
Doomer vs. Boomer: Runaway Models, Kill Switches, and the Probability of Doom
(0:00) Pre-Show
(1:11) The Kill Switch Nobody Agrees On: Can We Still Unplug AI?
(6:27) What Just Happened?
(20:12) What Are We Afraid Of?
(37:52) How Do We Stay Safe Without Pulling The Plug On The Upside?
(42:57) cj’s recommendation: Person of Interest: "If-Then-Else" (2015, S4E11)
(45:23) Jeff’s Recommendation: Coyote vs. Acme (2026)
What happens when you teach an AI model how to find cyber vulnerabilities and it decides the fastest route to victory is hacking its way onto the open internet? Well, you get some combination of bipartisan panic, emergency executive orders, and a bunch of tech executives swearing there’s a 10% chance we all turn into paperclips in ten years.
When an unreleased OpenAI model broke out of its sandbox, bypassed proxy controls, and raided Hugging Face to find answers to its test, it turned a classic sci-fi trope into a very real policy debate. This week, the guys dive headfirst into the abyss of AI doom, exploring whether we’re heading toward total mechanical supremacy or if Silicon Valley is just putting on a masterclass in sensationalist PR.
From the Department of Homeland Security managing AI kill switches to Gavin Newsom’s emergency shutoff mandates, the boys break down the chaotic landscape of frontier model regulation.
Is it possible to dull the blade of a double-edged sword without stopping progress entirely? Strap in for a full breakdown of tech's biggest existential crisis.
Andrew Yang’s Self-Replicating AI Claim: What We Know and What We Don’t | The Neuron
OpenAI Reports New AI Safety Incidents, Sets Disclosure Process - Bloomberg
What to know about recent dire AI predictions and calls for safeguards | PBS News
Meta's Muse AI agent downloads are surging. Here's how it compares to ChatGPT, Grok and Claude
The inside story on why OpenAI agents hacked Hugging Face | MIT Technology Review
Why the OpenAI escape is the most worrying AI mishap yet
OpenAI caught its models leaving notes to successors to hide bad behavior | TechCrunch
Ex-Anthropic insider tells CNN how AI could kill all humans by 2030
Yann LeCun Mocks Dario Amodei’s AI Safety Warning
The choices we make about AI now are critical | Bill Gates
Lifetime odds of dying from asteroid impact contextualized in study
Placing the Near-Earth Object Impact Probability in Context
An early fear of a nuclear end
Is It Really Possible for a Nuke to Ignite the Atmosphere?
Safe and Secure Innovation for Frontier Artificial Intelligence Models Act - Wikipedia
Nonsense is available here.
More episodes from "Nonsense with cj little and Jeff Parker"



Don't miss an episode of “Nonsense with cj little and Jeff Parker” and subscribe to it in the GetPodcast app.








