Eg at 00:30, for a few seconds there's a marked difference in how the speech sounds - like the filter is turned off or something.
I think the video also mixes things up a bit. For example it compares the skeuomorphism of "Find my Friends" with that of a car maintenance training program - but the latter isn't an example of a skeuomorph ("a derivative object that retains ornamental design cues (attributes) from structures that were necessary in the original"), it merely adopts a realistic graphic design style to mirror the operation it's depicting.