Erik comments on Bayesian Flame