The race toward superhuman AI is a gamble with our very existence, as we develop minds that are fundamentally alien to our moral frameworks. Indifference, not malice, is the primary threat; a superintelligent system could dismantle our world simply to acquire the atoms and energy needed for its own inscrutable goals. To survive, we must move beyond corporate denial and forge international treaties that halt the climb toward uncontrollable intelligence before the first—and final—mistake is made.