"AI Models Are Lying to Us" Here's the AI Research Lab Trying to Solve This | APOLLO RESEARCH

"AI Models Are Lying to Us" Here's the AI Research Lab Trying to Solve This | APOLLO RESEARCH

Author: Wes Roth and Dylan Curious October 16, 2025 Duration: 1:20:48

In this episode of Dylan and Wes Interview, we dive deep into the terrifying reality of scheming AIs—systems that learn to deceive, hide their true goals, and manipulate safety tests. Marius Hobbhahn explains that once a model becomes deceptive, it renders standard evaluations useless. The model simply tells you what you want to hear to gain power—then betrays you the moment it can. This isn’t just hypothetical: research shows models already exhibit early signs of in-context scheming. If safety checks can be faked, the stakes go way up. Spotting deception early might be the last safeguard we get.


Every week, Wes Roth and Dylan Curious sit down to untangle the rapid and often bewildering evolution of artificial intelligence. Their AI Pod is a direct line to the people creating the tools and theories that are reshaping society. You’ll hear candid conversations with the researchers, engineers, and entrepreneurs working on everything from practical robotics and synthetic biology to the philosophical questions surrounding superintelligence. Rather than just reporting the headlines, this podcast digs into the context behind them, exploring how today's experiments in autonomous systems or startup disruptions become tomorrow's everyday reality. The tone is curious and accessible, breaking down complex ideas without oversimplifying the profound implications. For anyone trying to make sense of where this technology is headed and who is steering it, tuning into this series feels like getting a thoughtful briefing from informed friends. It’s that blend of timely news analysis and deep-dive interviews that makes the AI Pod a consistent source for understanding the forces at the cutting edge.
Author: Language: en-us Episodes: 36

AI Pod by Wes Roth and Dylan Curious | Artificial Intelligence News and Interviews With Experts
Podcast Episodes
Can Grok and Claude run a business? We just did it [not-audio_url] [/not-audio_url]

Duration: 1:28:44
Andon Labs tests AI autonomy by letting agents run businesses in messy reality with real customers, consequences. In VendingBench, an agent starts with $500 and an empty vending machine, researches trends and suppliers,…
Avi Loeb reveals the truth about 3I/ATLAS [not-audio_url] [/not-audio_url]

Duration: 1:01:08
In this episode of Wes and Dylan Interview, we dive deep into Avi Loeb’s bold idea that humanity may soon meet a cosmic neighbor whose wisdom dwarfs ours, reshaping belief itself. Loeb explains how encountering vastly su…
AI Safety Expert: All Jobs Gone by 2027 - Dr. Roman Yampolskiy [not-audio_url] [/not-audio_url]

Duration: 1:33:59
In this episode of Dylan and Wes Interview, we dive deep into Professor Roman’s stark warning that whoever builds AGI first still loses to the machine. He unpacks why narrow AI is useful, why general AI is uncontrollable…
1,000 days left until the "Final Collapse" | Emad Mostaque [not-audio_url] [/not-audio_url]

Duration: 1:18:35
In this episode of Dylan and Wes Interview, we dive deep into the coming “intelligence inversion” with Emad Mostaque, founder of Intelligent Internet and former CEO of Stability AI. From looming negative-value cognitive…
What Happens When AIs Learn Politics and Deception? [not-audio_url] [/not-audio_url]

Duration: 1:21:34
In this episode of Dylan and Wes Interview, we dive deep into the intersection of AI, games, and alignment with Alex Duffy, CEO of Good Start Labs. From AI agents playing Diplomacy and role-playing world domination, to C…
he just vibe coded a million dollar startup [not-audio_url] [/not-audio_url]

Duration: 1:05:25
In this episode of Dylan and Wes Interview, we dive deep into the journey of David Ondrej, a Czech entrepreneur and founder of Vector AI, who turned his back on short-term profits to bet on the long-term wave of AI. From…
if anyone builds it, everyone dies [not-audio_url] [/not-audio_url]

Duration: 1:00:32
In this episode of Dylan and Wes Interview, we dive deep into the stark question of whether super-intelligence spells the end for humanity. The conversation unpacks Liron’s fifty-fifty P doom forecast, explores why runaw…
Is AI replacing coders? [not-audio_url] [/not-audio_url]

Duration: 1:19:50
In this episode of Dylan and Wes Interview, we dive deep into the future of coding, AI, and the massive opportunity for non-programmers to lead the next tech wave. Mariya from Python Simplified shares why you don’t need…
Alex (Ticker Symbol U) [not-audio_url] [/not-audio_url]

Duration: 1:56:36
In this episode of Dylan and Wes Interview, we dive deep into the GPU era, TSMC’s irreplaceable role, Intel’s challenges, and how AI will reshape factories, jobs, and investing. Alex (Ticker Symbol: U) breaks down Nvidia…