dsj comments on Empirical risk minimization is fundamentally confused