[SEQ RERUN] Abstracted Idealized Dynamics — LessWrong