strictly required ("get on a fellowship or a grant as fast as you can so you can completely absorb yourself in research"), or (b) mention that a professor was granted/denied tenure for anything besides research.

Don't get me wrong, departments pay a lot of lip service to this stuff. And yes, at many places the professors have to write a (brief) essay explaining how incredible they are at inspiring young students. But as a determining factor in advancement at research universities, my strong impression is that it's practically insignificant.

You'll probably think that Sean Carroll speaks with more authority than me, though. Here is how he puts it:

"When was the last time Harvard made a senior tenure offer to someone because they were a world-class educator, rather than a world-class researcher? Not only is the answer 'never,' the question itself is somewhat laughable."

http://blogs.discovermagazine.com/cosmicvariance/2008/05/29/...

Oh please, guy says linked articles are crap because they are only about US and UK and there is still rest of the world where the situation looks different. But he himself says only about Norway, Denmark, The Netherlands which are at the very top of the richest countries in the world list (he mentions also Slovakia which apparently has good PhD programs funding) and excludes poorer European countries such as Poland and Ukraine because they are too poor. I am sorry, but what kind of argumentation is this? Article is bad because it is only about US and UK but his blog entry is OK even though he compares only three richest countries in the world?

I am also doing a PhD in Norway, but in the humanities (ABD now). I agree with a lot of what is said here: PhD students in Europe, and especially in places like the Netherlands and Scandinavia, have it unbelievably good (not that you'd believe it if you followed the debate here, with "PhD oversupply" and "temporary contract abuse" a prominent topic; to be fair, most graduates from my program have found academic jobs both in Norway and elsewhere, though there is a tendency for people who have spent over ten years in temporary positions to call it quits). However, places like the Netherlands and Scandinavia, even if they punch above their weight in many fields, are not too big in the grand scheme of things. Again, I don't know that the situation is like in the hard sciences or in CS, but certainly in my field the fact that Europe has well-funded PhD programs and jobs does not help the pretty dire (as far as I can see) situation in the US. There are relatively few people who go on to positions in Europe after US grad school, and even fewer the other way around.

The comparison between startups and PhDs is not esp. appropriate. The entrepreneur often is willing to assume greater risk for greater pecuniary reward, whereas the PhD often is not assuming any greater risk but willing to accept likely lesser pay to do what he loves.

Which isn't to say that there is a class of entrepreneurs that start companies simply because they love starting companies, but that this isn't unique to entrepreneurs (e.g. there are lawyers who love doing lawyerly things).

I agree that it's not a perfect comparison, but there are undoubtably similarities. In both cases, you're working mostly on your own, on a project that you have to define and that is entirely dependent on your personal level of self-motivation. One of the biggest challenges to running a startup seems to be correctly defining the problem on which to work, and the single biggest hurdle to finishing a PhD is coming up with your dissertation topic.

I have long thought like this. I think the key similarity isn't in the particulars, it's that both are self driven, creative works, and in most cases, no one else will care what you're doing until you succeed.

Money isn't everything. Article is comparing absolute rewards, not pecuniary rewards. That said, I would say that a PHD is a much more modest risk than a startup, and has similarly modest rewards.

I believe that most of the articles that been posted to HN that are negative in regards to doing a PhD are referring more to the liberal arts types of degrees, and the author here points out some reasons that these people are going on to do PhDs in the first place.

Sure, getting a tenure track professorship in English or the Humanities would be great for a lot of these people, but isn't doing a PhD better than the alternates that most of them have available? I would rather pursue a PhD than go do phone marketing, myself.

While I think this article makes some decent points, I have to say that in all the recent discussion around the value of a phd and whether or not the system is broken the only really strong defenders of the way things are have been those who are currently pursuing a PhD.

As someone who has been able to swing a faculty position without one (wanting to remain anonymous I'm not going to say what exactly is the case, but I will say it's definitely not because I'm particularly clever). It saddens me to see friends who were brilliant PhD students at very good schools, doing post doc work at mediocre schools that don't even do research in their area of interest (and these are some people in the math and sciences). Almost none of the people I know with PhD work in the humanities are doing anything even close to what they studied.

I don't think the people who are pursuing PhDs are necessarily the ones who are broken, but there is definitely a problem with the system, and I don't think most phd student will realize it until they've graduated.

I completely agree with this. I've just gotten my PhD, and could have gotten some commercial job with a higher paying salary. Yet I've chosen to stay in academia, doing something I love, travelling all over the world and having fun doing it.

I should also note that, at least in the Netherlands, young researchers will be on a temporary contract for quite some time. Circumstances (laws, funds, budget cuts, etc.) are against us and unfortunately create a very bad situation for us.

Same here. I recently started a PhD after working in consulting for a few years. It's less money, but I'm much happier.

    Normal people don’t study philosophy, and if they’re into computer science,
    they don’t care whether P!=NP — they just learn Java, Objective-C, Python or
    whatever else they find useful for becoming a successful software engineer.

I resent that. Most programmers I've met find theoretical CS interesting, and learn things simply for the intellectual value of it. They just want to build things people use.

From my anectotal experience, they might find it interesting, but would only rarely go in-depth and pursue it with the needed diligence to further the state of knowledge (cf. P!=NP).

That's exactly the difference the author means. Programmers like to know about this as a means to better achieve things or purely as a hobby, exercise for the mind. People who pursue a PhD go beyond that, and are simply interested in the absolute truths (in the Platonic sense).

But he said "they don't care", as if pursuing a career writing software usually means you aren't interested in those absolute truths. And I'm saying maybe you are, but you want your job to be about something else.

Purely scientific questions like P!=NP are interesting to more than just academics. A Brief History of Time has sold over a million copies, and I bet mostly not to physicists.

My experience is almost exactly the opposite. Most programmers care little to nothing about any sort of CS or theory. I've met successful programmers with years of experience who'd never even heard of even the simplest CS concepts like linked lists. How their language of choice actually works below the surface is a complete black box to them, and they're quite happy to keep it that way.

If you meet a programmer who cares in any way about theory, then they are with all likelihood comfortably in the top 50th percentile and probably also in the top 25th.

I think what you're saying can easily be extrapolated to any vocation...

There are people who are passionate about what they do, and then there vastly more people who are doing it simply because it's a job.

Exactly. My point was that programming isn't too different than most other fields in this respect. This is easy to forget if the only other programmers you actually meet are top silicon valley hotshots.

To add to the discussion, the ability that the PhD gave me was the confidence to "roll my sleeves up" and delve into the muck to try and figure out how to get something working, or figure out why something is not working.

> Normal people don’t study philosophy

Well, I don't want to sound like a douche or anything, but they really do know how to pick their words, don't they? Following the same logic only Plato would qualify as a true philosopher (he's the evil mind behind all this "academia" concept anyway), while Democritus, one of the few philosophers who actually knew what he was saying, would just be seen as an amateur.

After weeks of PhD bashing here is one article which successfully defends it against a lot of arguments against a PhD.

I especially like the comparison of an academic career with the startup track.

While it is true, that BENELUX and Norway have more budget for research as it stands, opposed to say the cutbacks in the UK, but I would still be interested in the blog entry after he has finished (the last small piece is more than 90% effort). It is just like mountain climbing: you can only state that you made it if you make it home. Everest is littered with - I guess happy - bodies.

btw .no, .be, .nl are among the richest in Europe (petrol, gas, trade, EU admin. overhead etc).

Universities in the Netherlands are getting hit by budget cuts pretty severely as well. This means that permanent contracts are getting harder and harder to come by in academia.

Three main reasons for this in the Netherlands: * A permanent employee of a university is considered a civil servant, which means that it is impossible to fire him. * A temporary employee is "protected" by law; after three consecutive temporary contracts, the next contract is by law a permanent one. * Most funds at the university come from limited time projects.

This means that universities are very hesitant to hire people on a permanent contract and labour protection laws sort of turn against you.

Wow. In Norway, there is a similar problem with the "four-year" rule, which means it becomes very hard to fire someone who has worked for an employer for over four years. Now the PhD of course also takes four years, and PhD students are employees. In practice, almost no-one claims the four-year rule following, say, the PhD and a short-term research-assistant engagement (there was a recent case where two people in Oslo did something similar, and the university settled out of court, hiring them permanently, but it was quite unprecedented), but this means that universities much prefer the fixed-term contracts to be shorter than that, again reinforcing the glut of temporary positions; but if we had the three-contract rule, quite a few people I know would have been permanently hired already.

The PhD contract is just a single contract, and a position only becomes permanent after a third contract in a number of years I think.

Once you get your PhD, and you're rehired at the university, you get a different position, starting the counter from 0 again.

But no, if Norway had the three contract rule as I described, the people you know would have been fired already.

> the people you know would have been fired already.

I guess you're right, though who knows what those pesky unions might have said.

As pointed out elsewhere the comparison don't hold up.

Startups are high risk high return.

That can't be said about PHDs

You are comparing top startups to mediocre Ph.D.'s. Top academia is very high-risk and very competitive. You take on very difficult research problems in the hope of becoming famous (but you may never be able to solve them). The rewards are accordingly high - having a large impact on the world, getting tenure to do what you love for the rest of your life, being surrounded by the best people in your field.

Reply on news.ycombinator.com