The Cognitive Costs to Doing Things — LessWrong