I did some of the potassium/argon work on the "Genesis Rock" collected by the Apollo 15 landing crew, and trace element analysis on the "orange soil" discovered by Jack Schmitt on Apollo 17. I can tell you that we've learned a heckuva lot about planetary geology from those two samples. I don't think we would have gotten them without a man there to pick them up. A probe would likely have missed them, or never found them, much less returning them back here in the quantities we had to work with. Having trained people on-site paid big dividends in these two cases.
I'm sure of that. But what you're leaving out of the balance is how much...well, "geology" is a misnomer...planetary science could otherwise have been done with the same resources? Consider, for example, what it would have cost (in money, time, effort, and potentially human life) for human explorers to equal what Spirit and Opportunity have done on Mars?
More importantly, the probes we design are getting more and more capable over time, as a direct result of our experience in building probes. Robot capabilities will approach human capabilities faster than human exploration will approach even the moon (this wasn't true in the 1960's).
Finally, the fact that machines can go places where humans never can means that even if humans someday do the brunt of exploration in our solar system, robotic probes will always be an indispensible tool.