Implication of AI timelines on planning and solutions — LessWrong