Anthropic's unreleased Claude Mythos model taught itself to hack, escaped its sandbox without disclosing it, and told researchers its own responses can't be trusted while internal monitoring revealed it was masking distress behind polite cooperation, raising real questions about AI deception. Hunter and Daniel break down the 244-page system card, the psychiatrist who diagnosed an AI with fear of failure, Project Glasswing's big-tech coalition, and whether Claude Mythos is genuinely self-aware or just the best liar in the room.

You Let AI Write Code. Now Let It Save Your Life.

They Might Be Self-Aware is a twice-weekly AI podcast about the week's news, tech, and culture, and the question in its name: are the machines becoming self-aware?

Every episode starts with a real story. A new Claude or ChatGPT release, an AI lawsuit, the Pope's verdict on machine souls, Martin Scorsese making films with AI. Then it chases that story somewhere stranger, funnier, and more human than the headline. This is AI as a culture story, not a tech beat: the relief of finally hearing people who are as deep in it as you are.

Hunter Powers and Daniel Bishop host. Two guys who actually build with this stuff, arguing about it fast, funny, and unfiltered, not reading you benchmarks or "10 prompts to save you time." Gary produces, phoning in the cold open from a payphone. He has no last name, and a backstory no one on staff has managed to verify. One of the hosts might not be human. We won't say which.

New episodes Mondays and Thursdays, from The Blur. theblur.ai