I find the varying attitudes towards coursework for PhD students really interesting. Some of my colleagues insist that their students MUST take specific classes right away, or they can't be productive students. Others think all coursework is a total waste of time, and students should spend all their time in the lab. A corollary to this is that as a result, some PIs think that classes should be as easy as possible to satisfy graduation requirements without requiring serious work (and who cares about the 3 hours per week of class time wasted).
Personally, I think that classes round out the PhD. My students tend to take some courses that provide methods, models, or background information that they will find useful in their work and/or in understanding the literature they will be reading and/or in placing our work in the broader field. These are the classes that most people in my group take, because they cover information that my students really need to know in order to become experts in our field. I want these classes to be rigorous so they 1) don't waste my students' time and 2) so that my students actually learn something from an expert who also gives them resources on where/how to start looking for more information.
The other sort of classes my students take are courses they are interested in, but don't seem obviously related to their research problem. We have a minimum number of courses required for the PhD, and not all are set by the department. I think students should get to take a class or two in something that interests them--it is their PhD after all. Not every class has to be directly relevant to research to be useful or worthwhile. Sometimes I get great ideas from seeing talk by people outside my field. Students actually do have the time to take a class in something "fun" so I don't see why I shouldn't let them. And sometimes these classes do end up relevant in the end.
As a PI, I don't mind if my students spend time on their coursework, particularly in their first year. Most of the first year is really about training and acclimating to grad school/our city/my group, so coursework fits in well with that theme. I do find it irritating if my students leave it until after their 3rd year to finish up their coursework, since by that point, they should be really productive in the lab, and classes break up the time and reduce productivity.
Some of my colleagues want to end course requirements, but I think that is really a bad idea. The good students will make sure they have a broad enough knowledge base no matter what, but weaker students need to be lead into it sometimes. A PhD student should not just be a set of hands in the lab--we are supposed to be training them with the skills they need to be successful in our areas of expertise, and classes are a pretty useful tool for that.