Hacker News new | past | comments | ask | show | jobs | submit login
Ph.D. 2.0: Rethinking the Ph.D. Application (jeffhuang.com)
187 points by masterofmasters on Nov 14, 2013 | hide | past | web | favorite | 142 comments

This reminds me of YC's "Apply without an idea" experiment. Both experiments seem to arise from a innovator's insight that a whole class of otherwise highly qualified candidates self select out because of a self perception that they don't "fit the mold". This is especially true of YC, where there are so many articles every day that tell the story of a founder who "saw a problem, and set off on this quest to solve it" that people who don't have an idea on hand when YC applications come up can immediately self exclude. Never mind that many great founders pivoted more than once on their way to success.

I'm excited to see this experiment "meet the marketplace", and see what pans out. Given the small number of PhD students these top programs take each year, just yielding one or two great people into the PhD program that otherwise wouldn't have applied seems like it'd really move the needle.

[Disclosure: I went to undergrad with Jeff and we're good friends. He really is a great "mold-breaker" himself and I'm excited to see how a great "hacker of systems" in the best sense of the word changes academia during his career]

+1000, and latching on to this top comment for more visibility ...

also check out jeff's related post on bringing startup culture to academic research! http://jeffhuang.com/adopting_the_startup_culture_for_resear...

An offer letter from Google or YCombinator (where you are the tech co-founder) which serves as evidence that you passed a challenging programming interview at Google, or that YCombinator believed you would be successful at developing your company product.

Interesting. Because I have no intention of moving to a Google or YCombinator location, I never applied to either, but this makes me wonder if merely having been offered an opportunity at such a place could have value. Even if one never intended to follow through.

Which might not be a great thing for the folks reviewing applications...

In the UK in some areas such as finance and business there are many people who apply to be an Army officer, and go all the way through selection, but never intend to join up. They do it just so that they can put on their CV (resume) that they've passed Army officer selection.

Not to say that it's not true, but in a decade in finance I've never heard of anyone having gone through Officer Selection to help their career prospects.

If you want to have a good start in finance, having a degree (preferably a 1st) in Maths, Physics or Economics from Oxford, Cambridge or LSE is probably the most useful thing you can do!

Isn't Officer Selection where candidate for officer training are selected, not after the full training? According to this link it looks like a ~5 day process:


[NB Not sure I approve of people using up forces resources just to pad their CVs]

I think the Army likes it because it reinforces the idea that Army officers are an elite, and it exposes more future leaders to a little bit of Army life, so they're more positive about the Army when they make decisions when they're running a business or a government department.

Of course. I have no intention of moving to Cali, but one thing I do plan to do when I go to vacation West next year is to apply to Google, possibly Facebook or/and Amazon. If I get the offer letter, I can frame it and use it as negotiating tactic here in the midwest. It's almost as good as saying "I worked at Google"

Unlike employment, there's very little to stop this probably-currently-useful tactic from being rendered valueless by fakers.

Possibly better: you don't have to actually work at Google. 80% of the benefit minus 1% of the opportunity cost.

That's a very interesting perspective on working for Google. That is, that it's a waste of time. In my experience, working at Google is a blast both in terms of compensation and exciting projects.

Depending on what you want, what you can do, and from a certain perspective, any job is a "waste of time". For one thing, you will almost certainly spend the majority of your time working on someone else's grand idea. (By necessity, most of the people working at Google are doing the same kind of things that everyone is doing everywhere. The environment may be better, but it's the same stuff.)

As one of my favorite supervisors said, repeatedly, "It's not supposed to be fun. That's why they call it a 'job'. Now get back to work."

If you were luckier than others it doesn't necesserily make you an example to be followed.

I wouldn't call it luck. Google is consistently ranked as one of the top workplaces. Out of a large number of my co-workers at Google, almost all of them seem to enjoy their work. Very few people are unhappy or seeking to leave. The main difficulty is getting in.

Idea: Apply before you plan your vacation, then extend you stay when they fly you out to interview. Free flights!

The other question is- if you have an offer letter from Google or YCombinator in hand, would you really want to go to Brown? At Google it will be like going to grad school except you will be better funded, working with more consistently smarter people, and you will be getting paid instead of racking up debt. As a YCombinator founder you'd basically already be in the position that a Brown computer science graduate hopes to one day get to. I like universities. I hope they figure out how to stay relevant.

> ...racking up debt

This is FUD. While there is arguably opportunity cost involved, it is not at all difficult to avoid "racking up debt" in grad school, at least if you're at the PhD level and in a STEM field. It is standard in such programs to get tuition waived (so, no cost) and to receive stipends for assistantship or fellowship (so, you get paid).

To get into specifics: I got my PhD in CS from Brown ten years ago, and my stipend was around $20K (I think about $22 or $23K my final year). During my time there the students clamoured for, and received, a healthcare benefit (free for the student, and they could pay extra if they had spouse and/or dependents). Because I was lucky enough to be debt-free going in, I actually finished grad school with about $15K in savings, but even my friends who had debt from ugrad were not accruing additional debt (and at the time were able to defer payment interest-free, although I gather that's no longer true). I was not exactly living the high life but was able to afford a decent apartment, ate out fairly regularly, and kept myself supplied with geeky tools and toys (laptop, wifi, etc).

So no, "racking up debt" is not a worry here.

How exactly would you be racking up debt as a computer science Ph.D student? You don't pay for tuition, you get paid a pretty decent stipend + health insurance and you probably will do many industry/research internships (where you get paid very well). I say this with experience, as I am a CS Ph.D student and make more than enough to live on with a modest lifestyle and I have no debt. Plus if you get an outside fellowship, which most successful students do, you'll essentially be able to work on just about anything you want (I don't think this can be said for a random Google employee, however this is probably true for a YC founder). The Ph.D lifestyle is pretty great imho...at least in CS (non-CS, that's a whole other story).

Google pays a decent salary- not the highest, but fair when considering the cost of buying a home and living in a city near a Google office. I have never heard of a Grad student getting a stipend in the amount someone who studied CS for 6-8 years should be able to get outside of academia.

I'm not saying doing a CS Phd program is not worthwhile- it still is. But, if you have an existing offer at Google or YCombinator its like going to the NBA straight out of high school. Certainly there are good reasons to go to college. But at the end of the day, if the NBA is where you're trying to get to- you take the chance when the opportunity presents itself. Grad school will always be there.

so you are saying that the only reason to do research in CS is to land a job at google? huh. strange.

also, won't google always be there?

Get paid a pretty decent stipend + health insurance? Most Ph.D student got less than $2000/month before tax, even less than $1500/month and should pay for health insurance by themselves. OK, decent, how to define decent.

The NSF GRFP provides ~3k/month plus tuition and health insurance are covered. If you do a 3-month internship each summer on top of that you can get an extra $20k-30k easy. I am not saying you'll have the same lifestyle as a Google employee but I think it's hard to argue that you will go into debt. Maybe I live a more modest lifestyle than others. Also all the schools I got into when I applied to grad school would pay health care/tuition, that's standard.

If I may, I found the NSF GRFP process unsatisfactory. For example, I was rejected for two reasons: 1) because I did not publish prior to applying, and 2) because the reviewers thought that my research, although it had intellectual merit, did not have broader impacts (i.e., "the potential to benefit society and contribute to the achievement of specific, desired societal outcomes").

I proposed research on how to improve upon string searching, which according to the reviewers, did not benefit society (??).

The reviewers all have PhDs, I believe. But somehow, I do not feel they are qualified to review the applications. They are looking for specific things cross the check-boxes.

The second point is that stipends are taxed. So, $3k looks like a lot, but post-taxes, it's trimmed. Also, if you look at the OP's posting, $300k for ~5 years for a student. That does not mean the student receives $60k/year for living.

The third point is that someone immediately out of undergrad at 20 might find PhD / below poverty level stipends enticing, but once you make 150k+, you really don't want to hand control of your life over to your advisor's whims.

The fourth point (which is a corollary to the third) is that I know quite a few miserable PhD students, whose advisors control large aspects of their life. For example, a few of them were house-sitting their advisor's newly purchased mansion overnight in sleeping bags. These were foreign students.

In which case you move to #2 which (to me) seems much easier to get: A Github or other online repository of your source code demonstrating contributions to open source projects or impressive projects of your own.

True - if the open source project you've contributed significantly to is, say, a well-acclaimed machine learning library, and you're applying to a machine learning Ph.D., then the sum of your accepted pull requests might be a reasonable proxy for a recommendation letter from the (presumably well-acclaimed) maintainers of said library.

It does not seems so to me. To show that your interests align, you do one of the challenges that Jeff lists in part 2. Part 1 is only to show you are a strong programmer/ambitious so your projects can be in anything it seems.

if you do that you might as well forget the PhD. Schools don't want to teach you, they want to accept the best applicants. Applicants who work on open source projects and can get a job anywhere. Applicants that became great learning from somewhere that is not the school that he is applying to. Whereas as a post doc you'll be someone's bitch working for 40K per year. Screw the PhD and whatever it is supposed to mean--that you did all your stupid homework assignments and kissed people's asses for three letters of recommendation and that you "gamed" the GRE.

Sarcasm detector might be failing, but PhD work has very little to do with homework....

to get into a phd program, i mean. a phd from a "top school" means you did your homework in college and high school.

That is an interesting dilemma, yeah. To be a sharp homework-hammering undergrad is not to be a good researcher & fuzzy situation manager. It's why there's such focus on undergrad research these days, it helps identify the capable people and build their skills (as well as getting some amount of budget help).

The professor at Brown clearly enunciates some of the issues with the current system quite well.

But to make it a meaningful experience (and get out of it with a PhD) you need a wholly different skillset. It's clear that you have never seen a kid who destroyed the productivity of a whole office because he treated his colleagues Just Wrong, and that you have never seen a fellow who just couldn't transfer what he learned in class to real-life application.

i've seen people who didn't know what a lac operon was get a PhD in biochemistry.

That's harmless. But if you have someone who continuously messes up shared equipment, leaving it for the next person to fix, and who won't improve their attitude, no matter how often you tell them, that's deadly for morale.

then you should leave your phd program for a garage.


That's not as true anymore. If you did some interesting independent-research projects as an undergrad, which you got published as conference papers, that strongly overrides your undergrad course grades nowadays, at most schools. Basically publications trump everything, and grades are only used if you don't have any publications.

I'd be curious if the folks at Google and YC would be flattered by this, put off, or both.

If it works, it's clever. But I am definitely not flattered and wouldn't be personally happy about knowing someone did it. The 45 minutes I spent interviewing you and the more than an hour I spend carefully writing up my interview feedback is totally wasted. Well, that sucks for me.

Thanks for sharing. I think you hit the main point - it winds up wasting a lot of people's time, and also gives short shrift to others in the process. In a sense I put this up with the moral equivalent of practice interviewing with real companies. It's not illegal, but it isn't behavior one would advertise, and it's hard to call it ethical.

It would be different if the answer were, "You've worked 2 years for Google or started a company that went through YC" because that's no different than a company asking for Big 4 Accounting experience when hiring a controller. (Having a Big 4 offer letter wouldn't suffice)

I think this is a very interesting idea, and I look forward to the results of the experiment!

Allow me to take a moment to reiterate some of the comments I left on the Georgia Tech master's thread [1], specifically as a student who has taken graduate level computer science courses at Brown University with the majority of his experience working in industry.

* I went directly into industry in 1998 working at a software/systems engineer for a telecom working on Solaris deployments. I worked full-time and paid for my undergraduate education which I pursued part-time and which took me 9 years to finish. I majored in "information technology," as there was no computer science online program at Northeastern University. I didn't take any theory courses at all, and the majority of my course work was programming in languages like Java, C++, and C# (also COBOL, and the like...)

* I applied at Brown University after talking to the admissions and computer science departments multiple times, in which they told me I wouldn't be able to pursue a master's due to lack of independent or undergraduate research or an undergraduate education in computer science. The process of being told this was rather unfortunate, as the responses I received via e-mail telling me it wouldn't be possible were ended with "Sent from my iPhone."

* At the time I talked with them, I had been working at Basho Technologies for a year on Riak, an open source distributed Dynamo-style data store, as well as serving as a maintainer of rubygems.org.

* I was finally able to get accepted as a "non-degree" or "special" graduate student, which is allowed to take courses at full price. This role exists primarily to allow students to determine if they would be a fit for graduate school, at which point I would then have to re-apply for degree seeking status. This was possible because of an independent meeting I scheduled with Shriram Krishnamurthi, who, based on my industrial experience expedited the process along.

* Since starting as a "non-degree" student, I've been heavily motivated to attempt to stand out from other candidates for when I eventually re-apply to be a degree seeking student by independently publishing papers, publishing a blog, organizing a podcast, and speaking at conferences.

[1] https://news.ycombinator.com/item?id=6510142

what kind of research are you interested in for Ph.D.? i can't seem to find your academic info from your HN profile. email me if you want to chat more.

Distributed systems.

My website is located here: http://christophermeiklejohn.com.

If you haven't already, you should really talk to http://cs.brown.edu/~rfonseca/

... And go fellow Huskies!

Love the co-op that NEU is so well known for. Excellent way to earn a degree while gaining real-world experience on the job :-)

Hmmm. This is an interesting spin the Brown Professor has taken on the PhD application process.

For those interested in a career in research and development, a PhD is almost a must. Getting on to a prestigious PhD programme is incredibly competitive, and as someone who has gone through the whole process and seen it from both sides, I can say that I've seen individuals who'd make great computer science researchers not being given an opportunity because they don't tick the right "boxes".

Sure it's not perfect, but worth a shot if you're interested.

For those interested in a career in development, a PhD is rarely helpful. For the PhD candidate, it can even be counterproductive, as any gains in initial salary will be more than offset by the lost salary during the 5-6 years of the PhD.

Only for a career in research is a PhD important.

i will vehemently argue that a Ph.D. in CS is one of the best ways that a young hacker can spend his/her 20s, regardless of which career they end up pursuing in the future. pursuing a Ph.D. + interning at several companies during summertimes = amazing combo for a wide variety of future careers.

the ONLY enduring downside (which you mentioned) is the lost income.

see: http://vimeo.com/73178238

> i will vehemently argue that a Ph.D. in CS is one of the best ways that a young hacker can spend his/her 20s

No need to argue that point, because I'm not arguing against it. Look at the comment that I was replying to:

> For those interested in a career in research and development, a PhD is almost a must.

A PhD is a must for a career in research. It is not a must for a career in development. We use the term "R&D," but those are quite different things. There's a lot less "R" around, and a lot more "D."

When I say that the PhD is "rarely helpful," I do not mean that it is totally useless. I mean that the person who would otherwise have done the PhD is probably a self-starter who would've done other equally-valuable things in the meantime. Different things, certainly. But unlikely to be worse than what you would've done in the PhD. Thus, it is not helpful compared to what else you could've done in the meantime.

For a career in the "D" side of computer science "R&D," it's about your technical skills and experiences. Doesn't matter if you acquired them while doing a PhD, or in a company, or working on open source. This is very different from a career in research, where it is much more important to have a piece of paper with "PhD" written on it.

"I mean that the person who would otherwise have done the PhD is probably a self-starter who would've done other equally-valuable things in the meantime." -good point! i agree with that a lot. i see what you mean now.

this video comes from a professor, good joke!


(also, my first job out of grad school was in industry and had nothing to do with my Ph.D. work)

Hey, I read you book! I thought you will never go academia... things change...

whoa thanks! then you probably know that i've thought a lot about this decision to dive back in :)

The parent comment should probably be read as 'research and development' - working for IBM's research arm, Microsoft Research etc.

And I read it as exactly that.

I was pointing out that the PhD is useful for "research", but not for research "and development."

In my experience, the people who can do the D are better at the R.

> In my experience, the people who can do the D are better at the R.

Agreed. But less so, the other way around.

Which is why development jobs don't reward the PhD enough to pay for the opportunity cost. It counts as 6 years of industry experience -- which you could've gotten in 6 years of working in industry.

That is true to an extent. But, most people don't do PhDs to end up doing purely software development.

Do you have any thoughts on how to get started on this road? I feel like I missed the boat sometimes, though I'm using doing interesting-ish work professionally and in open source. I'm not interested in chasing the framework-of-the-week, I want to work on tools that advance how we think about software.

This may be something I do later in life, but I feel very wired for R&D + tooling, and a bit locked out of these types of jobs without a PhD.

Its never too late to get started down the PhD road. At least in my department, most PhD canidates aren't fresh out of school, they've usually taken at least a year or two and worked in industry of some sort and then applied.

My own experience: I had had no research experience in school and worked in industry for a bit before decided that it bored me to death, but I felt shut out from PhDs at the time. I started looking into research that might be interesting and sent out a bunch of emails to professors at universities explaining who I was, my goals and asking if they were looking for a research tech or anything. I probably sent out like 30 letters, got a handful or responses. I took a bit of a pay cut, but I got a fantastic job that I loved as a research programmer for a professor. I worked there for a while, worked on some interesting research, read a ton of papers, got a handful of co-authors on papers, and then was able to apply to grad programs with a strong application.

Pick a specific area, read lots of papers, find out what the bleeding edge statements 'we can do X' and 'Y would be nice and seems within reach but we're not doing it now' are there.

Do something interesting and novel based on that - not neccessarily a detailed implementation, but at least a quick&dirty proof of concept; and then also describe and publish it as a formal paper, it's not particularly much work compared to actually doing the stuff. For example, if you "want to work on tools that advance how we think about software" then picking some specific narrow area of how, say, Bret Victor's concepts could/should be implemented in real practice would be an interesting topic.

All of this can be done w/o a PhD as a hobby/side-project, and such a project+publication would also be a ticket to a decent PhD spot.

mattgreenrocks it's never too late! But, I agree without a PhD R&D jobs are difficult to come by.

It sounds like you know what you want to work on. My advise would be to do some more reading in the area and find research papers in your field. Soon you'll know how to extend or improve on prior work. At this point it's a good idea to identify schools and supervisors you want to work with, especially those whose papers you've read or reach you are interested in extending. You might even propose a research collaboration with them based on the papers you read. This is a good way to open channel of communication and get your foot in the door.

There's a great deal that can be said on how to get into a PhD programme. Let me know if you want to continue this discussion offline.

I'm surprised that admission into Computer Science PhD programs is so competitive. I would think that with the strong job market for computer scientists, there would not be as many students who would want to make the commitment to a PhD program over other job and entrepreneurial opportunities.

I am not surprised that it is competitive (PhD's in general are kind of oversupplied), but the form the competition takes:

"Maybe you didn't publish as an undergraduate.... Maybe you can't even write a very compelling research statement yet."

When I was applying, no undergraduates had published research. And the "research statement" could hardly be described as "compelling"; that was one of the points of the breadth requirements for the graduate program---my eventual research topic was completely unrelated to my interests or knowledge coming out of my undergraduate program.

It's a positive development that admission into a decent PhD program requires research experience. Some people somehow cannot make the jump from classroom exercises to production work, and you would rather that they don't get accepted, instead of having to throw them out a year later.

Besides, a research lab is a workplace after all, and some kids are completely missing the social skills that are needed in a working environment. You can't have that sort in a PhD program either.

> It's a positive development that admission into a decent PhD program requires research experience.

No, it really really isn't. The vast majority of undergraduates have no effective access to research experience, even many that are fairly inquisitive and are receiving a pretty good education. This is especially true of undergraduates who are unable to afford tuition at top institutions (and increasingly even in-state public university tuition is unaffordable for the flagship universities that would be considered "top institutions).

Requiring research experience in advance of grad school turns grad schools into an elitist echo chamber.

I have to say, I agree with this. You risk optimizing to early around a local max, not a global one. Primarily due to lack of priors. However, I will be the first to admit true research is seriously a different aptitude than acing a lecture course. Advanced degree candidates should certainly be screened somehow for the latter. But my gut is that this is just optimizing for grant extraction. Truly powerful work is always second fiddle to the "more money" hampster-wheel that is peer-reviewed academic research.

I would argue that it isn't, because it simply pushes the boundary back, and possibly making it even less appropriate.

I can't see having valid or "compelling" research experience as an undergraduate while taking a full load of classes (ok, so you have to take advanced placement courses in high school, which my school did not offer even if my school district did; I'm not sure one way or the other about that) and possibly working to support yourself (I'll leave the alternatives as an exercise to the reader).

At this point, you're filtering on many criteria that have absolutely nothing to do with successful research.

It feels like requiring 5+ years of experience in entry level positions...

Entry into elite PhD programmes is as competitive as ever. Others, not so.

...and of course Brown (where the author is) is top-twenty, in fact tied at #20 [according to US News & World Report[1]] with places like Rice. It's pretty elite.

[1] http://grad-schools.usnews.rankingsandreviews.com/best-gradu...

Wow, my school's not even on the list and our graduate program for CS/CE has less than a 13% acceptance rate.

This lists are basically meaningless. However, I wouldn't argue with the top 10 or so.

I believe someone who is not a good software engineer or even can't code at all, can still be admitted into some CS PhD program.

How about theoretical computer science? That probably doesn't need one to be a good software engineer.

Well, you really don't need to be a good engineer, but the skills required to do anything meaningful in theoretical computer science PhD program are more hardcore and rare than good SW engineers - from what I see, most of the people who can do theoretical CS are already doing it; and most engineering-oriented CS researchers couldn't/wouldn't cross to theoretical CS no matter what.

I mentioned only some subset of PhD applicants, which might not be small. So there are many other types of applicants who apply with different and various reasons.

As for your question, no, I don't think that scientists in all of the branches of CS need to be good engineers. But that is just my personal opinion.

I'm not sure how these students are filtered out at other places, but in Brown's CS department, PhD students are required to pass a "programming comp" that involves a fairly intense week-long programming project. You don't need to be a superstar coder, but you do need to be able to program effectively in order to pass.

i'm sure that's extremely rare though. maybe like 1 person per year may have no idea what coding is and get admitted to a program for CS.

It completely depends on what area you're going into. Software engineering, systems, and AI students obviously need significant coding experience. HCI and theory (ie math), not as much.

i don't think it depends what area. almost every single person in all of those fields will have done some coding before. Even if that means cs 101

Right, but what I wanted to highlight - a person might be able to code quick sort algorithm or breadth first search, but unable to get paid doing real world coding - when it is the client or employer who needs to be satisfied.

So basically when you have a choice of being unemployed vs doing PhD and getting some studentship, you obviously prefer the latter path.

And I believe there are quite many people who can "code" but unable to get paid.

If you can "code" so poorly that you're unable to get paid, I can't imagine how you'd get even close to getting into a PhD program, even if your specialization is math of finite state automata or human interface guidelines.

Coding (or being DBA or analyst, depending on your area) to the level of getting an okay salary is quite trivial compared to that - from the PhD students I've seen the 'most unemployable' ones would be so not because of their ability to code but because of social/mental issues.

Theory cs could care less about coding abilities. It's never a bad thing to know how to code, but I can go though the entire PhD program without coding a single line of code. (Well not really, one of the professors in my PhD program committee recommended me to do one systems course. other than that, I just prove theorems. )

Yes, we've got a bunch of people like that in our PhD program as well - but 100% of them, despite not writing a single line of code for years, could still easily get a paying job in industry and be mostly okay (though less happy) there.

If you can prove theorems, then you can also code CRUD websites by following a tutorial, and get paid for that.

If you can't write code if your life depended on it but can write papers on human interface or methodology issues, then you can get a job as a system analyst in any large software development company and keep not writing code there.

There is also another moment: yet another CRUD job in 3rd world country vs PhD in USA. And no need to struggle for H1B.

It's hard for someone in anything remote to math/science without coding anything in their entire undergraduate life.

It's like you say no one can be admitted to a PhD program in US without first learning some English.

Many US students do not go into PhD, because, as you mentioned, due to the strong job market (and also a 20k/year stipend). Most of the PhD students are foreign.

Why would you be filling out a PhD application if you already had an offer letter from Google or YCombinator?

Because you always wanted to do academic research? Because you someday want to teach at the college level? Because you have no idea what you want to do and you want to keep your options open a while longer?

I left grad school to work in industry, but I’m not so naive as to think that everyone shares my preferences. I have friends who turned down Google offers c. 2001 to go to grad school, and they’ve never regretted that decision. (And I don’t regret the time I spent in grad school before leaving, either).

Personally, I'd want to get a PhD to be able to work in such places as Microsoft Research. I don't want to become a teacher/professor. Other than that, I really hate some perils of academic world - bureaucracy, nepotism/favoritism/elitism, ...

Google often also hires people to do work they are overqualified for. Had a friend turn down Google because it was boring compared to research.

ha, i've been giving talks on this very subject ;) http://vimeo.com/73178238

btw i'm a big fan of jeff's approach and hope it works out well.

Like Jeff I am also a young faculty member at Brown (http://cs.brown.edu/~kraskat/) and I couldn't agree more with him. The Ph.D. application process has a lot of flaws. We (the Big Data Management Group at Brown) are also always looking for very talented people and similar to Jeff we started to hand out small challenges to candidates interested in doing data-centric systems research (see also http://cs.brown.edu/~kraskat/phd14.html). In addition, last year we started a research internship program (http://database.cs.brown.edu/big-data-internship) - it is quite competitive to get in, but still easier than to be admitted to the PhD program, and the best way for both sides to determine if there is a good fit with the group and the PhD program. Finally, I would like to mention, that we actually do consider MOOC courses and github portfolios as part of the candidate evaluation.

While I agree the application process needs improvement, I think this entry greatly exaggerates how hard it is to get into a PhD program. Since when do you need to be published to even get into a PhD program? When I went through the application process, I got accepted by 6/6 schools that I applied to, with no publications and no real ideas for what I wanted to research on. In fact, most PhD students actually don't even have a research topic until the second or third year of grad school. I think more than anything, your personal statement is the deciding factor in a lot of cases. What had worked for me was to have a list of research topics and faculty from each school that you're interested in working with, and directly address them in your statement. Other than that, of course whatever supporting evidence to differentiate yourself is helpful. I thought proposing to use offer letters from Google or YC (both have much lower acceptance rates than a PhD program) just to prove you can code is pretty hilarious.

When did you apply, and to what schools?

These days, good luck getting in a top tier institution for CS without a solid track record (I.e. a publication or two at an ACM/IEEE conference, and letters of rec from reputable professors/researchers in your field). The letters of rec are what make or break your application.

I agree that it's a bit over the top, but that's how it is.

2005, Berkeley, MIT, Princeton, Stanford, UIUC, and my alma mater UT Austin. Ended up at Cal. Granted, I was EE and admitted as EECS, so maybe a bit different.

Usually the Ph.D. is aimed at being a college professor and working toward tenure. The CS Ph.D. now seems to be aimed at such a professor but also at a career as an employee.

Sorry to say this, but Ph.D. or not, it is getting clear that in the US being an employee is no good for a career.

Can get hired as an employee in your 20s, but the chances go down in your 30s, and the chances go to near zero long before your 40 year career is over. Exception: If you rise high into management, then you might be able to continue to get hired until, say, 50. And high management positions commonly don't last very long.

E.g., a big tech company might hire 100 Master's or Ph.D. degree holders, promote 1 to management, and at age 35 or so fire the other 99. Then the other 99 can wish that they could convert their Ph.D., say, in electronic engineering, to an electrician's license or had followed the path of a friend in high school who was mowing grass and now has 5 crews mowing grass and is getting into landscape architecture and commercial instead of just residential clients.

In broad terms, for a long career in the US, be a sole proprietor with a geographical barrier to entry. If want to do something technical, then be a CEO of a startup that takes advantage of your technical background. For being an employee, regard that as a temporary slot that will have to be replaced by owning part or all of the business from which you get your income.

Then, a problem with a Ph.D. is that you spend in grad school most of your 20s when you are most employable. Then to go into the job market in your late 20s or 30s can be a big disappointment because, really, the jobs are for subordinates, not narrow subject matter expert researchers. Actually, a Ph.D. can be highly resented, can be a black mark on your resume.

Be careful.

I'm applying as we speak, but in Bioeng. When I email professors and get to talking they always coach that they do NOT have control over the admissions process, that a committee acts as Maxwell's demon. Though I am not applying to Brown and do not know anything about the internals of their process, this seems to indicated that professor Huang has at least some control over the admissions. If I were professor Huang, I would be very very careful about this, as sparks of racism, sexism, and homophobia can quickly ignite into a fire. I want to be clear, I am not accusing professor Huang of this at all. Heck I don't know the guy a bit and I do like this alternative approach to the application a lot. Still, he needs to be careful.

Are you kidding?

Of course a professor has control over the admission process. If he wants you on his team, you're in, unless you have a criminal record, or GPA below 3.0 (or whatever is absolute minimum for that particular grad school).

Think of it as a hiring process, with the professor as a hiring manager.

No, I am not kidding.

Based on the conversations I have had, it seems to be typical that a committee selects the applicants, not the professors. That is what all the professors tell me.

That being said, I would be stunned if this were the truth. There is no frickin way that this happens. Here's how I imagine it:

'Admissions committee is in session' 'So, which of these bozos we gonna pick?' 'I want Joe, Schmo, and Dingus' 'Sweet, I want Harry, and Fiona' 'Ok, we all cool?' 'Yep' 'Aite, lets get beers'

Well, the committee also has a job of selecting the rest of the applicants, if there are still places left after professors made their recommendations.

Cool idea-- people inventive enough to go down an entrepreneurial route would probably be interesting/compelling candidates for academic innovation.

I actually think that creating many interesting projects + one's corresponding Github would be a better metric than getting a Google (or even YC) acceptance letter. You want people who are constantly tinkering and thinking of new ideas/projects/approaches and inventing just for the sake of inventing.

If you go too entrepreneurial, you might also be selecting for students who might be more amenable to dropping out from the program and doing a start-up (which then wastes the resources/time that Jeff mentioned he had to give to every PhD candidate).

I love this. Especially the part where it gives you a few problems that you can try to solve. Not only is a good way to find talent that may otherwise be detected. It increases the chances of finding new ways to solve some problems. (Edit: grammar)

As a person who was about to apply to PhD programs but changed my mind due to insufficient recommendations, I'm glad somebody else realizes that PhD admissions is utter bull crap.

185 dollar GRE that I can get in the 99th percentile on with ZERO studying? Grades that aren't normalized across institutions weighed heavily? Need recs from established professors who can slip in a good word for you?

Oh sorry, you were actually doing things that interested you, not necessarily things that would impress the right people. Oh sorry you took tons of hard classes at a prestigious institution and so have something less than 4.0. Oh sorry you hopped labs for a bit instead of staying with that boring CS guy that would've gotten you into any PhD program in America for 4 years. Oh sorry you spent a year or two out of college doing startup tech work without building credentials with other PhD grads but nevertheless solving research grade problems.

I've realized that the PhD admissions does not select candidates who actually want to do research. It selects candidates who want to be admitted into a PhD program. No thanks. I love science and math, and I love research, but I think the PhD application has been so harrowing for me, I won't consider it again unless drastic changes are made.

As a person who reads PhD applications and works with graduate students, bullshit.

I care a little about your GRE scores, it is a filter, you need some kind of standardized comparable. Congrats on getting in the 99th percentile (that's expected). I don't care much about your GPA. I might look at what classes you took. I will probably look at your reference letters briefly. They might be from professors, they might not. I don't care. I love the fact that you have some startup or industry experience, you are slightly more likely to actually know how to get things done.

The very first thing I read is your cover letter, then straight to your research statement which I will read in great detail. The only thing I care about is do you have a passion for research, are you likely to be successful at it.

At most research institutions in most areas of science, we have to pay our graduate students stipend and tuition, either right away or after a year. This is a damn good incentive to get students in who are going to do great research.

All I care about is getting great researchers in the door and (eventually, hopefully) into my lab. So, my advice: focus on your research statement. Tailor it to the department your are applying to, talk about the research they are doing there. But, have your own agenda. At least know what research areas you are passionate about.

What you are telling me is completely at odds with what others have told me (including faculty on admissions committees) so I imagine it varies from institution to institution. I had a pretty compelling application a year ago and my statement (in my opinion anyways) demonstrated that I was caught up with the academic literature on a particular topic and I had done some work myself implementing and attempting to improve existing techniques. I realized that I may not have worked on that topic in particular, but my hope was that it would demonstrate research aptitude. My application was rejected from 8 schools and I contacted several to see what the problem may have been.

The GRE point I was trying to make was that since 99th percentiles are expected, the test shouldn't exist at all as it is clearly a poor filter.

Did you have a low GPA by any chance? In academia, this is considered a black mark :(

Honestly, I think you dodged a bullet. Academia isn't that magical (CS PhD here). I think you get a better education reading Hacker News some days.

Perhaps this was the case. I did not do well the first two years. I hated school in fact and averaged a 3.3 or so (the engineering courses felt very soft and uninteresting). This changed when I discovered pure math and theoretical physics and I averaged almost a 4.0 for the last two years with the most difficult courses so I didn't think it would matter. I like engineering now, but I wished that the fundamentals were covered more rigorously as the lack of rigor made me initially very disinterested.

HAHA. As a PhD student, I agree with most of your sentiments. The PhD application process is nothing, try applying for fellowships. Go to a prestigious institution, pull allnighters, to get a subpar gpa, and have no time for volunteering activities. These are the criteria you need to satisfy to get something like NSF GRP (Broader Impact is 50% of the application). You might have a better shot at a lesser prestigious university with a not-so-great engineering program and being a super smart person, killing it. Then, apply for fellowship --> you will get it. It is all a rubric.

Now on to the game of academia. It is one that is based on reputation, papers, and not money. No one cares for efficiency either and no one is building crap. All they are doing is trying out different "solutions," which may or may not make any sense. For instance, "hey guys, let's apply game theory to problem X, it will be a good/novel experiment and we can get a paper out of it." Waste of tax dollars. The logical answer that solves problem X might have already been solved and can be solved using ordinary methods, but academia doesn't want that. Research labs have business models, keep proposing new complexity on complexities, very unnecessary. Overcrowded fields like computer vision reek of this.

Don't even get me started on funding and fellowships haha. I think trying to get funding in math must be the worst. Every formula has to boil down to "security this" or "missile navigation that."

I'm still waiting for a REMOTE PhD position :)

remote Ph.D. is a bad idea; the Ph.D. experience is isolating enough as-is, and pushing it remotely will exacerbate those problems. plus, innovation often happens in serendipitous in-person encounters in the hallway, lounge, and lab with students, professors, and research staff incidentally bumping into one another. see:


I doubt if it would happen at all.

I'm a full time Ph.D. student. You might, of course, know that professors interact with the students very often (except if the professor is super busy always). I interact with mine almost every day. I do work with couple of collaborators in different countries. That is because my advisor knows them in-person before deciding to collaborate. I work with lab colleagues frequently. Remote makes all this extremely challenging.

Right, that is the difference between PhD in USA versus mainland Europe - where PhD is regarded as a regular salaried job (albeit very low-paying). Basically, after some intro time with the professor, you're really on your own. Interactions happen, but nothing that can't be done remotely.

If you work remotely, there is a good chance you will be left out. Of course, "interaction" can be done remotely. But many other things cannot. Showing your face every now and then around other people is quite important. Somebody is organising a seminar? Somebody is organising a conference? Somebody is looking for someone to teach a couple of their lessons? If you are the person that never shows up, you will be the person who won't get asked, simply because you have no connection to other people. Who am I going to ask if I want to organise something? The person I have never seen and who I know nothing about, or the person I've met, where I can gauge what they are working on, whether I get along with them, what they are interested in and what their plans and goals are?

"Disclaimer": I'm not in STEM

I did my dissertation remotely, and to keep up with people, make sure my network was strong, push through the things that sometimes just need you there with a piece of paper and keep everyone happy, I was in the car an awful lot.

It's not a tactic I'd recommend to anyone unless their circumstances were strongly dictating it, and there's definitely stuff lost in the process. And my department wasn't particularly "collegial" - for the places that are, it'll be worse.

I did my Master's work part-time and semi-remotely - I lived and worked in the same geographical area as the University, and it was hard.

I would recommend against it in exceptionally strenuous terms for the PhD. Full-time on-site PhD or don't bother IMO. It just decreases the educational intensity and quality so much not to do that.

Yes, there is a trade-off. But the choice should exist, which is not the case so far, unfortunately.

What about it exactly is challenging? How is doing computer science research (probably what many people here would do; I don't know what your research field is) with a team of remote individuals harder than doing software development with a team of remote individuals?

remote software development is very challenging, and when it works well, it's due to super-clear specs and project management. by its very nature, research isn't as amenable to crystal-clear specs.

I haven't worked in industry so far. So, cannot comment on that.

In academia however you can possibly build connections with your peers (other labs) and lab colleagues through your work. One can attend on-campus seminars and exchange ideas with potential collaborators. Being remote voids these options. Opportunities to work on interesting projects (other than research. For instance, contributing to student organizations) will be lost as well.

Edit: Regarding my research area. On most occasions, students will have to register for classes (not all classes are offered online). Once you complete classes, it is potentially viable to work remotely. However, it again depends on the funding scenario. How a student is supported etc. If I'm a Teaching Assistant, I'm expected to be on campus and available to students. Being a Research Assistant gives the flexibility to work remotely (if advisor agrees however not all universities might allow that).

I essentially did a remote PhD for the last three years of my doctorate.

While it was possible, and I'd defend my decision to do it, it's a tough road, and not one I'd start out on if I could avoid it.

First two years in the US PhD is like a master's degree studies - regular lectures/courses. That is way US PhD program is usually 5 years long and you apply with Bachelor's degree.

In Europe, mostly, you're required to have a master's degree to apply for a PhD.

That's not particularly generalizable to U.S. programs - a good 3/4ths of the PhD programs I applied for expected a Master's before hand.

I'm also not sure how its relevant to what I said?

I see. I meant that hardest part can be that initial 2 years of US-esque 5-year PhD. Overall, in US, PhD is student - teacher relationship, while in Europe, mostly, it is employee - employer relationship.

Application looks too involved for Brown. When applying to 10 schools I wouldn't do this one because it has no overlap with my other application. I can reuse my letters of recommendation, I can't reuse this application.

I graduated quite some time ago. Getting letters of recommendation from professors would be impossible for me. I doubt I'd have one professor that knew me when I graduated let alone now years later. And yet, fulfilling these requirements would be very easy for me. I think it's very nice to have multiple routes in. Even if very few candidates avail themselves of this latter route, they might be interesting candidates worth strong consideration.

I'm in the same boat applying to a Master Program. It doesn't help that i spent the better part of my 20's as a business owner with no boss that can write a letter for me.

I think this would be useful for three groups of people:

(1) People who were planning to go into industry and already have, say, an offer letter from Google, or just a strong open source profile, but now aren't sure if they want to go into industry or academia. Being able to apply to Brown with their offer letter could reduce a lot of friction in their consideration of the academic route.

(2) People who've already tried to get into a top Ph.D. program the traditional route and failed. If this is your dream, and you don't have the grades or publications to get in otherwise, but you do have top programming chops, this could be a way for you to still get what you want.

(3) People who are already in industry and want to go into academia. They may have lost their academic connections at this point in time. Even if they had great grades, they may struggle to reconnect with old teachers to write the requisite stellar recommendation letters.

I wonder how would this work in fields other than CS. I'm in physics, and I got my D.Phil, but has been thinking seriously about the state of academic research, including the admission process & finding good student/lab fit.

Looking at the examples in the original post, I think the "demonstrate scholarship" section could directly apply for physics as well. Offer letters, github, are more more marginally useful...

Tough one. I'd love to hear how this works out, he sounds like a very thoughtful professor.

That sounds right up my alley! I'm currently tracking the geolocation of computer science professors by tracking the gaze of the NSA's agents who are watching them through the myriad of security cameras in our country.

My academic resume isn't going to get me noticed for a PhD program but it sounds like someone is finally looking at ability (I had great scores on the Computer Science GRE in 1990, but they apparently only last five years and that test has been discontinued)!

The GRE Computer Science exam was pretty heavy on theory and systems. This made sense for 1990, when CS departments were heavier on theory and systems. It has grown less relevant as the field of CS grew to encompass additional areas.

Thus, CS programs stopped requiring it, so fewer people took the exam, so the ETS decided to drop it.

I understand the rationale for dropping the exam, but it's the last academic endeavor on my resume. I've got a ton of work experience since then and many of my projects would be worthy of today's buzzwords. My lament is that only forward-thinking advising professors would see the value in what I have done.

Note that no professors have actually been tracked or harmed.

I'm almost certain that a Ph.D. in the UK takes 3-4 years max. Why does it take so much longer in USA?

You still need to do your masters, "along the way". If I remember correctly, in England, first you do you masters, and then start a PhD. 2 masters + 4-5 phd = 6.

Most European countries require a masters degree before starting, but the PhD is officially supposed to take 3 years (but usually takes 3-4+)

Submit your GitHub portfolio and join the Ivy League university that also develops Racket? Awesome!

This got me to consider applying to brown whenever I'm looking at grad schools.

even better: don't apply at all. simply work/study/design/research/build whatever you wanted to do anyway. then share/publish/collaborate as you desire. or not. rinse, repeat. be a genius. or a crank. or an entrepreneur. or some mix. just do things and get it out there. see what sticks. not all smart people are in academia and not all people in academia are very smart. and you can publish, collaborate, ship and show accomplishments without/outside a university paradigm.

Please don't underestimate how much it helps to have an adviser who has been through the ropes, or how awesome it is to have peers that are brilliant and working on similar stuff. I'm a self-taught programmer, so I know exactly what it's like to wing it, and I loved the PhD experience. The anti-PhD attitude on here is so weird, considering YC seems to be largely the same thing on a smaller timescale (weed out everyone but the very best, have them make things, give them connections, ...)


personally I think HN has both a pro-PhD, pro-academic, and also an anti-PhD or anti-academic segment. Sometimes within the same people! I certainly do.

I think it has both strengths and weaknesses. I think to the extent that universities and PhD track stuff serves an unnecessary and embarrassing anachronistic echo from the Middle Ages then that is a bad thing. But to the extent it helps teach disciplined thinking, cultivate a new generation of teachers and researchers, and helps to cull out noise or invalid/false/dumb contributions to the advancement of an art, it's a very good thing. The weed-out effect is good. I think some antipathy comes from the folks who perceive there to be a lot of non-weeds outside academia, and weeds within it. All discipline and curation is good. But it's not the unique providence of universities or professorship tracks. A lot of programmers have a libertarian bent too and dislike profit-maximizing monopolies or priesthoods, and universities/academia has at least a little of that aspect to them.

Classifieds: "Master is looking for more cheap power with experience from Google to pay minimum wage". Good luck, professor Huang.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact