AI #15: The Principle of Charity — LessWrong