gjm comments on Bayesian Flame