The networkist perspective — LessWrong