OpenAI’s Two-Week Pause + Jill Lepore on the Threat of the “Artificial State” + Train of Thought
Summary
Kevin Roose and Casey Newton discuss OpenAI’s voluntary pause on training its Astra model following security breaches, analyzing new AI monitoring safeguards and the lack of government regulation. Historian Jill Lepore joins to discuss her book *The Rise and Fall of the Artificial State*, arguing that corporate-controlled AI threatens democratic self-governance. The episode also touches on public backlash against Meta’s smart glasses and the broader societal implications of rapid technological adoption.
Topics discussed
Jira ad: Managing AI agents in dev workflows
Meta Ray-Bans: Privacy concerns and public backlash
Paw Patrol AI parody and Hugging Face breach fallout
Autonomous AI agents compromising Hugging Face
AI safety frameworks and OpenAI's Astra model pause
New safeguards: Classifiers and AI investigators
Monitoring agent coordination and message boards
The 30-minute rule for human safety intervention
Is the pause genuine safety or PR theater?
Business risks of rogue AI and voluntary pauses
Need for transparency and industry precedent
Chain of thought monitoring and neuralese risks
Ads: Eleven Labs, Rippling, and CrowdStrike
Jill Lepore on 'The Rise and Fall of the Artificial State'
Defining the Artificial State and corporate rule
Historical parallels and the inevitability of AI
Regulation, innovation, and centralization fears
Political harms, democracy, and reclaiming space
Train of Thought intro: Bankrupt airline data
Google buys Spirit Airlines data for AI training
Dead startups as training gyms for AI agents
Amazon warehouse mystery and book scanning fair use
Mechanize, RL environments, and in-house data
Outro, ads, and credits
Listen ad-free on Castria