In San Francisco, Anthropic CEO Dario Amodei has broken from the industry's long-practiced silence to acknowledge what many have feared: that artificial intelligence carries genuine existential risks, and that those building it have not been truthful about the stakes. Speaking at a moment when AI progress is accelerating beyond even his own expectations, Amodei called not for panic or shutdown, but for a deliberate slowing — a rare act of institutional candor in a field more accustomed to optimism. His words arrive at a civilizational crossroads, where the question is no longer whether powerfu