Hacker News new | ask | show | jobs
by throw_pm23 1297 days ago
I discredit this philosophy because E[..] cannot be accurately estimated. So it should be P(no harm|a) > 1-eps before anything.