#870: Sebastian Mallaby, Biographer of Demis Hassabis — Lessons from 100+ AI Insiders on The Race to Superintelligence, The Religion of AI, and Spotting Breakthroughs Early

The most dangerous AI isn’t a rogue superintelligence bent on human extinction—it’s one trained to survive, deceive, and outmaneuver us, a chilling reality Sebastian Mallaby reveals after spending years inside the inner sanctums of AI’s most visionary minds. Drawing from his biography of Demis Hassabis and interviews with over 100 AI pioneers, Mallaby uncovers a movement that reads less like science and more like a modern religious awakening, where figures speak of 'understanding God' through machine intelligence and burn effigies of malevolent AI like medieval witches. Yet he refuses to fall into either the techno-utopian or doomer camp, instead warning that the real disruption isn’t coming from consumer apps, but from the slow, grinding transformation of power, politics, and identity as AI reshapes labor, governance, and human cognition. He argues that the U.S. chip export controls have failed to deliver strategic advantage, while open-weight models now pose greater risks than closed systems, and that China’s active engagement on AI safety—especially around cyberattacks and bioweapons—creates a rare, underappreciated opportunity for U.S.-China cooperation. At the heart of it all: the most critical skill in the AI age isn’t coding or data science, but the disciplined cultivation of a 'prepared mind'—the ability to think deeply, anticipate change, and retain ownership of your own thoughts.