A thin semiconductor layer at low temperatures behaves like a pot of honey: electrons flow not as individual particles but as a thick liquid. If you run an alternating current through such a layer, the resistance responds with a smooth curve. Add a direct current, and a dip appears in that curve, deepening with increasing current.
This dip is not a new mystery but a consequence of heating. The direct current heats the electron liquid like a fire under a pot. The hotter the liquid, the lower its viscosity, and the resistance drops more strongly. Measurements showed that the viscous part of the resistance is inversely proportional to the square of the electron temperature. The same effect as in Gurzhi's experiments years ago, but now it was induced by current rather than by heating the sample from outside.
The practical significance is that the current itself controls the properties of the electron liquid. By changing the strength of the direct current, one can smoothly tune the resistance of a device without touching the thermostat. This is a step toward electronics where the charge flow behaves like a controllable liquid rather than a swarm of independent particles.
At liquid helium temperatures, electrons in gallium arsenide can flow as a liquid with viscosity exceeding that of water by hundreds of times.