The study of the Gaussian limit of linear statistics of eigenvalues of random matrices and related processes, like determinantal processes, has been an important theme in random matrix theory. I will review some results starting with the strong Szegö limit theorem, and also discuss the possibility of non-Gaussian limits.