Superintelligence 27: Pathways and enablers — LessWrong