Ask HN: How does one find a research avenue? (retry)
Backstory: I completed my bachelors in '16 with some significant side/hobby projects (kernel/systems) and I'm now working as an engineer at one of the Major Ones. Although it pays well for now, I am slowly losing motivation and creativity and it is replacing my 'engineering for the fun of it' mindset with 'do what you're told to'. I'll need to adapt heavily to be able to thrive in this environment as a generic software engineer. I'm not learning, I'm not specializing, and I start to realize that I want to put in concentrated effort and work towards something for myself. Hence the idea of pursuing research. Or working for/on a startup.
My reason for going into academia to pursue a PhD would be, to be able to focus X years of independent research towards an idea I think can turn into a product later. But that depends on the idea/domain, the supervisors I work with and how success is measured (number of publications v/s real-world outcome, which may have a gap). Of course if I can't turn it into a product later, I have specialized and can come back to the industry.
Please correct me if my perspective is horribly wrong : ).
Thanks.
2 comments
[ 3.4 ms ] story [ 18.7 ms ] threadThe opportunities (scale, compute power, funding, how close you are to the edge of the state of the art) for research in industry are currently uniformly better than the opportunities for research in academia, and industry pays an order of magnitude or so more too.
If you followed ten "famous" CS profs five years ago, I'd guess that at least eight of them quit academia (despite having tenure) and now work at Google or Facebook.
So I think it's clear, at least to me, that if you want to do research it would be better to do it at a Major One. But I understand that it's hard to pull that off early in your career. Still, I would keep trying before taking the second-best option of academia.
Secondly, it's extremely unlikely that research in academia aimed at getting a PhD will be suited to being turned into a business. The purposes of "research for degree" and "research for product" are totally different. If you try to achieve one by doing the other you'll likely succeed at neither.
To answer the question you started with, the problems you work on in getting a degree are ones agreed between you and your supervisor during your preliminary work. For a PhD you need something sufficiently mainstream as not to be dismissed as a toy problem, obscure enough that no one else is likely to be working on it, hard enough that results are worth the degree, and easy enough that you can make progress in a severely limited time. As I say, finding a problem that suits those criteria and then also asking that it's research that can be turned into a product is all a bit much.
There are avenues to explore other than taking a PhD. Spare time projects, contributing to Open Source, reduce your hours at BigCo and start a consultancy on the side, go freelance.
A PhD is probably not the best option, but in the end all I can do is provide an opinion for you to take as grist to the mill.