Exploring Values in the Future of AI and Humanity: A Path Forward — LessWrong