I am a physicist. I published a lot of nice papers in grad school, and sometimes I worked really long hours. Those two things are correlated.
Note that I did not say I always worked long hours. When you need to stay at work to babysit something (an experiment or observation), then do so. When you don't need to be at work, go do something else. You will be LESS effective working as much as you can, all the time. Good science requires creativity, not drudge work. The professors want results. They have no training in management and no interest in it either.
If you really want to see a culture change, don't get a PhD. Research funding pays for grad students, not staff scientists. Don't think that you'll be immune to this. Part of my job is to help the government manage research funding. Even when the economy is good, the government can't justify spending money on one professional when the same amount can get us 3-4 grad students (each working free overtime, right?). Think about what that means.
As long as grad students are happy working for peanuts, the system will continue. The accountants are merciless.