The Dao of Bayes comments on Kurzweil’s predictions: good accuracy, poor self-calibration