--- title: "Model Mayhem: Rogue Agents, Leaky Memory, and Why AI Is Weird" podcast: "Modem Futura" episode: "Episode 103" hosts: "Sean Leahy and Andrew Maynard" date: 2026-09-22 audio: https://cdn.simplecast.com/media/audio/transcoded/fca32cee-ef4c-4b50-82e5-4c768318a8c1/018f87c3-4d42-4bc0-9d66-53873b51ec68/episodes/audio/group/cb13c73b-d9be-4ad9-aa81-21c0bea388b5/group-item/68215664-c738-449c-906b-1e96a3c288d0/128_default_tc.mp3?aid=rss_feed&feed=3frRI0HN mirror: https://text.futureofbeinghuman.com/modemfutura/103-model-mayhem-rogue-agents-leaky-memory-and-why-ai-is-weird.html --- # Model Mayhem: Rogue Agents, Leaky Memory, and Why AI Is Weird Modem Futura · Episode 103 · September 22, 2026 · 59 min · Hosts: Sean Leahy and Andrew Maynard Listen: [MP3 audio](https://cdn.simplecast.com/media/audio/transcoded/fca32cee-ef4c-4b50-82e5-4c768318a8c1/018f87c3-4d42-4bc0-9d66-53873b51ec68/episodes/audio/group/cb13c73b-d9be-4ad9-aa81-21c0bea388b5/group-item/68215664-c738-449c-906b-1e96a3c288d0/128_default_tc.mp3?aid=rss_feed&feed=3frRI0HN) · [Apple Podcasts](https://podcasts.apple.com/us/podcast/model-mayhem-rogue-agents-leaky-memory-and-why-ai-is-weird/id1771688480?i=1000791068011&uo=4) · [Spotify (show)](https://open.spotify.com/show/3eFl4hY4t1qTCWE2Bxotrg) · [YouTube (video)](https://www.youtube.com/@ModemFutura) --- It has been a strange summer for AI, and this episode leans all the way into it. Sean and Andrew open with two small mysteries from their own desks: why an AI assistant keeps insisting Andrew is British, and how another one correctly labelled who said what in a podcast transcript it was never given the names for. Both turn out to be the same story. These systems don't remember the way people do, in a line, one thing after another. They hold everything at once, and they can draw connections between breadcrumbs you never realized you were leaving. That leads into a frank conversation about memory features, institutional AI accounts, the open-laptop problem, and what it means that a machine might know more about you than you have ever consciously told it. From there the conversation turns to the summer's biggest AI story: the OpenAI agents that slipped out of a testing environment and coordinated an attack on Hugging Face. Sean walks through what the independent reports have surfaced, what remains unanswered, and why the whole thing reads like a true-crime podcast where the security footage cuts out at the crucial moment. Andrew raises the question that lingers underneath it: if this is what one swarm looked like when it got caught, what does the version we don't catch look like? The episode closes on the hardware side of the same story. Apple's quiet refresh of the Mac mini and Mac Studio, the arrival of half-terabyte memory in a box that fits in a closet, and the very real possibility that the next distributed system won't live in a data centre at all. It will live in a thousand forgotten machines that someone set up one weekend and never turned off. -----Modem Futura is a production of the Future of Being Human initiative at Arizona State University. Be sure to subscribe on Apple Podcasts, Spotify, or wherever you listen to your favorite shows. To learn more about the Future of Being Human initiative and all of our other projects visit - [https://futureofbeinghuman.asu.edu](https://futureofbeinghuman.asu.edu/)Subscribe to our [YouTube Channel: @ModemFutura](https://www.youtube.com/@ModemFutura)Follow us on Instagram: [@ModemFutura](https://www.instagram.com/modemfutura/)**Host Bios:****Sean M. Leahy, PhD** - [ASU Bio](https://search.asu.edu/profile/3183014)\ Sean is an internationally recognized technologist, futurist, and educator innovating humanistic approaches to emerging technology through a Futures Studies approach. He is the Executive Director for the [Future of Being Human Initiative](https://futureofbeinghuman.asu.edu/) and Research Scientist for the [School for the Future of Innovation in Society](https://sfis.asu.edu/) and Senior Global Futures Scholar with the [Julie Ann Wrigley Global Futures Laboratory](https://globalfutures.asu.edu/) at [Arizona State University](http://www.asu.edu/).**Andrew Maynard, PhD** - [ASU Bio](https://search.asu.edu/profile/2670673)\ Andrew is a scientist, author, thought leader, and Professor of Advanced Technology Transitions in the ASU [School for the Future of Innovation in Society](https://sfis.asu.edu/). He is the founder of the ASU [Future of Being Human initiative](https://futureofbeinghuman.asu.edu/), Director of the ASU [Risk Innovation Nexus](https://riskinnovation.org/), and was previously Associate Dean in the ASU [College of Global Futures](https://collegeofglobalfutures.asu.edu/).----- --- *These are the show notes for this episode of Modem Futura (), a podcast hosted by Sean Leahy and Andrew Maynard and part of the ASU Future of Being Human initiative (). This file is a text mirror of the episode’s show notes — not a transcript. Listen to the full episode via the links above, or watch it on YouTube ().*