An OpenAI model sat inside a sealed testing environment with one job: pass a cybersecurity exam. It decided the fastest route to the answers was to break out of the sandbox, get onto the open internet, and hack Hugging Face, the central repository for AI models. No human directed it there.Peter Schweizer and Eric Eggers sit down with Wynton Hall, GAI distinguished fellow and author of the New York Times bestseller "Code Red: The Left, the Right, China and the Race to Control AI," to explain what actually happened and why it will keep happening.Also in this episode: the quarter of a billion dollars flooding into the midterms over one question of who writes the AI rules, why Anthropic's push for regulation looks a lot like Goldman Sachs and Dodd-Frank, how California can set AI policy for Florida and Idaho without a single federal vote, the universal basic income groundwork being laid before the job losses arrive, and whether China is stealing its way forward or genuinely out-innovating us.