At first, when I read this, I thought they were talking about some kind of new interaction between matter and dark matter. But as George Jones mentioned, they're not: only gravitational interactions are considered. So the models are still considering dark matter to be a collisionless fluid.The abstract concludes with the sentence: "Our results suggest that, to leading order, dark matter is a cold, collisionless, fluid that can be kinematically 'heated up' and moved around."
My intuition is that a "cold, collisionless, fluid" that can be kinemtically "heated" and moved around.", is almost trivially a contradiction of terms.
As for cold, that means that the intergalactic medium of dark matter is cold. Within galaxies dark matter will always be warmer because it gains energy as it falls into gravitational potential wells. The temperature of dark matter in galaxy clusters, for instance, is likely to be similar to the temperature of the hot cluster gas, which emits radiation in the x-ray range. In fact, the dark matter is probably hotter if anything, because normal matter has lots of mechanisms to cool down that dark matter lacks.
This paper is based upon earlier simulations which demonstrate that when a burst of star formation occurs and changes gravitational potentials, dark matter in the area gets a tiny kick in energy. Each star formation burst imparts a small amount of energy to the dark matter, but repeated star formation bursts over an extended period of time have an incredibly significant impact on the entire shape of the density of dark matter near the center of the galaxy, making for an entirely different shape of the galaxy's dark matter profile (having a less dense core).
What the paper George Jones linked does is applies the earlier models to galaxy observations, and shows that galaxies which have evidence of a history of recent star formation are a fundamentally distinct population in terms of their mass profiles than are galaxies which do not have evidence of recent star formation. And while the error bars are pretty large, they broadly match the simulations which are based on cold, collisionless dark matter.