Ask HN: Are there any other reasons to accept a PhD offer? | Hacker News Reader

Ask HN: Are there any other reasons to accept a PhD offer?

12 pointsToPhDorNotToPhd9 comments

Out of a passion for both brain research and computation, I've been working towards getting into a PhD program in computational/cognitive neuroscience (my work would focus on applying machine learning techniques to fMRI data). I've done this at a later stage than most people, having left a career in data analytics (I'm currently 31).

When I started out, I really thought that I'd pursue a career as a scientist/academic, but I've since read plenty of articles (one example among many: http://www.economist.com/node/17723223) about how abysmal the academic market is. I'm also aware that in cogNuro and neuroscience, even the best people (with very few exceptions) post-doc for 4-5 years.

So, I'm at a strange stage. 1) I've been accepted to a top 10 program (private midwestern university) where I'd be working with a well established and very active advisor. 2) Yet, I'm fairly certain that I no longer want to be an academic. Not only do I think my chances are slim of landing a decent job, even attempting to go the academic route would put me at 40ish before my first job.

My questions: -Is there any other reason for me to accept? -Would the PhD have any value outside of being intellectually very rewarding (as one professor friend of mine said "a 6 year adventure")? -Are there any applied routes that I could go? I do know that there are applications of this technology as "lie detection" and helping people with locked in syndrome, but it seems that these aren't nearly mature enough to produce jobs for PhDs. Of course, doing machine learning would probably be applicable in the data analytics world, but if I'm just going to go back to my old industry, I wonder if I'd be better off doing a masters in stats, cs, etc.

I ask these because I'm really trying to find out if there's anything that I'm missing. I just want to know if anyone is aware of some route that I don't see that would make this valuable.

If not, I may have to make the sad decision to turn down an offer to do some really cutting-edge research.

It seems like you're definitely going to get a good experience if you go with this lab. I'm currently enrolled in a grad program myself - this may bias my perspective.

The great part about where you're coming from is that you might already have a great sense about how the 'real world' works. I'm sure you've done some form of research before - however, the process of getting through a PhD program makes you work harder to understand what you think you know - and hones your ability to create new academic value.

Although a Master's in stats does prove useful, you're generally rehashing the same old techniques. You seem like you wanted more creativity in your job. I feel like value begets value - the academic value you create because you're curious can, if leveraged properly, become a company for you.

I'm sure you already knew all this - but regardless if you're thinking of doing a program, why not consider doing the PhD program? Just make sure to get the Master's on the way, and you haven't lost any time - you were going to do the Master's, anyways. In that time, you'll get a lot of research done and find more opportunities. Furthermore, you'll probably find it easier to be paid - most schools are better funded for PhD candidates than Master's candidates. Why not take a 'free' degree, then, if you're going to put in the time?

Quite honestly - the stuff you're doing is specialized, to some extent, but you can also link it to other imaging modalities. I presume you're correlating 'what am I thinking' to 'what do we see on fMRI?' - it's a classic multidimensional problem with a ton of potential solutions and a crazy number of permutations to handle. That's already some pretty useful stuff - you can take it, rinse, and reuse in another format - say, EEG, or applying new diagnostic fusion modalities to diagnosing various diseases automagically. Things appear different every time in different people - essentially a similar problem.

In addition, you've got a huge area for HCI (human computer interaction). Jobs already has a ton of people contemplating moving to the tablet for everyday usage. What if we end up going a step further to wearable computing - how will we have interfaces? Turn your stuff slick and sexy, and you've got a huge potential market.

You're landing a great research job in a great university. You're on Hacker News - wouldn't you consider starting your own company with all the crazy stuff you've discovered?

I have considered just going and seeing how it turns out (some people have given me the advice to go .... no better way to find out if I'll like it than to try).

I think you're right about the stuff coming up on the horizon. From what I know, some of the HCI stuff is problematic (I just talked to an EEG guy that was lamenting how problematic his machine learning/EEG research has been, given the noisiness of the data), but people are really starting to think about this stuff. There are also individuals using fMRI as a neurofeedback device (Chris DeCharms of Omneuron has started down that path: http://www.ted.com/talks/christopher_decharms_scans_the_brai...). Relatedly, some researchers have started to throw in classifiers such that one can focus on distributed brain states, as opposed to specific regions (which is what DeCharms does).

I read about this stuff and I'm fairly convinced that eventually the HCI, neurotraining, neurotech industries will take off, but for me it's a question of when. Will the growth of those industries happen in 5 years or 25 years? If it happens in 5 or 6 years, I'd be golden. If 25, then I'd be looking for a job when I come out.

As for the startup idea, I really have thought about that. The issue would be age again .... I'd be 37 with limited funds trying to crawl through the startup "Valley of Death." It wouldn't totally be off the table, but a lot would depend on the state of the industry.

So, all in all, I look at this PhD opportunity and think that there's definitely exciting work to be done, and the potential for really interesting jobs in the future. But it seems very high risk considering that I'd need to get the timing just right.

Are you married? If yes, DONT. If no, read on

Do you have lots of money stashed somewhere? If no, DONT. If yes, read on..

Are you fuckin' smart? Like 1 in 100 smart? If no, DONT. If yes, read on..

Do you have high levels of motivation. If no, DONT. If yes, read on..

Obviously this shit is easy to figure out when you are at 21; you have sometime to "throw away". Not at 31, AFAIK.

There must be companies out there who are doing the things you are interested in. Find em. If you are smart, you will figure out atleast 70% of what you would as PhD student; with the added bonus that you have a job.

You might want to read this too: http://www.bostonphoenix.com/alt1/archive/styles/97/03/06/de...

Answers:

Married: No

Cash: Not a ton (maybe a years worth)

Smarts: Maybe (on my best days)

Motivation: Yes (over the last 2 years at least....though over the last two years I had a clear goal: academia. Without that, I'm not certain that my doubts won't get in the way).

And agreed. If I were 21, this would probably be an easy decision. I'd just go, push through and see where it took me. At 31 though, I have other goals that I'm not sure if I'll be able to meet if I do the PhD. I'd probably have to put off getting married, having kids, etc. Not to mention, putting away cash will be almost impossible.

So, this seems risky. I could end up in my late 30s with some really interesting publications, but limited funds and no idea where I'd be finding a job.

As a PhD dropout from an east coast Ivy League school who just turned 24 this year, I would say a PhD isn't worth it if you've already decided not to follow the academia path for the rest of your life.

Besides, PhD programs in the states has been filled with Chinese and Indian smart kids these days. While you may be looking for your unique contribution to the frontier of the fearless intellectual exploration in Mother Nature from those programs, those smart Asian kids are just looking for a ticket into the states and they probably would sacrifice as much as it takes to outperform you.

So take a pause and pay attention to the facts. I know your advisor may be extremely warm and friendly and his recent publications may be extremely exciting. But the facts are that they expect you to follow the academia path while you don't want to, and that your grad school classmates are competing for an immigration stamp which you don't care.

The good news for me is that cognitive neuroscience doesn't face those pressures (yet). Most of the students doing cogNeuro and even computational neuroscience are from the US and Canada. In general, I got the impression that competition between students in these programs is much less important than collaboration.

That said, your point is very well taken: cutting edge research will still be very competitive. And, since the stuff I want to do comes very close to bio-medical engineering, I might face more of these sorts of competitive pressures than most people in my field.

PhD isn't worth it if you've already decided not to follow the academia path

I know a least a dozen people with PhD's in science, medicine or engineering and only one of them is following the academic path, and he gets regular offers of big bags of money to work for various commercial companies. The only PhD's I know who are 'stuck' in academia are people with PhD's in 'softer' subjects like architecture or social anthropology.

the facts are that they expect you to follow the academia path

Again this varies a lot from advisor to advisor. I know at least one guy (and know of several more) working at a company started by him, a classmate and their advisor during his PhD program. I also know other people who's advisor happily used their contacts in industry to place their students in great jobs.

These are the opportunities that I'm still trying to explore.

Even though my PhD would be in basic science, there are lots of potential applications. In fact, the program I'd be at has very, very strong links with the medical school. I'd probably be able to use some of these techniques on clinical populations and doing some translational research (i.e., stuff that has direct applications). For example, using machine learning with fMRI has some potential in mental health diagnostics.

The big question for me is if these things will take off in the next few years. Currently, I haven't seen any companies really growing in these sorts of areas (though I look at it and think the potential is immense).

the facts are that they expect you to follow the academia path

The bad news here is that while my potential advisor is really great (and probably okay with people not becoming professors) the people that haven't jumped into the post-doc route haven't had any of these sorts of outcomes. The few that haven't gone the academic route have become staff scientists or gone into government (though, those individuals graduated 5+ years ago, and my work would possibly be more technical since the lab has gone down the machine learning path).

In my opinion getting a PhD is a lifestyle choice.

Take a sheet of paper and write the three most important things you want to do. If a PhD won't help you achieve one of them, perhaps it is not the right route. If you are unsure find people doing what you want to do and just ask them if a PhD was valuable.

In the meantime, keep your options open until the last responsible moment. Remember there is no wrong answer.

Reply on news.ycombinator.com