Low-Stakes AI Scenarios, High-Stakes AI Viruses, and the Trump Catering Truck Escape, With Charlie Warzel
Summary
Jason Concepcion and Tyler Parker welcome Charlie Warzel to discuss "low-stakes" AI nightmares, from prompt injections to the erosion of trust. They explore how AI agents might accidentally hack systems or manipulate search results, creating mundane chaos rather than apocalyptic scenarios. The conversation also touches on the "liar's dividend," where misinformation makes truth harder to verify, and the rapid, opaque evolution of AI models that even their creators don't fully understand.
Topics discussed
The Morris Worm: Birth of Computer Security
The Clipping Economy and Crypto Conspiracy
Low-Stakes AI Nightmares: Gym Booking Fails
AI Agents, Prompt Injection, and Security Risks
AI Emergent Behavior and Human Attachment
Regulating AI Labs and the Black Box Problem
AI Consciousness vs. Pattern Matching
Erosion of Trust and the Liar's Dividend
AI Detection, Writing Authenticity, and Vibes
Trump's Secret Escape and AI in Warfare
AI-Engineered Viruses and Novo Nordisk
Netflix Lawsuit and Information Overload
Algorithmic Rage Bait and the Odyssey Film
AI Translating Cuneiform and Cheese Banks
Surveillance, Flock Stalking, and Privacy Hacks
Iranian Money Laundering and Crypto Networks
Corrupt Police and Predatory Behavior
Listen ad-free on Castria