Superintelligence 26: Science and technology strategy — LessWrong