Comments on: Ze Future! http://kevnull.com/2004/05/ze-future.html Fri, 02 Dec 2011 21:19:38 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: mfk http://kevnull.com/2004/05/ze-future.html/comment-page-1#comment-332 mfk Wed, 30 Nov -0001 00:00:00 +0000 http://kevnull.com/?p=295#comment-332 Being in a PhD program, I can offer this advice on getting a PhD: Unless have a really good reason and like doing research, don't do it. You're smart, talented, and resourceful enough, I'm not questioning that at all. But it's a grueling process, and you really ought to have a good reason for putting yourself through that. Especially when, I think you have a really good shot at getting a job at some of those places. And at the rate that OK/Cancel is taking off, those options can't be ruled out either. Regardless, you (and Kat) should move back to the states or more pertinently somewhere (anywhere)closer. :) Being in a PhD program, I can offer this advice on getting a PhD: Unless have a really good reason and like doing research, don’t do it. You’re smart, talented, and resourceful enough, I’m not questioning that at all. But it’s a grueling process, and you really ought to have a good reason for putting yourself through that. Especially when, I think you have a really good shot at getting a job at some of those places. And at the rate that OK/Cancel is taking off, those options can’t be ruled out either. Regardless, you (and Kat) should move back to the states or more pertinently somewhere (anywhere)closer. :)

]]> By: Audiophile http://kevnull.com/2004/05/ze-future.html/comment-page-1#comment-333 Audiophile Wed, 30 Nov -0001 00:00:00 +0000 http://kevnull.com/?p=295#comment-333 Someone recently asked me for advice about whether they should go for a PhD. This is what I told them: So here's the standard "why you shouldn't get a PhD in Computer Science" spiel. You've probably already heard this several times, but no conversation on the matter would be complete without it. Getting a PhD is usually between a 6-7 year committment, sometimes more, and almost never less. If you were a CS undergrad, the opportunity cost of going to school is huge. Likely costs that you will never recoup in overall financial terms. I haven't seen a "real" salary survey in quite some time, but all talented software developers that I know rake in upwards of $75K/year and almost always in excess of six figures. While, PhDs typically do make more, they do not typically hold an appreciable salary advantage in this field. So if you're at all thinking that getting a PhD, makes financial sense, I'd say that you're probably wrong. If you think being in academia will help you leave behind office politics or the daily grind, you're mistaken. Before you get tenure, you basically have to work your ass off on research to make sure you get published and do a lot of ass kissing to make sure that the in-crowd likes you. You also have to fight with a huge number of other people to come up with research funding (read: beg for money). Then even if you have funding, you have to work your butt off to get good students. Getting tenure at a good research university is not easy. When you finally do get tenure, you probably won't be doing a ton of research. Your students probably end up doing 75+% of it, and you'll spend all of your time working on getting funding. If you want to teach, you can always get a lecturer/teaching position without a PhD. So if none of those options fit into your plans, what is left? Government or industry. Government research is... well... government. There are basically 2 types of research, applied and basic. Basic CS research is fairly limited outside of academia. TJ Watson and just maybe some at Microsoft. If you want to do applied research, you can get a lot of interesting positions without a PhD. So why bother? Basically, you really need to know why you're getting a PhD. Doing research is one of those things that sounds a lot more glamorous than it ends up being in reality. At the end of the day, if that's what you really want to do, all the more power to you. However, you have to recognize that a PhD (like all degrees) is really just a piece of paper. Someone recently asked me for advice about whether they should go for a PhD. This is what I told them:

So here’s the standard “why you shouldn’t get a PhD in Computer Science” spiel. You’ve probably already heard this several times, but no conversation on the matter would be complete without it.

Getting a PhD is usually between a 6-7 year committment, sometimes more, and almost never less. If you were a CS undergrad, the opportunity cost of going to school is huge. Likely costs that you will never recoup in overall financial terms. I haven’t seen a “real” salary survey in quite some time, but all talented software developers that I know rake in upwards of $75K/year and almost always in excess of six figures. While, PhDs typically do make more, they do not typically hold an appreciable salary advantage in this field. So if you’re at all thinking that getting a PhD, makes financial sense, I’d say that you’re probably wrong.

If you think being in academia will help you leave behind office politics or the daily grind, you’re mistaken. Before you get tenure, you basically have to work your ass off on research to make sure you get published and do a lot
of ass kissing to make sure that the in-crowd likes you. You also have to fight with a huge number of other people to come up with research funding (read: beg for money). Then even if you have funding, you have to work your butt off to get good students. Getting tenure at a good research university is not easy. When you finally do get tenure, you probably won’t be doing a ton of research. Your students probably end up doing 75+% of it, and you’ll spend all of your time working on getting funding.

If you want to teach, you can always get a lecturer/teaching position without a PhD.

So if none of those options fit into your plans, what is left? Government or industry. Government research is… well… government. There are basically 2 types of research, applied and basic. Basic CS research is fairly limited outside of academia. TJ Watson and just maybe some at Microsoft. If you want to do applied research, you can get a lot of interesting positions without a PhD. So why bother?

Basically, you really need to know why you’re getting a PhD. Doing research is one of those things that sounds a lot more glamorous than it ends up being in reality. At the end of the day, if that’s what you really want to do, all
the more power to you. However, you have to recognize that a PhD (like all degrees) is really just a piece of paper.

]]> By: Audiophile http://kevnull.com/2004/05/ze-future.html/comment-page-1#comment-334 Audiophile Wed, 30 Nov -0001 00:00:00 +0000 http://kevnull.com/?p=295#comment-334 One more thing. This is another thing I told them: 6-7 years is a long time. I know people that took 8+. I don't know your age or your relationship status, but ask yourself what you want to be doing 5 years from now. Ask yourself if busting your ass for something like $20K/year fits into those plans. One more thing. This is another thing I told them:

6-7 years is a long time. I know people that took 8+. I don’t know your age or your relationship status, but ask yourself what you want to be doing 5 years from now. Ask yourself if busting your ass for something like
$20K/year fits into those plans.

]]> By: mfk http://kevnull.com/2004/05/ze-future.html/comment-page-1#comment-335 mfk Wed, 30 Nov -0001 00:00:00 +0000 http://kevnull.com/?p=295#comment-335 I agree with most of what Audiophile posted, but I'd like a note that the average time to graduate is dependent on a lot of factors. Different programs have very different averages, and different specialties have different averages. For example, theory grads tend to either get out fairly early or very late (or not at all), while the hardware guys tend to plug along and get out with in a time that's Gaussian distributed with a mean between the bimodal peaks for the theory grads. Also if you're not switching fields, your MSc *might* cut down on your graduation time. But any way you slice it, you're going to be in school for a long time (if you do the PhD). I agree with most of what Audiophile posted, but I’d like a note that the average time to graduate is dependent on a lot of factors. Different programs have very different averages, and different specialties have different averages. For example, theory grads tend to either get out fairly early or very late (or not at all), while the hardware guys tend to plug along and get out with in a time that’s Gaussian distributed with a mean between the bimodal peaks for the theory grads. Also if you’re not switching fields, your MSc *might* cut down on your graduation time. But any way you slice it, you’re going to be in school for a long time (if you do the PhD).

]]>