No, I didn't hook it up that way.
Were you showing up an example of a incorrect connection or was I supposed to connect like that?
The diode is reversed in my setting.
What happens now is that, the pin on the MCU is simply 4.9V when the fan is idle, which is good. And, when the fan is running, the PWM gets sent to the PIC but its PWM signal has a DC offset of about 0.3V or so. So, its mean is read (by an oscilloscope) as 2.8V instead of 2.5V (which is what it should be).
I also cut the RPM signal from the fan in half so I could measure the current flowing in the line (without any components attached, straight from the fan to the PIC). The current measured is about 12mA, if that gives anyone any idea.