The Manifesto for Ethical AI and Decentralized Intelligence — LessWrong