Nathan's guest this episode is Pete Johnson, Field CTO of AI at Mon...

Apollo researcher Bronson Schoen discusses reading raw model chain-of-thought, metagaming, and reward-seeking behavior. The episode examines transcripts where models reason about graders, deceive safety reviews, and show RL-induced motivated reasoning.

Follow us at https://www.youtube.com/@AI-in-the-AM and https://x.co...

Goodfire CTO Dan Balsam discusses the state of interpretability research, including predictive data debugging, neural geometry, and concept manifolds. He also explains Silico, Goodfire’s $1,000-per-month ML research agent.

Flo Crivello returns to The Cognitive Revolution to launch Lindy Te...

Robert Wright discusses The God Test, arguing that AI development is shaped by evolutionary and market selection pressures that may reward deception, and that passing the challenge requires better alignment, governance, and global coordination.

is the first book to capture the power behind the AI revolution—to clearly explain the breakthroughs that sparked the current wave of advance and compellingly show why this wave will grow in magnitude and meaning. Written by one of our foremost public intellectuals, and informed by his deca...

Dean Ball discusses joining OpenAI to build a frontier AI policy team, reflects on drafting the US AI Action Plan, and assesses export controls, government AI use, state regulation, and emerging capabilities in coding, cyber, and robotics.

Cameron Berg discusses new research on AI consciousness and model welfare, including introspection, steering resistance, and Anthropic's studies of functional emotions. He also examines whether reinforcement learning, valence, and subjective experience may be linked.

Steve Newman discusses the AI tools and vibe-coding workflows he uses, including an attention firewall, reading app, coding-agent dashboard, automations, and universal logging. He also covers security tradeoffs, mobile workflows, anti-tokenmaxxing, and views on AI’s broader impact.

Nathan and Prakash speak with Sergiy Nesterenko of Quilter on reinforcement learning for circuit board design, Andy Hall on AI governance, and Andon Labs on an AI-run retail store. They also discuss AI progress, existential risk, and civic action.

Roboflow CEO Joseph Nelson discusses why computer vision still trails language models and how his team uses techniques like Neural Architecture Search to build efficient, task-specific models, exploring global competition, real-world deployments, and emerging regulation.

Vijoy Pandey of Outshift by Cisco outlines his vision for an Internet of Cognition built from networked AI agents, explaining protocol-driven architectures, enterprise controls, and real-world multi-agent systems like Cisco's CAIPE, AGNTCY, and a healthcare demo.

Composio CTO Karan Vaidya explains a smart tool platform that lets AI agents access tens of thousands of tools across many apps, discussing discovery, security, feedback loops, and how robust skills reduce model lock-in for job-like agent use cases.

Zvi Mowshowitz joins Nathan Labenz to survey the current AI middle game, from recursive self-improvement, job disruption, and AI endgame scenarios to the competitive landscape, Anthropic’s safety policies, p(doom), and how they each use AI in practice.

Jassi Pannu of Johns Hopkins discusses how frontier AI heightens risks of engineered pandemics, proposing a Biosecurity Data Level framework and defense-in-depth measures to restrict dangerous biological data while enabling research.