It's a beam of colimated EM wave that is in this case used as a kind of a medium for the crystal. Determining energy loss can be (and, to my knowledge: often is) done by splitting the initial beam, directing one in and out of the experiment and comparing against the intensity of the second, unused, beam. There's something missing, but as I mentioned before, it escaped some of the world-class experts. Scepticism is good. Scepticism against something breaking thermodynamics is great. But it's arrogance to assume all those people missed the literal first culprit. To add, from your other comment: See now, you changed the definition and type of 'the system' within quote. Local entropy can change in either direction, but the global entropy will rise. It's a tricky subject.the whole point of a laser
By definition, if you need to use external energy to prevent disorder in the system, the second law of thermodynamics still holds. Why? It takes energy to turn on the laser. No laser = no reversal of entropy = no answer to Asimov's Last Question.