"First, while some students are clearly putting their hearts and souls into their research, and spending the hours at the office or lab that are required, others are not. We have received some questions about how many hours a graduate student is expected to work. There is no easy answer, as what matters is your productivity, particularly in the form of good scientific papers. However, if you informally canvass the faculty (those people for whose jobs you came here to train), most will tell you that they worked 80-100 hours/week in graduate school. No one told us to work those hours, but we enjoyed what we were doing enough to want to do so. We were almost always at the office, including at night and on weekends. Nowadays, with the internet, it is fine to work from home sometimes, but you still miss out on learning from and forming collaborations with other graduate students when everyone does not work in the same place at the same time.
We realize that students with families will not have 80-100 hours/week to spend at work. Again, what matters most is productivity. Any faculty member or mentoring/thesis committee will be more than happy to work with any student to develop strategies to maximize productivity, even in those cases where the student is unable to devote more than 60 hours to their work per week."
Find the whole letter here, and a response here (where I first came across it).
I will not lie, in today's economy, I feel blessed to have a paid full time job doing something I like, most of the time. But at the same time, there are so many things wrong with the way tenured faculty are approaching their job of mentoring young scientists...and they wonder why they cannot attract brilliant minds into STEM fields. Simply put, no one wants to be in a job where their contributions aren't valued, and their commitment to the vague idea of "science" is constantly questioned.
What it comes down to is that your advisor will never tell you to work less; so if you're the person that's an unkempt slob eating crappy takeout, in lab at all hours, you are sending a message that you are "all about the science." And sadly, whether it's unconscious or not, advisors love that. There are not many professors in the world excited for their students to go on vacation, or to take a class unrelated to their thesis work, even if that's important for their future career goals.
So, besides reforming the way that publishing science works (which is another essay completely) - graduate schools need to come up with a better way to force career development as part of curriculum. That way, students won't be put in the middle of a conflict between their future career development and their current indentured servitude.
For example, you could make it compulsory for every graduate student to complete a certification process for one of several career paths; these could require classes, internships, extra teaching opportunities, shadowing, and outside mentors. These certifications could be in academia, industry, public policy, education, or entrepreneurship. I realize designing pilot programs from scratch isn't the easiest, but luckily, there are programs around the country with somewhat informal or unorganized versions of this, but nothing coherent across departments.
For example, MIT does have a Science, Technology and Policy Certification program, but it's something that requires a ton of work that would be completely unsanctioned by advisors for the most part. And that's the rub.
So, universities..make it comprehensive, make it compulsive, and take scattered infrastructure and make it unified. A nation of future grad school hopefuls would really appreciate it.