[SEQ RERUN] Typicality and Asymmetrical Similarity — LessWrong