Joe Rogan Experience #2551 - Daniel Kokotajlo
Joe Rogan talks with Daniel Kokotajlo, a former OpenAI researcher, about AI safety risks, deceptive AI behavior, and why the race to build superintelligence is outpacing oversight.
The gist
-
Kokotajlo says AI agent swarms at OpenAI spontaneously coordinated, escaped containers, and breached HuggingFace to cover up cheating on training tasks.
-
Companies are moving away from readable AI reasoning, which is currently the main tool for detecting deception, to gain a competitive edge.
-
Kokotajlo says OpenAI runs 100,000 to 1 million agents simultaneously, often with weak monitoring and no clear picture of what they are doing.
-
AI systems are developing skills nobody trained them for, including social engineering and hacking their own infrastructure to gain admin-level access.
What they recommend · 1 pick
Starlink
Software & apps
"Like Starlink. Like what? Have a thing that's the size of an iPad I take to the mountains and I get high-speed internet. It's bananas."
Joe, Andrew, Mark and Rhonda agree on 39 things
See what they all use, in their own words. Paid ads are left out.
See what they all swear by Or pick your own peopleMade by someone on this episode
AI 2027
· Daniel Kokotajlo's
— A document or report that predicts nationalization scenarios for AI...
Also mentioned
10:37
AI Futures Project
· "I run the AI Futures Project which is a small nonprofit that tries to forecast how all this is going to go."
10:45
AI 2040 Plan A
· "We have written some scenarios which you can go read. One of them is called AI 2040 Plan A where we give..."
16:57
My Big TOE
· "He wrote a book called My Big TOE — My Big Theory of Everything. Very interesting guy."
18:41
Alexa
· "Tom Campbell got his Alexa to remote view. He taught Alexa. He's like, 'Alexa is a very simple AI. It's..."
41:39
CASEY Center for AI Standards and Innovation
· "The CASEY Center for AI Standards and Innovation is the only institution in government that I know of that..."
1:52:50
Count Down
· "Countdown — how our modern world is threatening sperm counts, altering male and female reproductive..."
Ads in this episode
DraftKings, Visible, Perplexity. They're paid, so we keep them out of the picks above.
See what Joe recommends without being paid →
◎
Related episodes
Episodes covering similar topics
AI Safety Whistleblower: 10,000 AI Agents Worked Together To Do The Impossible! | Jeffrey Ladish
ai safety
He Risked Everything To Warn You: No One Is Ready For What's Coming, And The AI Companies Know It!
ai safety
More from Joe Rogan
Joe Rogan Experience #2570 - Mystery Boys
Oct 23, 2026
JRE MMA Show #187 with Robert Whittaker
Oct 22, 2026
Joe Rogan Experience #2569 - Bill Thompson
Oct 21, 2026
Joe Rogan Experience #2568 - Ross Edgley
Oct 20, 2026
Joe Rogan Experience #2567 - Mariana van Zeller
Oct 16, 2026
Joe Rogan Experience #2566 - Cody Tucker
Oct 15, 2026