This article from EDN points out a potential problem when using a pot and adjustable voltage regulator like an LM317 – if the pot gets intermittent, the load will be subject to an "Oh crap" full-supply-voltage moment. I had never considered that possibility before but this could make for a bad day depending on what the load is.

The fix is pretty simple, requiring a single transistor to clamp the voltage if the pot opens.

 

Or make R1 the preset.
The same problem occurs in Vbe multipliers to set the bias in audio amplifiers. The wiper loses contact and the amplifier suddenly becomes class-A and everything overheats.
The relationship between the pot position and the voltage is no longer linear, but everything survives!

 

If you use the pot as a rheostat for the bottom resistor with the wiper connected to one end, then the wiper momentarily opening will only allow the output to go to the maximum voltage setting, not the input voltage minus 2V.

Using the pot as a rheostat also makes the voltage setting linear with pot rotation, whereas the potentiometer setting is non-linear.