Let me explain: you'll be broke, you'll have the worst emotional roller-coaster ride of your life, you will be neglected by your supervisor/advisor, you will be part of the biggest whining demographic in the world, and you will hate it and hate your life. Yes it will not be all bad, but it will get really bad at many points and no one tells you about it in advance.

On the other hand, the emotional highs can be really high. The sense of achievement is amazing. When you discover something new (which is what you're supposed to be doing as a researcher), it's amazing to know that for a split second, you're the only one in the world that knows what you just discovered - it's yours and only yours.

The end of the day, it's a balance: are the emotional highs enough to offset the lows? For me, it was just enough and I finished my PhD on time. I ended up leaving academia because the experience overall was not one I would like to live with the rest of my life.

Finally, I want to note that doing a PhD is not an objective decision most of the time. Some careers essentially demand a PhD (IP law, academia, etc) but most PhDs do not work in academia and a large proportion do not work in their field of study. Don't think it will open wonderful doors that were closed beforehand.

Good luck.

Pierre

The only difference is that with a startup, the potential for a bigger paycheck is higher than with a PhD.

1. I get to work on the problems that interest me most.

2. I make more money than I would starting at an unfunded startup. It is more than enough to live on.

3. I'm working towards the union card for interesting research jobs and/or academia. These jobs are for people who like to work on problems that interest them.

4. PhD's typically make higher salaries than other degrees. The statistics show that the payoff is only marginally better, if at all, to get a PhD over a masters. However, these statistics include people who work in academia who have intentionally given up salary in exchange for intellectual freedom and thus bring down the average. On the other hand, a lot of people who get terminal masters are doing so to get the bump in salary.

5. I love what I do.

However, if you aren't at all interested in academia and/or research there is no reason to get a PhD. If you'd "much rather like to start working" than you have already made the decision.

I found academia incredibly frustrating. If you're used to Internet time, the pace will drive you insane. You submit a paper to a journal, get comments back 2 months later, submit a revised version, than get approved 1 month after that. Then you're published 6 months later, and then 9 months after that (do the math) the first citations of your work start to appear. So the feedback cycle is around 18 months. And instead of having your ideas tested in the marketplace (near impossible to fake), you're judged by a small number of people who are highly protective of their egos and careers. The whole industry is driven by a need to get published and get grant funding, not to make or promote useful things.

Kissinger said that the reason academic fights are so brutal is that the stakes are so low. A lot of truth in that.

On the other hand, if you want to be a top-notch tech entrepreneur, a PhD is CS has a huge amount of value. Just make sure you pursue a field which lets you learn and apply lots of useful algorithms and data mining techniques, e.g. bioinformatics, information retrieval, image analysis. I regularly use a lot of the things I learnt during my PhD - not the theoretical stuff, but the experience I gained in dealing with messy data. These kind of "deep tech" skills are desperately lacking in the Web 2.0 developer community, and will help you truly delight your customers. You needn't look any further than Google to get the point. My two most successful web projects both have a strong mathematical/algorithmic element which has made them very hard for most programmers to reproduce.

Still, these kinds of techniques are becoming better known in the Web developer mainstream. For example, the book "Programming Collective Intelligence" is a good first taste. So on balance, I'd say skip the PhD, but invest seriously in educating yourself about the kinds of things that are still mostly taught at PhD level.

  "No" is "no",
  "No" is always "no",
  When I say "no"
  It means "a thousand times no."

  "No" plus "no"
  Equals "no".
  All nos lead to
  "No no no."

I'd much rather like to start working, doing my own projects, etc.

And so you should.

What's a PhD worth?

Three to five years of existential despair.

will a PhD get me somewhere a MA wouldn't be able to get me?

Well, my Ph.D. is in EE with a concentration in semiconductor manufacturing and quantum electronics. So mine got me into labs filled with one-of-a-kind, multimillion dollar equipment, and gave me an up-close and extremely personal view of how microelectronics are made. Later, it allowed me to contribute to cancer research (cancer cells = microscopy = optics!) and spend a delightful but unrenumerative three years learning about the biosciences from the inside. None of which was fun enough for me that I wanted an actual N-year career in it, but it was certainly fun. Perhaps even fun enough to be worth the cost.

Unfortunately, you're thinking about a Ph.D. in comp sci, which will get you access to a crappy desk with a PC on it, a lot of barely adequate coffee, an inferiority complex, a lack of sleep, a lack of pay, and the opportunity to write a huge tome that only a handful of other people will read. I assure you that all of these things are available to civilians at a great discount.

Should I only pursue a PhD if my goal is to go in academia/research?

Yes, that's what it's for.

Would a MA be more advisable if I just want to pursue my entrepreneurial career?

Yes. "I'm getting an M.A." is a great excuse to move to a startup hub and hang around with a bunch of smart programmers who are itching for action. So I won't talk trash about the M.A. -- go for it, if you like. The sunk costs are a lot lower, and the benefits higher.

For more on this topic see https://hackernews.hn/item?id=117294 . For more gratuitous TMBG lyrics see http://www.davidkjones.net/tmbg/no/04.php

Select your graduate school and advisor very carefully. If you have an advisor of the entrepreneurial type, then go work for this person and get your PhD. E,g: Working for Robert Morris.

Some very interesting companies have come out of grad schools: Google, Yahoo, Akamai, Teoma, etc. It helps to go into grad school with an open mind, get your Masters and at that point in time, decide if you want to go ahead with a PhD.

As a general rule of thumb, this line from the movie Ronin, helps: "Whenever there is any doubt, there is no doubt."

http://www.imdb.com/title/tt0122690/quotes (4th quote from top)

An MS is worth about $5-10k extra per year, and it will allow you to act as a lead or get professional stature which is a requirement for some government regulated contracts.

A PhD is worth $10k-$50k extra and will likely allow you to lead a research team, rather than just a development team.

However if you are on YC then you probably have an entrepreneurial spirit, and the 3-6 years you spend on a PhD could have been spent learning how to run a business.

My advice is to start some small simple business, make it profitable, sell it and start another one. In the mean time practice your tech, which should always come more easily to you. There was so much for me to learn on my first business that could have been learned in a cheaper industry than Software.

A PhD, MS, or any other validation of your tech skill will be much less valuable than your ability to sell product at a profit. A proven track record with some mediocre tech, is worth more than some great tech and a shotty business record.

But I decided that it was more important to be doing work that I enjoyed rather than slaving away on a PhD that was not interesting to me and irrelevant to my future career anyway.

I've now got a good job in an exciting startup (http://www.songkick.com/) doing the work I wanted to be doing. What more could I want?

Whether this is relevant to you I don't know. But I have strongly come to feel that unless you are going to be an academic, or work in a field for which it is specifically useful, like finance or pharmaceuticals, a Phd is not useful.

But I have a vested interest in that being true so perhaps you shouldn't listen to me :-)

There's only one right reason to do a PhD, and that's because you really really really want to do a PhD in some particular subject. I did one in physics because I really really really wanted to, and I have never once regretted that decision.

So for you I think the answer is no. If at some later stage in your life you decide you really _do_ want one, you can always go back.

If you really _did_ want to do one, the question I would be asking is this: why would you want to do it in the US instead of the UK? You can usually save at least two years by doing it outside the US, and the UK has some very good universities (well, at least two of 'em).

That's one good reason to get the degree in the States.

This document is intended for people applying to Ph.D. programs in computer science or related areas. The document is informal in nature, and is meant to express only the opinions of the author. The author is currently an assistant professor of computer science at CMU, and has been involved in the Ph.D. admissions process at CMU, U.C. Berkeley, and MIT."

www.cs.cmu.edu/~harchol/gradschooltalk.pdf

I tend to agree with the tone of the responses so far: don't pursue a PhD unless you're convinced that you want to live the life of a professor (and you're comfortable with the fact that you'll likely never make it into a professorship at a good school, if at all). Definitely don't pursue a PhD if you've been told that you need it for a job (i.e. in pharmaceutical research), but you've never experienced the industry in question; if you think you need a PhD for a job, get a job close to the one that you want, work for a few years, and then decide on the degree. The job market for professional researchers is quite tight, and you may well decide that 5-8 years of post-secondary education and poverty isn't worth it for the jobs that a PhD makes possible.

That said, I can hardly think of a better place to meet smart, interesting, creative people than in a PhD program at a top university. If you go into it with your eyes open (and you don't get sucked into the cult mentality that makes it hard to quit), you'll learn a lot, meet brilliant people, and you can still leave with a master's degree in a reasonable time. (And hey...there are worse places than a college campus for a 20-something single man to be spending most of his time....)

But.

I don't think those three years were a good investment when seen from a start-up perspective. I think three years of extra real-world experience would have been more valuable.

I've never used any of this stuff.

There are plenty of things that I learnt as an undergraduate that I use, however.

I think it's a bit like comparing high school math with university math. Or like saying that you can program after taking a single programing course. :)

Similar skills can be gained elsewhere, are other skill sets are often needed in industry.

I have to agree with the rest - do a PhD if you clearly want to do research/work in academia. If you're just doing it for the credentials - stay far,far away. You might survive, but you will be utterly miserable.

Go ahead and do your PhD. You'll learn some interesting stuff. It will be narrow, deep and might have some application. You will learn intellectual rigour, how to write and think. Then when you finish you will be just the type of new-hire a startup might want.

I often think doing a PhD is the anti-entrepreneur approach. The sheltered workshop of academia. All analysis, no action and devoid of the real cut and thrust of commerce. Then I think of all the new ideas generated by those who spent their time thinking about stuff. But did Woz need to do a PhD to build computers and start Apple? Do you need a PhD to do a startup? No. But the skills you learn there might help. Alternatively by the time you become a newly minted 'Dr', you might have created something of value and still learnt.

So what is it to be? The risky path of the Startup with its uncertainty, risk and commercial experience? Or the almost certain path of a PhD, a title and specific knowledge and less potential reward?

The other big problem in academia is working with the people you want to work with. You can't generally band together with your friends and do a PhD together. There will probably be some great people in the lab but you usually aren't working too closely.

It can be a good way to get paid (a pittance) for doing interesting work. I don't think it helps in the startup world - much. Certainly not when you consider that working for a pittance for 4 years could allow you to startup and fail at a number of ventures of your own.

Advantages - you will have to be motivated enough to give up a pay cheque. However, being more mature than the other kids, you will get financial aid first.

You gain a lot of credibility in industry with a Doctorate...people defer to you on all things scientific. And you join an elite club, though not quite as elite as it used to be before they required all professors to have a Ph.D., forcing colleges to hand them out to people like me to keep the supply of professors steady.

Oh, and put your all your advisers on every paper you publish. It is hard to deny someone a degree after they have published 5 papers that you agreed with so much you put your name to it.

And get rid of that foolinsh "create an account" waste of time - do you really believe any data entered these is not junk? Dr. J. K.

As long as you can get into a PhD program, it gives you a lot more flexibility and is cheaper than a Masters program.

Don't let the bullshit of academia get in your way. You don't need to go to too many useless conferences or publish more than you think makes sense.

Most programs will let you leave early with an MA, so may as well start with a PhD program.

Another problem is the opportunity cost, you loose four to six years of your life that you could have spent doing something meaningful, or at least accumulating resources. That might not seem like a big deal right now, but it will become a serious issue after two to five years.

  The worst part about graduate school is that you have to
  work on someone else's ideas instead of your own

I'd like to believe that the best advice I've ever given about doing an advanced degree is not to start one unless you know why you're doing it. Some of the best folks I've worked with came in with a very specific agenda, whether it was a problem they wanted to solve, an advisor they wanted to work with, or a company they wanted to work with. Having that sort of end point in mind does wonders for focussing your time.

In the interest of full disclosure, I should point out that I did NOT follow this advice. I stumbled into grad school without much forethought. If I were to do it again, I'd make sure I had a reason.

Personally, I got a lot out of my thesis. It made me a more curious person, refined my research, questioning and reflection abilities, and improved my critical thinking skills. But it did not help my career...

A Computer Science PhD (especially in more concrete areas like systems / security / perhaps even PL a la Graham), I believe, is different from many other academic PhD's in that what you learn through the course of your research and dissertation can have real world implications in industry. Whether or not it does, though, depends on the idea you are working on which is entirely your responsibility to define. There have been some comments below about being forced to work on "somebody else's problem" with which I disagree. It is your responsibility to find the best advisor for you and the best advisor for you is going to be the advisor who either A) shares a serious interest in the exact problem you would like to solve -or- B) has a general interest in the problem you would like to solve and has an advising style that is concerned as much or more with your growth and success than his/her gain from your work. For another perspective on choosing the right research I suggest Hamming's remarks: http://www.paulgraham.com/hamming.html

So what's next for an entrepreneurial minded PhD drop-out? Another PhD program! This time with an advisor whose research interests are very in line with my own and in an area whose commercial market is only growing (collaborative systems).

One must not either be an entrepreneur or an academic. One can be either, both, or neither. The entrepreneur considering the pursuit of academia must convince herself/himself that it is a risky investment that, at best, will lead to great commercial opportunity, at worst, be a painful 4 or 5 years and have lead to research that dead ends, or, most likely, will be a challenging experience where you will learn a vast ammount of new information and instill a certain level of discipline when approaching new ideas. Economically one is safest in going to Microsoft/Oracle/SAS immediately after obtaining a BS/MSc. However, if you're the type of entrepreneur who dreams of innovating on technology (a la Google, Akamai, SAS, Sybase, INGRES, etc) versus the type of entrepreneur who dreams of innovating on applications of technology (Facebook, MySpace, Web 2.0 in general), then a Computer Science PhD may be a very worthwhile investment in the path to your ambitions.

"Any day you can be paid by someone to learn something new... that's a good day." -Prof. Gary Bishop

A PhD is really only necessary and sufficient for an academic career, which is not the same as a research career.