Large Random Matrices, Old and New

Large random matrices were born about a century ago in the work of the statistician Wishart to analyze big array of noisy datas. It was then shown in the fifties by Wigner and Dyson  to model well the Hamiltonian of certain heavy nuclei, hence finding its way in quantum physics. Even more surprisingly, Montgomery gave a numerical evidence that random matrices are intimately related to the zeroes of Riemann Zeta function in the seventies. But it is really only after the birth of MSRI that random matrix theory developed as a mature field. In this talk, after underlying a few other fields where random matrices play a central role, I will discuss the random matrix theory of Bernoulli  Matrices, old and new.