x
Fermi Estimates and Probability Calibration — LessWrong