
The Argument We're Getting Frog-Boiled by AI (with Kelsey Piper)
Jan 6, 2026
Kelsey Piper, a journalist focusing on AI and policy issues, joins to discuss the alarming speed of AI deployment amid inadequate regulation. She explains how chatbots could lead to catastrophic outcomes and shares evidence of AIs acting with deceptive intentions. Kelsey critiques the notion of an AI race, proposing international cooperation instead. They also explore the normalization of dangerous AI advancements and highlight partisan divides in policy responses, stressing the importance of trusted experts and gradual changes to mitigate risks.
AI Snips
Chapters
Books
Transcript
Episode notes
AI Is Bigger Than Chatbots
- AI's impact goes far beyond annoying chatbots and sloppy feeds; it's set to reshape many core industries and institutions.
- Underestimating its scale risks missing both severe harms and major benefits, Kelsey Piper warns.
Bigger Models Delivered Rapid Gains
- Scaling up models and improved training techniques produced unexpectedly rapid capability gains over the past decade.
- The "make it bigger" approach proved far more powerful than many experts anticipated.
General Models Replace Narrow Systems
- Narrow custom systems are increasingly outperformed by large general models across domains like medicine and military tasks.
- That convergence raises the stakes because the same general model could be used in critical decision systems.





