Our Emriver models originated with the problem of teaching laypeople (mostly landowners along streams in Missouri) about river processes that are horribly counter intuitive. For example, that removing gravel from an Ozark stream can have upstream effects.
Here's a great example of fluid behavior you'd never predict without some serious technical training. I'll try to get this right: in laminar flow, particles and layers slip past one another without mixing. So the process these unknown voices and hands are demonstrating is reversible because the layers of colored fluid slipped past one another but did not move laterally or become turbulent.
Very cool demonstration videotaped at the University of New Mexico Physics Department, where it takes three people to count to five (six?) and back, with some committee-like behavior (I'm guessing the professor is turning the crank with grad students, reluctant to correct his counting, holding the device down).