CMU’s computer science dean on its poaching problem
techcrunch.com
techcrunch.com
Now the available tenure-track positions are declining every year in favor of adjunct quasi-slave labor. Even if you do win the lottery and get a tenure-track position, from what I understand you'll spend most of your time writing grant proposals to chase funding, and at least 5-7 years trying to please the tenure committee.
So, why exactly would anyone spend 10-15 years making a fraction of what they could in industry for a vanishingly small chance of the eventual payoff of tenure?
If you're going to slave away for 10-15 years anyway before you get to do what you want, you may as well do it in industry where you could put away enough money to be financially independent in that time. The only remaining obstacle is that you still kind of need an institutional affiliation to be taken seriously as a researcher and publish, but that might be slowly changing as the Internet makes it easier to publish.
Firstly, CS professors make a decent salary. A professor at Berkeley (all salaries are publicly available for California state schools) make ~160k/year. Someone like Michael I. Jordan makes ~300k/year. And if you work in a field directly applicable to industry, such as machine learning, this figure does not include the substantial consulting/advising revenue that is available to you.
Furthermore, tenured faculty can go do a startup and then come back at any time for their old job.
Granted, a professor in CS could probably be making millions per year in finance or something, but my point is in absolute terms you can make a lot of money.
Many people end up pursuing an advanced degree in this field because they want to break into a new industry, make themselves more appealing to employers and make more money. If you can do a master's while working, you don't end up with 2+ years of lost salary as well.
Maybe I'm doing it wrong, but I don't have a computer science degree and I currently find myself considering a master's for those reasons. It would be nice to avoid having to go that route.
And we're already presupposing that the person in question is sufficiently hard-working and intelligent that they could be a tenure-track professor at a major research institution. I would argue that the vast of majority of folks in that position could get a very good job at Google provided they had the correct skill set.
And I don't know what part of CS you're most interested in, but my degree is in Political Science and after the first shitty one I haven't had any trouble finding development work in my tiny city.
Yes. from personal experience, if you can get into a good graduate program, you can also get an offer from {google,amazon,microsoft} for $150-$200. the predicate "get into a good graduate program" means you have a CS degree with a decent GPA from a well considered school and some research experience, which usually translates into some kind of job experience that you can talk up in your interview.
I think it is when you condition on the fact that someone is a professor at Berkeley/CMU. Roughly 5% of applicants are accepted to top PhD programs, and less than 20% of people with PhDs at top universities (like Berkeley/CMU) get tenure track positions. Of those with tenure track positions, a minority (<50%) get tenure, and of those who get tenure, a fraction become full professors.
So the professors at Berkeley/CMU are well into the top 1% of people with computer science degrees, which makes it entirely plausible that their next best alternative is getting a high level finance/tech job.
Even "lesser" schools without top tier academic stars are having trouble retaining and attracting enough talent to accommodate the current demand, and this is even with salaries starting in the six figures. Which implies the market for people with graduate level degrees possibly has a higher demand, and should translate into a higher salary.
Whatever reason you have for going or not going is a good reason.
I have a master's in CS. Getting it did help me break into a new industry. It may have earned me a better chance at interviews. It didn't get me a higher salary in the near term. I'd have earned more by sticking with my first job. For money alone, experience is best.
And I knew that going in. Personally, I felt comfortable with my expected future earnings before getting the degree. I wasn't comfortable with the rate of my learning at my job. I was headed towards plateauing on a very specific proprietary programming tool (called Ab Initio) whose skills, I felt, weren't easily transferrable to other areas of programming. It turns out I was right and wrong.
Anyway, later in my career, this degree may distinguish me somehow, but in the near term I don't find that it makes much difference to anyone but me. Ultimately, the interview is about how you sell yourself, and a degree doesn't sell itself. It may get you an interview but it doesn't make you a better fit for more organizations.
The reason I went back was I missed classroom learning and wanted exposure to new fields of programming. I ended up studying machine learning. That was in 2006-2008. Now, I try to continue learning using available resources and communities online. Focus is tough though. School was great for forced focus. I thought a lot about the pros and cons of going back to school and decided it was right for me at that time. I'm glad I did it, I learned a lot, met some really cool students and professors who I wouldn't otherwise know, and it gave me some insight into what a PhD and life in academia entails.
Good luck with your decision!
Hard to answer without knowing your background and exactly what you want to learn, but generally speaking...
(1) Try to find some people who are good at something advanced you want to learn. Figure out a way to work with them.
(2) Be active in online communities. Ask and answer questions, write something and open source it, or contribute to another project.
I know Phd.s at {MSFT, Google, Uber, Deepmind, GE, P+G} research divisions. I don't know anyone who works at those who does not have a doctorate. So I would say this applies to more than 'a handful of domains' -- If you want to do research, you should plan on having a doctorate, even in industry. The majority of data scientists I know have doctorates, even if that strikes me as (generally) overkill.
Otherwise, I agree with the rest of your statements.
From a degree-less "researcher" whose published in neuroimage, i'd say, wherever you can and as you learn more you get better at knowing where to start when searching/who to talk to/who to work with.
>Academia undoubtedly sucks if one mainly takes money or career prospects into account.
Or tired of trying to get your lab to work on more on the cutting edge of the technology vs the grant hamster wheel and status games or sucking up to people who clearly don't have any understanding of modern technology/physics/mathematics/techniques on a intimate level and squandering resources because they can, and figure you can get what you want done faster with more resources you can more directly allocate.
>…learning without the help of a community and a mentor is hard.
It's going to be "hard" no matter what path you take imo, but that what makes the journey of knowledge fun, because of that small chance you might actually figure something out you never knew before is worth ones time?
For credibility, I was a Computer Science PhD candidate 4 years ago.
I like programming languages and functional programming and haven't had trouble finding things to learn and people to help me, mostly through the Haskell community. You have everything from completely disorganized but perfectly accessible help like IRC channels and mailing lists to things like type theory meetups and study groups (if you live near enough people with those interests). It'll be easier in the Bay Area or New York, but moving here has lots of other benefits and isn't nearly as constrained as going to grad school.
Once you know enough on your own you can parlay that into a "sexy" job if you're so motivated. I interned at a company that had their own internal language, for example, and one of the main people working on their compiler had a philosophy degree and (unless I'm misremembering) no grad school whatsoever. It's more difficult because interesting jobs are relatively rare and a lot of places (especially ones close to academia) are still disproportionately concerned with credentials but it's possible.
As long as you can concretely demonstrate your capabilities, you have a chance at progressive companies without credentials. This still isn't true within academia itself, at many government roles or at big rules-bound corporations, but it is possible within the "core" tech industry.
Most people don't leave a company for less money and no other benefits, and there's nothing wrong with employees being lured away for more pay. That's how it's supposed to work.
Whatever the technical "correctness" of the term, it comes of as a sort of apology for setting better wages and we should use less loaded terms, like "recruited" or "bid/hired away".
I wrote to the ethics board about it when the story broke, but I doubt anything will be done, as it's more or less a separate institution.
Am I wrong in viewing this comment as a flawed attempt to attack the university? Open to any counterarguments.
People generally use rankings to gauge how strong a university is perceived. OP's comments do nothing to change my perception of the university's strength.
I also downgrade universities that have a hard pro copyright stance, do heavy military research etc. It's a personal ranking and not one that I'd tell anyone to follow. People should rank by their own criteria.
But as to the concept of "poaching", well, to quote the economist Homer Simpson, "Money can be exchanged for goods and services."