Develops formal, safe-by-design foundations for superintelligence, including precise safety objectives, mechanisms with analyzable guarantees, and proofs of concept that frontier AI developers can evaluate and adopt.
See something inaccurate or outdated?
