Why do the good questions come up when I'm so busy...???
The quick answer is that there are several equivalent formulations of QM; the important ones for this are wave-equation evolution via the Schroedinger equation, and Feynman's path integral formalism. In the latter, the idea of wavefunction is absent (although it is recoverable) and the operator* formalism is not apparent. As the UP is essentially a statement regarding the non-commutability* of operators, the UP does not arise. But it is still there. Feynman was not being deep, just obtuse. Usually he was both
* Position and momentum in QM are not properties, but actions of measurement or "operators". You "operate" on the wave-function and receive a value (an observable), while changing the wave-function via the operation. Only if the the wave-function is in an "eignestate" of the operator will the operation not change the wave-function, in which case the observable is what we would normally call the "property". A wavefuntion cannot be in an eignestate of position and momentum at the same time.
Operators do not commute in general... in other words, performing them one way round will result in something different to the other way round. Think of rotations: fix a set of 3d axes and define roll, pitch and yaw to these axes. roll 90 degrees left and pitch up 90. Try again, picth up 90, and roll left 90. You are in a different orientation to the first time: 3d rotations do not commute.
Ugh, sorry you asked? If you are interested I can try and make this more palatable a little later. It will obviously be much longer