> Quickly put a mirror in the photon’s path right where the wave function superoscillates, keeping the mirror there for a short time. Varying things quickly in

> Quickly put a mirror in the photon’s path right where the wave function superoscillates, keeping the mirror there for a short time. Varying things quickly in space or in time requires a lot of energy. The energy could be coming from moving the mirror too quickly. For example, if you modeled the mirror's movement as being driven by a force then you might find the mirror's motion is damped by reflecting the photon; losing energy. Instantaneously switching the driving Hamiltonian can easily spread energy around, unless they commute. The Hamiltonians for "no mirror" (H1) and "yes mirror" (H2) won't commute. It sounds like they've arranged for the eigenstates of H1 to overlap with a huge range of eigenvalues of H2, and vice versa. So when you switch from H1 to H2 you end up in a superposition of all kinds of different energies. Evolve there for a bit so switching back won't destructively interfere you back down to exactly where you started, and voila. I think if they account for these kinds of "changing the Hamiltonian ain't free" effects, they'll find where the energy came from.