I have seen several mentions of James-Stein estimator being almost an empirical Bayesian estimator, but this article made it more clearer. Thanks for sharing.