This podcast discusses the idea of predicting errors in short computer programs by enumerating all programs that produce the same output with added errors and using that enumeration to predict future errors.