ControlAI's Connor Leahy warns of superintelligence as adversary

controlai's connor leahy believes superintelligence is not a weapon but an adversary, posing significant risks that cannot be managed through alignment and containment alone. he advocates for stopping companies from developing superintelligence and supports legislation such as the sanders-casar 'ban superintelligence act' in the us and parallel legislation in the uk. leahy views frontier ai labs as political actors and warns of the dangers of ai self-improvement. this affects ai safety, policy, and the trillion-dollar data center buildouts.