Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Why (Not) Do a PhD in Computer Science?

blog.skanev.org

Post reply on HN
Print view View original

11–20 of 45 posts

Re: Why (Not) Do a PhD in Computer Science?

#11

Post by drcross »

sc0rb wrote: post #8
wjnc wrote: post #3

Never do a PhD for the money. Ever. It's not worth it, compared to using your skills to get a good education BSc / MSc and then start your career. You miss so many years of good income, plus income growth and will always stand out as the theoretical type. That being said: it's all about curiosity and people. So if you're curious and have the time / money... go for a PhD!

I know of three PhD's in comp sci that went straight into Goldman Sachs (and one hedge fund) where they immediately started earning far more than those of us that left Uni with our Bachelors. They work in high frequency trading mainly using Java.

and in doing so sold their souls to the devil.

Re: Why (Not) Do a PhD in Computer Science?

#12

Post by arethuza »

sc0rb wrote: post #8
wjnc wrote: post #3

Never do a PhD for the money. Ever. It's not worth it, compared to using your skills to get a good education BSc / MSc and then start your career. You miss so many years of good income, plus income growth and will always stand out as the theoretical type. That being said: it's all about curiosity and people. So if you're curious and have the time / money... go for a PhD!

I know of three PhD's in comp sci that went straight into Goldman Sachs (and one hedge fund) where they immediately started earning far more than those of us that left Uni with our Bachelors. They work in high frequency trading mainly using Java.

Were they actually using anything specific from their PhD? The research they did was directly relevant to some part of the HFT application?

Re: Why (Not) Do a PhD in Computer Science?

#13

Post by liljimmytables »

simonh wrote: post #10
sc0rb wrote: post #8

Earlier quoted context omitted.

I know of three PhD's in comp sci that went straight into Goldman Sachs (and one hedge fund) where they immediately started earning far more than those of us that left Uni with our Bachelors. They work in high frequency trading mainly using Java.

Interesting. I wouldn't have though Java was suitable for HFT due to garbage collection pauses playing havoc with your system determinism and performance reliability. I supported a derivatives trading system written in java that was canned due to performance issues, and where I am now we had java FIX protocol gateways that we've replaced with C++ components, again for performance reasons. I'm not saying java has perf…

> garbage collection pauses

AFAIK, mark-and-sweep GC was removed from the JVM years ago (except on Android, where Dalvik used it until 2.something). I can't help with the rest of your post though.

Re: Why (Not) Do a PhD in Computer Science?

#14

Post by thomasjames »

vladoh wrote: post #5

I like the reasons described in the post. I would like to add one more that was really important for me when deciding to do a PhD in CS - learning to write papers. While it may seem like a burden for some people, I think that describing your research in a very limited space (something like 8 pages) teaches you very valuable skills. You should also write it in a way that most of the readers will understand your ideas…

This is important to me. It is the reason that most of the jobs I am interested in require a PhD. Outside of the start-up and web space, there are actually a lot of industrial research jobs that will not even give your resume the time of day without a PhD because you are expected to produce technical literature and patents. I thought it was kind of funny in the article how they still mentioned the "I want to be a professor" reason. No interest in that rat race.

Re: Why (Not) Do a PhD in Computer Science?

#15

Post by arethuza »

thomasjames wrote: post #14
vladoh wrote: post #5

I like the reasons described in the post. I would like to add one more that was really important for me when deciding to do a PhD in CS - learning to write papers. While it may seem like a burden for some people, I think that describing your research in a very limited space (something like 8 pages) teaches you very valuable skills. You should also write it in a way that most of the readers will understand your ideas…

This is important to me. It is the reason that most of the jobs I am interested in require a PhD. Outside of the start-up and web space, there are actually a lot of industrial research jobs that will not even give your resume the time of day without a PhD because you are expected to produce technical literature and patents. I thought it was kind of funny in the article how they still mentioned the "I want to be a pro…

Don't those kinds of industrial research jobs also usually require the kind of publication track record that you have to acquire in academia to survice?

Do industrial research labs really employ freshly minted PhDs - who probably haven't acquired the most important skill in modern research: the politics of attracting funding/backing?

Re: Why (Not) Do a PhD in Computer Science?

#16

Post by thomasjames »

arethuza wrote: post #15
thomasjames wrote: post #14

Earlier quoted context omitted.

This is important to me. It is the reason that most of the jobs I am interested in require a PhD. Outside of the start-up and web space, there are actually a lot of industrial research jobs that will not even give your resume the time of day without a PhD because you are expected to produce technical literature and patents. I thought it was kind of funny in the article how they still mentioned the "I want to be a pro…

Don't those kinds of industrial research jobs also usually require the kind of publication track record that you have to acquire in academia to survice? Do industrial research labs really employ freshly minted PhDs - who probably haven't acquired the most important skill in modern research: the politics of attracting funding/backing?

It is actually very common to go to a research facility after a PhD. Anyone with a doctorate from a reputable program will have a list of publications that demonstrate a clearly defined interest and quite possibly even grants. For those that aren't at the level of a research scientist, there are often post-doctoral fellowships. If the companies are large enough, grants are less of an issue because there is a built-in R&D budget. There is less of the panhandling you see in academia.

Re: Why (Not) Do a PhD in Computer Science?

#17

Post by seanmcdirmid »

arethuza wrote: post #15
thomasjames wrote: post #14

Earlier quoted context omitted.

This is important to me. It is the reason that most of the jobs I am interested in require a PhD. Outside of the start-up and web space, there are actually a lot of industrial research jobs that will not even give your resume the time of day without a PhD because you are expected to produce technical literature and patents. I thought it was kind of funny in the article how they still mentioned the "I want to be a pro…

Don't those kinds of industrial research jobs also usually require the kind of publication track record that you have to acquire in academia to survice? Do industrial research labs really employ freshly minted PhDs - who probably haven't acquired the most important skill in modern research: the politics of attracting funding/backing?

> Don't those kinds of industrial research jobs also usually require the kind of publication track record that you have to acquire in academia to survice?

Yes.

> Do industrial research labs really employ freshly minted PhDs - who probably haven't acquired the most important skill in modern research: the politics of attracting funding/backing?

Yes. How do you expect freshly minted PhDs to learn the art of attracting funding and backing?

Re: Why (Not) Do a PhD in Computer Science?

#18

Post by barry-cotter »

lutusp wrote: post #2

You need to be aware that PhDs often make less or the same income as a "professional" degree holder does (for a higher student loan burden). That means if you feel that the experience will be personally enriching, the fact that it might not be financially enriching may not matter.

1. Professional degree does not need scare quotes.

2. US professional degrees have dropout rates of 3%, approximately. The Ph.D. with lowest dropout rate, engineering, has a 35% dropout rate. Pretty much everyone who can be admitted to a Ph.D., funded is a guaranteed admit to a pro programme that will on average make them more money.

One should do a Ph.D. if one will not regret it even if one never holds a faculty position, i.e. it's more consumption than investment, or if one got into a TOP programme in one's chosen field. Economics seems to be in the process of collapsing from top 5/6 to top 2, CS has been 4 for a long time, but the principle is the same everywhere.

Re: Why (Not) Do a PhD in Computer Science?

#19

Post by YPetrov »

I like some of the reasons the OP has listed (great skill set, intellectually curious) and because of those I was planning on doing a PhD after I finish my BSc (4th year now).

However, after careful consideration, I realized that: 1. I am not entirely sure if my enthusiasm will be enough to push me through a PhD 2. You can obtain a great skill set, meet loads of smart people and work on intellectually stimulating problems in industry, too.

That's why I plan to work for some time and then re-evaluate my situation and maybe apply for a PhD, who knows.

Re: Why (Not) Do a PhD in Computer Science?

#20

Post by daigoba66 »

simonh wrote: post #10
sc0rb wrote: post #8

Earlier quoted context omitted.

I know of three PhD's in comp sci that went straight into Goldman Sachs (and one hedge fund) where they immediately started earning far more than those of us that left Uni with our Bachelors. They work in high frequency trading mainly using Java.

Interesting. I wouldn't have though Java was suitable for HFT due to garbage collection pauses playing havoc with your system determinism and performance reliability. I supported a derivatives trading system written in java that was canned due to performance issues, and where I am now we had java FIX protocol gateways that we've replaced with C++ components, again for performance reasons. I'm not saying java has perf…

Check out LMAX. They've built an exchange that runs on Java. Their engineers are constantly talking/blogging about getting Java to perform well. "Mechanical Sympathy" is the buzzword they've coined.
Post reply on HN