> I mean, do we really think Google will be able to pull off a context-aware "hmmm" as a vocal trigger signifying desired interaction with a pop-up alert?

We have no idea what they have currently built, but from an article a few weeks ago it seems that they integrate with gestures, not audio cues. It's entirely reasponable that the head nodding you saw when he said "hmmm" was what caused the popup.