gwern comments on Thinking Bayesianically, with Lojban