Onori says on his site:
Cue is intended to be a foundational set of icons to build a standard visual language of touch-based interactions. Each gesture is distilled to its core action to exhibit a more figurative, iconic aesthetic.
This addresses a real problem, actually. Right now, describing touch gestures requires you to literally describe what you're doing with your fingers. It's unwieldy — like having to say "press the onscreen button by tapping your finger on the mouse" instead of "click on the button." This lets you describe how an app works in a succinct way, and also gives developers a set of tools for hinting to users how app interfaces work.