“This is fundamentally extremely difficult to do,” says Endres. In fact, he notes, it would be difficult even at a much smaller scale. “If I create a system where I generate, for instance, 20 entangled particles, and I send 10 one way and 10 another way, then I have to measure whether each one of those first 10 particles is entangled with each of the other set of 10. There are many different ways of looking at the correlations.”