The second law of thermodynamics just follows from statistics and a large number of interacting particles. It's not actually inviolable as the article says. There are whole books written about far-from-equilibrium thermodynamics, and the fluctuation theorem quantifies the probability of entropy increasing https://en.wikipedia.org/wiki/Fluctuation_theorem

Right, the second law simply is a result of selecting a result from a probability distribution that increases entropy, but drawing from tails of the distribution aren't impossible, only highly improbable.

Decreasing entropy locally is usually called a fridge, and I am currently betting my dinner that it is not bullshit. (I don't have the slightest idea what the article is actually talking about.)

Haven't read the paper, but it sounds like they just rediscovered the microcannonical ensemble (but for quantum mechanics). Probably works, probably hard to do anything interesting with it in the near to medium term.