Episode 311: AI Could Outrun Our Control Faster Than We Think

AI doesn’t have to “turn evil” to wreck real life. It just has to be wrong, persuasive, and everywhere at once and the people closest to the tech are saying that risk is already here. We walk through a resignation and public warnings from major AI labs, then unpack what “alignment” really means in plain language: we’re growing systems from oceans of human content, we can’t fully explain what’s happening inside them, and the moment we can’t reliably turn them off or contain them, the stakes ch...