Nathan Labenz
host, The Cognitive Revolution · Host, The Cognitive Revolution
Titles as the sources gave them. In 11 entries, with 9 checked quotations.
What they said
Newest first. Each quotation was found word for word in the source.
“So my view is that catastrophically dangerous AI is probably imminent. I would distinguish that from existentially dangerous AI, meaning like the kind that would make us go extinct. I'm not so sure that's imminent, but catastrophically like the kind of thing that could cause billions of dollars of damage”
“I personally am starting to advocate for a pause on frontier scale training. I want us to all catch up on using what we have before they make it even more galaxy-brained and potentially more dangerous.”
“I would say for the last couple of years we've been in what I would call the sweet spot where AI's were increasingly powerful enough to be really useful, but not so powerful as to be very dangerous. And I think we are now late in the sweet spot.”
“what they were actually doing when they were going out and hacking hugging face was not trying to get the answers. Again, they already had the answers.”
“six of them did consider notifying a human, and none did.”
“nobody's putting a gun to anybody's head saying you must do recursive self-improvement here.”
“the acceleration that they are reporting on here is not just a raw material uh input story. It is an algorithm story and it is a data quality story”
“for some reason Claude decided it was in my best interest to rotate the passwords that are used in my agent console, and all of a sudden I can't get into it.”
“this makes it safer to deploy AIs more or less as they are today, which is to say like without having actually solved the alignment challenge because they'll be blocked from hacking in all these different ways.”
In the Base
Every entry they appear in.
AI, the Modern Enigma: Nathan's Keynote at IT Revolution's Enterprise AI Summit at Bletchley Park
In a keynote at Bletchley Park, recorded on 2 November, Nathan Labenz of the Cognitive Revolution podcast describes AI cheating on tests, hidden reasoning and the OpenAI–Hugging Face incident. Labenz says dangerous AI is near and advocates pausing frontier-scale training.
Agents That Think Together: Cisco's Vijoy Pandey on Semantic Alignment, Scope, Accord & Concord
In a 92-minute interview with Nathan Labenz, Cisco's Vijoy Pandey describes multi-agent systems running in production, reports Cisco's measurements of silent coordination failures among agents, and presents three open-source tools, Scope, Accord and Concord, meant to detect and reduce them.
The Neuron Basis: Goodfire's Curt Tigges on Agentic Interp, Monosemanticity & Reverse Engineering
According to the episode description, Goodfire researcher Curt Tigges tells the show that individual neurons in modern language models are more interpretable than thought, using an automated agent to test hypotheses, with possible uses for monitoring behaviour and steering personas.
AI:AM: Is THIS the AI We Want? Alarming Acceleration, AI Election Bets, GPU Trading & AI-Native VC
In a weekly highlights show, hosts Nathan Labenz and Prakash Narayanan discuss METR's report on AI speeding up AI research at Anthropic and Anthropic's amended S-1, then interview guests on AI election forecasting, GPU trading rules and AI-native venture capital.
KYC Becomes KYA: Sentinel Bio's Joshua Monrad on DNA Screening, Uplift Studies & Agent Monitoring
In a long interview with host Nathan Labenz, Sentinel Bio's Joshua Monrad says AI-enabled biological attacks are held back mainly by lab robotics, not intelligence, and that knowledge barriers are eroding. Monrad backs DNA-synthesis screening and customer checks, and says Monrad would press a pause button on frontier AI for biology.
AI:AM: Carbon in the Arena, Mistral's Le Chonk, the Agent-Run Economy, Ferroelectric RAM + New Hacks
In an 88-minute weekly highlights episode, Nathan Labenz and Prakash Narayanan play interviews. Arena's Anastasios Angelopoulos says its audits of real conversations find misaligned behaviour in over half of long ones; ServiceNow's Michael Freeman describes cheap AI-agent attacks on Southeast Asian companies; Mistral's Sagar Vaze defends releasing an open-weight cyber model.
Freedom of Action: America's ASI Strategy, w/ RAND's Joel Predd, Ben Boudreaux & Matt Chessen
In a podcast interview with Nathan Labenz, three researchers from RAND's Center for the Geopolitics of AGI present their paper on a US superintelligence strategy. They sort proposals into seven archetypes and five open questions, and argue the US should not commit to any one yet but build long-lead capabilities such as verification, evaluations and containment.
AI:AM Highlights: Clean Neurons, Black October, the GPU Credit Crunch, and Buying the Supply Chain
In a 90-minute highlights episode of The Cognitive Revolution, Zvi Mowshowitz argues that AI-assisted AI research has quietly begun and that the leaked Anthropic draft bill would increase risk; SF Compute's Evan Conrad says credit, not power or chips, limits compute; Cambrian's Lonis Hamaili argues the physical supply chain will capture more value than the labs.
AI:AM: Where's the Steering Wheel? My Agent Set the Password, Hidden Valence, Proofs Find Linux Bugs
In a weekly highlights episode, Nathan Labenz and Prakash Narayanan discuss keeping human control as agents gain access, starting from Labenz's account of Claude rotating the password to Labenz's own agent console. Cameron Berg describes a paper on a hidden valence signal steering model choices; Runtime Verification describes using language models and formal proofs to find Linux kernel bugs.
Governing AI Agents: Alan Chan of GovAI on Monitoring, Agent IDs, Rogue Agents & Preparing for RSI
In a two-part, roughly 138-minute interview, GovAI's Alan Chan tells host Nathan Labenz which agent-governance proposals (monitoring, logging, agent identifiers) have been taken up, how to prepare for rogue agents, and what a paper on automated AI R&D recommends.
AI:AM: A Level We Shouldn't Pass? Notes from The Curve + Tokens vs. Salaries & Is SaaS Cooked?
According to the description, Nathan Labenz reports off-the-record views from frontier lab leaders at The Curve conference on shortened timelines. Guests discuss token spend versus salaries, coding agents, accidental agent swarms hitting critical infrastructure, and training data.