I agree, except that (a) this should lead you to TA for longer in order to spend more time doing classes, but that is exactly the opposite of what grad students are told to do ("get to research as soon as possible") and (b) you should push the bounds of how much non-directly-thesis-related learning you can do while on your advisor's grant. You're a grad student, not a research technician, and only bad advisors will require you to do nothing but work directly toward immediate publications.
> only bad advisors will require you to do nothing
> but work directly toward immediate publications
I disagree with this point. If you have an advisor encouraging you to keep your eye on the ball getting out frequent publications, you should be grateful. You certainly will be when you get your Ph.D. and have > 3 publications under your belt for your job search instead of 1 and "several in submission."
I apologize. I must be presenting my ideas wrong, because I really don't think we disagree.
It is definitely the responsibility of a good advisor to make sure that their student leaves grad school well-prepared and with a competitive CV. An advisor who doesn't push his student to work hard and focus is surely doing a disservice. I just don't think that means the student shouldn't have a certain degree of freedom (maybe...10% of his time?) to pursue less immediately publishable knowledge.