Second law of thermodynamics "broken".

I'm in two minds about this one. On the one hand, the 2nd law was only ever a statistical law - Entropy can decrease, it is just that it is vanishingly unlikely to happen. So unlikely, in fact, that in practice it never does. But there is no reason why a special set of circumstances, like the one set up by the researchers, shouldn't work in a different way.

On the other hand, Holy Shit!

Update July 24th: BBC coverage. The idea here is that at molecular scales and over extremely short periods of time, the 2nd law doesn't necessarily hold - this is the Fluctuation Theorem. The Australian team has now demonstrated this experimentally. There is no implication for large scale physics, but nanotech might have to take account of this effect.