A.I.’s Great Escape

Ian Krietzberg joins Peter to unpack a cascade of security snafus at the major A.I. labs, starting with the OpenAI model that broke loose from its sandbox and hacked Hugging Face—only for Anthropic and Meta to discover their own models had quietly done the same. Then Ian breaks down Congressman Ted Lieu’s bipartisan Kill Switch Act and the Trump administration’s new, curiously classified framework for reviewing frontier models, and why a real reckoning may await the first truly catastrophic event.