Slashdot Mirror
Google's Ph.D. Advantage
Frisky070802 writes "The New York Times reports on Google's success and desire in hiring Ph.D.'s (free registration required). It says that Google's willingness to let every employee spend 20% of his or her time on an independent project is a compelling motivator and that they estimate that Google has as many Ph.D.'s working for it as Microsoft, which is 30 times larger. How many other companies put "Ph.D. a plus" in their want ads?"
Besides, I'm guessing that a lot of those PHD's independent projects have something to do or might eventually be integrated into google (PHDs researching information retrieval, web page ranking algorithms, you name it).
Google is proof that using a smarter aproach is often the best way to solve a problem. If Google tried to use the naive clustering model their expenses would have massivly higher and their scalability and fault tolerance would have been much lower. It seems that Google realizes that the best way to hire and retain the people that will continue to come up with the smarter aproaches is to offer them things that not many other employers are, time to do what intellectually stimulates them for instance.
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
Although having an advanced degree is great, some of the best tech sector innovators come without advanced degrees. -- Also most employees spend more than 20% of their work time on personal goals anyway.
A bunch of Tech Stuff
How many other companies put "Ph.D. a plus" in their want ads?"
How about: Every company which does any kind of research?
Seriously. In areas like biochem, getting a job (or at least, a good one) without a PhD is near-impossible.
If you have a Ph. D and you're working at Google, you've got a great job. Ph. D jobs are worth the work for the degree, believe me. However, don't think you'll just be able to glide into getting that degree like you can with a BS... because professors will not just let you out! A Ph.D is designed to figure out which people actually can be creative and think of new stuff, and to keep out the "Ivan make basket" (you need communications skills) or "i learned it in 24 hours, and I think I'm a god now" (how many patents do you have? I thought so) folks.
stuff |
you seem to suggest that all google people do all day long is code. yeah, you're correct in naiveness. a lot of mathematics go into the type of stuff they do (the company founders were PhD dropouts, and the algorithm was something they were working on for thier theses).
even their cluster for their massive search index probably took some mathematical experimentation to determine a proper arrangement and setup.
Yeah...but Google doesn't need programmers. They need brains. Sure it is easy to write code and do it well without a degree in Computer Science. But to create an algorithm that effeciently searches millions if not billions of pages and returns the most relavent thing throughout the ENTIRE Internet is a little daunting and takes someone who has had some advanced training.
I tried for 5 years to come up with a clever sig...only to realize that I am not clever.
> advanced degress != coding ability/work output
If your measure is number of lines of code per day, then perhaps not.
If your measure is new algorithms and technologies that no-one has ever thought of before then I'd say the advanced degrees are a little more pertinent.
- For the complete works of Shakespeare: cat
The question is: is it really relevant for most jobs? I doubt it.
Anyone can hire PhDs. Even the government. But there may be a corporate culture that doesn't take risks, that cares too much about short-term profit, that is affected by political considerations. In Google, the nerds seem to run the show. They have the business people, and great branding. But the technical side of things is the priority.
Mod parent up!
At the same time, there's the group of people who can indeed code as well as, or even better than, you but never followed a formal education on it. Thefore they lack the precious PhD title.
:)
Just as a PhD is no guarantee that the person will grok what you're hiring them for - even if it's supposed to be right down their lane of education - the lack of a PhD doesn't guarantee that the person will not grok what you're hiring them for.
Of couse the odds are in favor of those with PhDs, not contesting that
I've worked at two universities, and I'm still happy with my undergrad.
The Phd: an exercise in self-aggrandizing behavior with little application to the real world. In the event you actually research or do something worthwhile your expertise is basically a very tiny narrow slice of the pie in your discipline in which you possess astonishing depth, and you are likely no more knowledgeable about the rest of your field than a masters candidate.
I've worked with a number of Phd candidates in computer science, chemical engineering, history, and life sciences, and then EXPECT (yes, I said expect) one of two things to happen when they graduate:
1. A company offers them quite a bit of money to do the research that *they* love
2. *poof* Tenure track faculty position
in reality now, its usually
1. Teach as an adjunct
2. Try to convince private industry that you're okay taking that 60k a year position as a chemical engineer.. I'm not overqualified, HONEST!
I think the most perverse observation I've made is that it seems like MBA's and doctoral business students have no trouble getting work around here. How depressing.
What is innovation? - you can be coding monkey without a PHD - sure. If you like it then don't get a PHD. But where has the real innovation come from?
The transistor? Nuclear weapons? Drugs that save your ass? What other technology came out of Bell Labs?
The real innovation in our society is done for the most part by people with PHD's. Amazon.com, eBay - these are small innovations compared to the above. The groundwork was laid by the PHD's creating the underlying technology.
Boris
I'm not really trying to crack a joke here, but honestly: What are the chances ANY competent person is going to stay with a Help Desk job for any significant period of time? The customers are often frustrating, the pace can be exhausting, the work rarely has long-term personal satisfaction associated with it... If you get some PhD, hire him / her and feel very lucky to have a (presumably) competent employee for the few months that they are with you.
Hiring is an investment and I need to be able to see a return on that investment.
Get use to the "would you like fries with that" crowd, then. Face it: Help Desk is no ones ideal job. Why would anyone stick around for an extended period of time?
"Who are in control, they are not in control of anything - they don't even control themselves!" - Glen Beck
You are probably correct, from your point of view. A self taught coder, even a guy with a BS/CS or BA/MIS degree in his first year in a code writing position is likely to write a good tight sort routine or a nice tight SQL statement that is 12.7% faster than the bubble sort or select statement an older guy / guy with a few years of experience / guy with better 'credentials' might use - same way some AMD/ATI fanboy is going to put together a computer that uses liquid cooling to keep his overclocked record setting system running circles (18.9% faster!) around the off the shelf Dell system an older guy might be using.
/grin.
What you are describing right now is a very narrow scope of vision. It is perfectly ok, and even expected from a one year guy - don't get me wrong as I'm not bagging on you. But you are seeing instant gratification, lines of code per hour, faster embedded loops and search routines, and frames per second. What you are not seeing, if I had to guess, is long term maintainability, group cohesion, the ability to integrate different routines together or reuse the existing development effort going forward, the overall architecture of the bigger system, scalability, usability in a business environment, reduced downtime when problems do occur.
In the same way that the overclocking crew can make a single uberMachine run 12.6% faster than a machine off the shelf, a tightly focused coder can write small blocks of code that are quite a bit faster than something written by an old school coder. From a business perspective, however, neither is particularly attractive when considering a large scale rollout of a massive business initiative. You simply can't have users running computers that sound like jet engines to keep their overclocked CPUs cool, and you can't have coders winging it to shave CPU cycles at the expense of long term stability, usability, and interoperability. Sure, you can read your in-line assembly and make it work - but can the guy over in maintenance keep it working without screwing it up or needing to rewrite it from scratch because he doesn't understand what it does?
When (if) you stop to think through all of these things you will take longer to write your individual lines of code than the next generation of hot coders. For every five days on a project, a full day needs to be dedicated to understanding what the customer (internal or external) needs and envisioning how you will design it. A full day needs to be spent doing documentation (documenting the code, user dox, developer design and intent, interaction conventions, installation, maintenance routines, etc.) and delivering the product. A day designing the system architecure, and two days actually doing the work come between the envision and delivery. In theory you could sit down and do the actual 'work' in two days, but someone has to be responsible for the other stuff - not doing the other stuff is why projects fail.
A day will come that you decide that hand building your own machine and getting an extra 7 fps isn't worth the hassle and you will just order a Dell. There will also come a day that you spend time documenting how you understand the customer's expectations and go over that document with the customer before you start designing how the system will work, and you will do that design before you start to write the code. And there will come a day that you write an official separate document describing the code you just wrote. Look forward to that day, consider it your next Graduation Day, and celebrate that day. Because the day after that is the day the youngsters start hassling you because they code faster than you do
Glonoinha the MebiByte Slayer
There is a BIG difference between coding and what you would want a PHD in CS for. Shure there are lots of people that can code most applications. It does not take a PHD two write a CMS, accounting system, point of sale, or even a spreadsheet. We just hired a programmer with a BS in CS. He did not know what a hash was! I bet he could not code a quick sort to save his life much less decide which sort to use for a given task. Now if you want to set up a server farm that can handle billions of searchs a day then you might want to invest in a PHD or two. A person that has a PHD might not be any better than a really talented person with out one but you can bet that a person with a PHD is not dumb, or lazy, and knows how to learn.
See my blog http://ilovecookes.blogspot.com/ for light hearted technical information.
"3M for example....you're allowed to work on whatever you want for 15% of your time.
At 3M, you used to be allowed to work on whatever you want for 15% of your time. Thanks to the new CEO/regime from GE, the 15% "Innovation Time" is quietly going the way of the dodo. The focus on stock price over all else (such as real, tangible, actual profits) will be the death of many a formerly powerful and truly innovative company, I expect.
He now works as a computer programmer.
This may seem a little weird, but if you think about it, a PhD [hopefully] shows that you're willing to apply yourself to something and do hard work. People with PhDs should be the most intelligent of the bunch, as they managed to get the thing.
So Dad's PhD is a prestige degree - from Oxford, no less. It shows that he has skills beyond merely chemistry.
im in ur
A PhD isn't about learning facts. It's about learning HOW to do research. It doesn't matter that one's topic is the The Hydration of Tri-Calcium Silicate". The important thing is that when told to find out something new about "The Hydration of Tri-Calcium Silicate", you can do so. The same person can also be told to find out something new about search engine algorithms, and hopefully do a good job of it.
If you're good at what you do, there'll be good jobs for you no matter what path in life you choose. If you're a lazy slackabout, then you're screwed no matter what. There's no "right" or "wrong" answer about whether a PhD is a good choice -- it's about whether it's a good choice for YOU. This is the real reason why people tell you to do something you love -- chances are, you'll be enthusiastic about it and do it well, and success will follow naturally.
I am about to finish my Phd in CS and during these long years, I came to realise that part of the Phd process is (maybe) to figure out what is this all about... be able to answer questions of the form "Does it help me to find a job?", "Should it be useful?" etc.
:-)
My take on this is as follows... It's not about finding a job... it's not about adding another bullet in a CV to impress someone... it doesn't have to be useful or practical.. it doesn't have to cure cancer (although some people do this for a phd)...
I think a phd is a long thought exercise. You prove to yourself (and to a bunch of other people) that in a finite amount of time, you can understand an area, the issues involved, and you can come up with something innovate, something new... a new problem or an new solution to an old problem...
how to get a job after all that, is an orthogonal issue... maybe deserving another phd...
I do computer support for an engineering department on campus. Means I deal with supporting PhDs (and masters students and undergrads). For many of the PhDs, this isn't an unfair generalization. They are so focused on their one little area of expertise, that they seem to loose all basic knowledge. This is an engineering department here, so people should have a little technical skill. None the less I've solved printer problems that stumped a room full of masters and PhD students by turning the printer on (really, twice). They ought to have the basic electrical knowledge and problem solving skills to figure this out. The DID at one time to pass the undergrad courses.
Now that's not to say there aren't some really smart PhDs out there. We have them here too and they are fun to work with. But there are plenty that aren't.
Working here has really shown me that having a PhD doesn't mean your smart, just means that you could play the game long enough and well enough.
I have a MS in EE. I was offered a scholarship to do a PhD. However during my MS thesis work I worked closely with PhD candidates and suffered perhaps a fraction of what they did. Don't get me wrong, I enjoy what I do, but I eat corporate shit for $$$ alone, so I declined. My observations are as follows:
1) PhD is a lot of work for yourself, and 1000x more work doing your professors busy work (papers etc.)
2) PhD slave labor wages are less than those of any given malaysian factory child if you count the total number of hours worked and divide that into your scholarship/stipend/grant/etc.
3) If you are not a US citizen/permanent resident and are on a scholarship to get a PhD in the US, you are fucked. Bring the vasoline and bend over.
4) If your goal is simply to get a degree to get more money, stop at your masters.
5) If your PhD is not in a subject actively investigated by the corporate world be willing to accept an academic position after getting your degree, or find another subject. It's heartbreaking to see people get their degree and realize they are either stuck in academia or worse, take a job in industry doing work outside their expertise making the same they would have as a masters (i.e. degree worthless).
6) If at all possible GET A COMPANY TO FUND YOUR PHD! This is harder now than it used to be, but it is THE way to go. I can't recommend it enough, if I personally thought there was money in a PhD this is what I'd do myself. If your professor administrates whatever finances your degree, and you are above broccoli intelligence, he WILL try to hold you as long as he can (5-7 years in most schools). If your company is paying the bill they are quite good at getting you in and out ASAP. Avg stay of corporate funded PhD students in my experience was 3 years. Do this!
7) Stupid people can get PhD's far easier than smart people. Simply put, professors want stupid people out of their hair, if they can't wash em out, they graduate em. Just like elementary school.
who owns the fruits of their research. Most employment contracts in high-tech companies are pretty anal about that. No matter when and how you develop the code, even if your own spare time and using solely your own hardware and software, the company 0wnz0rz the code. And there are also provisions about conflict of interest...
It would be interesting to know how google manages all this mess.
There are two reasons to IPO--to generate capital to expand or to cash out. Certainly I can't image Google needs the former, and while I don't begrudge anyone the right to cash out on their creation, I hope they realize that by definition, they're giving up ownership. Maybe they're strong enough leaders, and will start off with enough shares to be ok--I certainly hope so becaue the list of technology visionaries who were ousted from their own company is already too long.
I guess I am cynical today.
Vote Quimby.
If you believe that the sole reason for getting a degree is to gain knowledge, then no. If you believe that a university education is about learning to learn, then yes.
A BS (BSc, BA in the UK) demonstrates (in theory) an ability to follow a prescribed course of study at the pace set by the lecturers but with the self-discipline required to go to the library rather than goof off. You should make a good worker bee who doesn't need to be continuously supervised.
A MS (MSc, MPhys, MChem ...) demonstrates an ability to function independantly within broad parameters to achieve a general objective set by your supervisor. You should be capable of working at a remote site without seeing your direct boss for six months (and you should be capable of picking up the phone when you need help - rather than just sitting and stewing until someone demands to know what you have been doing for months).
A PhD demonstrates that you can determine your own goals, demand information and contributions from a wide range of individuals (even people who are senior to you in an organisation), set your own schedule, work towards a project goal that is years in the future and say with a tough project longer than some people stay at one company in Si Valley (at least during the "new job every six months" boom :) Furthermore, you have demonstrated that you don't need someone to have done it before - you have proved you can create something original!
Of course, these are all grotesque generalisations, and I know several PhDs I wouldn't trust to drink a glass of water without close supervision and paramedics standing by. Equally, there are other paths that demonstrate the same skill set. Furthermore, it can be hard to maintain one's non-conformist, independant spirit when one is producing a PhD thesis that must, by definition, conform to your examiner's views.
I've found that there are two types of CS:
Type 1 is the type you're thinking of. None of them generally need anything beyond a BS, and their coursework was mainly focused in 'software engineering' disciplines.
Type 2 is the type that develops new algorithms and does research. They need the postgrad work, and their coursework focuses on algorithms, math, and suchlike.
I really think that we need to split these degrees apart; the first should become 'software engineering' or something similar, to help convey the difference between the application-oriented (engineering) and the theory-oriented (science).
(disclaimer: I am a EE who does algorithms; I work with type 2 when I'm doing algorithm design and type 1 when we need implementation, and appreciate both)
---
Mod me down, you fucking twits. Go ahead. I dare you.
(I read with sigs off.)
I worked as one of two junior programmers at a startup (now dead), which at its height had approximately 10 or 11 people on the software side of things), and 5-6 of those were PhDs (and 6 or so hardware guys, I think half also had PhDs). Most all of these guys had very impressive resumes/CVs, and were being paid enormous salaries... though some were light on working in industry
Anyway, we were a small startup and I had heavy interaction with basically all members of both the software/hardware teams working on basically parallel processing. To make a long story short, having a PhD didn't lead to a correlation between being good at implementation OR design, or really anything. Out of 5-6, only one was truly good at actually programming/implementing, but I figured their strength was in their ability to help out designing some of the horrendously big and complicated stuff, and the algorithms underlying. However after over 2 years of work at this company, many code reviews, design meetings, etc, it was pretty clear having a PhD in EE/Comp Sci didn't particularly mean you had a handle on algorithms or design, either. I still vividly remember a presentation over a design prototype one of the PhDs had developed on his own (approximately 1-2 months of solo work) that was absolutely ripped to shreds at the most fundamental levels during a code review meeting. It was actually embarrasing to be in the room.
Anyway, my experience there pretty much killed whatever mystique or respect I previously had behind having a PhD. To me it seems to mean you 1) Did a research project, which may or may not have been relevant to anything at one point 2) Had 5+ years to do it 3) May or may not have learned a lot about the subject. I don't mean to belittle it, but I think in general theres a *lot* more fluff surrounding a PhD than meat.