He is at a good public R1, where he is presently teaching a graduate-level course. Alas, he has found himself disappointed by the quality of graduate students and fears he will never get any good ones. Luckily, he is probably wrong, but won't be able to see that for the next few years.
After having gone through the tenure track and a few additional years, I feel that one of the things that has most profoundly changed with respect to the early days of assistant professordom is my ability to work effectively with graduate students -- not the imaginary perfect ones, but the students you actually have around.
Sure, there is the initial shock -- not every potential graduate student is brilliant. Or competent. Or motivated. Actually, they may all seem hopeless at first. "Students here suck," you think, "I would be so much better off if I could only get my hands on one of those super smart kids at Top Notch Places. Sky would be the limit, with my brilliant ideas and their boundless potential!"
Several years down the road and many laments with colleagues later, you realize that everyone everywhere complains about the quality of students. But here's the thing: we weren't perfect graduate students either, it's all revisionist history; I bet most of us professors would not actually want our own selves as graduate students.
When you are on the tenure track, your own career is tied to your students. You are understandably very antsy to get things moving and are less able and less willing to ask yourself if maybe this student needs a different type of project or a different type of mentoring or more time or something else. I don't think this is selfish. You need to put on your own oxygen mask first before helping anyone else. Tenure track is a brutal, stressful time, and if you fail you will likely not get another chance again. For a student, working with a brand spanking new professor can be a trying time.
As a new professor, you don't really know what you are doing. You have no idea what you really need in a student (even though you totally think you have it figured out) and you don't know what it takes to advise someone. Mostly, you only know what you want done and you want someone who will magically understand what you want done and exactly how you want it done and just do it the exact same way that you would, preferably without asking you for help because you are unavailable on account of being crushed by teaching, grant writing, and service.
These days, I know what I am looking for in a student in terms of background, motivation, attitude, and aptitude. There are schools from which I like to recruit, as I've had excellent students from there in the past. I have enough funding and flexibility that I can offer a student several projects to choose from. Students usually start with a smaller project, where they take on an extension of someone else's work, the equivalent of one paper's worth, just to cut their teeth. By the end of this process, the student has a good first-author paper and we both have a better understanding of how the student likes to work, what their strengths and weakness are, and -- most importantly -- what gets the student really fired up and what types of projects they would really like to tackle. Here is an example: I recently had a student who spun wheels for a year on a project with a somewhat abstract, fairly theoretical focus; seeing that things were not working out, I restructured the project a bit to shift the focus onto a less lofty, very concrete and very computationally demanding problem. The student -- who had always been smart, hard working and conscientious -- took off in a particularly spectacular way and has made great strides on the revised project, as it turns out he is a very skilled programmer. The story has a double happy ending, as I soon thereafter recruited a student who loves the more theoretical work, so the original project is progressing well, too.
With the right project and proper guidance, I think almost any student who has the right background and motivation can do a very good job. It is true that sometimes finding the right project does mean changing advisors, but often it works out well with a little tweaking of the focus within the original group, or with a change in advising style: some students don't need the advisor much, some start out needing more time but gradually become more independent. As an advisor, you learn all these things through experience and you get better and more patient down the road.
Finally, I believe there are excellent students at every R1 university. It's just that they are fewer and farther between at lower ranked places, but they are definitely there. You may need to go personally through every applicant file, but there are many smart people all over the world. All it takes is one or two good students, with whom you can work productively, to make all the difference in your budding research program. And working with students definitely gets smoother and more enjoyable as you gain more advising experience.
Best of luck to all our junior colleagues in recruiting their first PhD students!