In this talk we study the problem of Bayesian linear regression, where the coefficients have an iid prior.