In a classical paper P. R. Halmos shows that weak mixing is generic in the measure preserving transformations. Later, in his book, he gave a more streamlined proof of this fact based on a fundamental lemma due to V. A. Rohlin. For this reason the name of Rohlin has been attached to a variety of results, old and new, relating to the density of conjugacy classes in topological groups. I will review some of the new developments in this area and its connections to model theory.