In 1946, as the atomic age dawned, the United States proposed something unprecedented: placing the world's uranium and thorium under international control. The plan, championed by Bernard Baruch at the United Nations, would establish a global authority with strict inspections and the power to police nuclear development. Remarkably, according to a recent LessWrong analysis, the US—holder of the atomic monopoly—offered to surrender its weapons in exchange for universal compliance. It was a genuine attempt at technology governance. The Soviet Union refused. They rejected the intrusive inspections, saw the restrictions as a plot to keep them weak, and doubted American promises of disarmament. By 1949, when diplomatic efforts finally collapsed, the USSR had detonated its own bomb, forever closing the window for coordinated control. Humanity settled instead for the weaker International Atomic Energy Agency and its patchwork safeguards. The parallel to today's AI governance debates is stark: mutual distrust among great powers makes binding agreements nearly impossible—even when existential stakes are at play.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton