At least it doesn't break conservation of energy. The second law of thermodynamics is already kind of suspect in that random fluctuations break it all the time on a microscopic level. It's closer to a statistical observation.

Turning heat into usable energy without temperature differential is still an extraordinary claim that demands extraordinary evidence, but I wouldn't dismiss it out of hand