A mechanistic model of meditation — LessWrong