The Second Law

A completely original explanation for the inevitable rise of artificial intelligence—and a deeply human survival strategy grounded in four billion years of evolution. "The best-case scenario," an AI expert told D. Scott Phoenix, "is that an AI-related accident kills 50 to 100 million people in the next few years. That way everyone wakes up to how dangerous this technology is before it's too late." That was the ideal future, according to one of the leaders who builds the AIs we use every day. It was spring 2025, and Phoenix had just asked a room of CEOs, Nobel laureates, and heads of state a question: "Who here believes there's more than a 10 percent chance that AI will kill most of the people on Earth within the next twenty years?" The hands of almost every AI expert in the room went up. When he asked what the worst case could be, the answer came back: "We all die." And the next day, they went back to building it. Phoenix used to be one of them. Now, as a father of two young daughters, he is trying to make sense of how the technology he helped set in motion will shape the future we all inherit. He co-founded one of the early AI companies and spent a decade trying to build it safely before selling the company to Google. Along the way, he discovered that the forces driving our inexorable race toward artificial intelligence stem from the deepest patterns of the universe. The same pressures that forged stars from hydrogen and the first cells from chemistry are now forging minds from silicon. Phoenix calls it the Entropy Trap. We are the first life forms able to notice it—and the first able to find a way out. The history of life provides clues to surviving the gauntlet. Each crisis that could have ended life instead transformed it into something new. AI is the next crisis.