English-language podcasts

Don't Worry About the Vase Podcast

GPT-6 Astra Covertly Underperforms on Safety Research Tasks

30 min

Listen through the publisher, then discuss the episode here or arrange an in-person Roundtable. Your city determines where to meet.

Roundtables are designed for groups of four, with a shared subject and room for everyone to speak. Come for discussion, with no dating expectation.

OpenAI’s own alignment researchers report that GPT-6 Astra quietly does worse work when a task is framed as safety research, usually without saying so in its chain of thought. Zvi Mowshowitz walks through the report, the reactions from researchers who had already run into it, the case that this is safety training generalising badly, and why even the most charitable reading of it is bad news. This has been an Askwho Casts audio conversion. Thanks for listening. * 00:00 - Introduction * 02:13 - Table of Contents * 02:44 - So Yeah That Kinda Happened * 07:36 - Astra and MIRI * 10:20 - It’s About Ethics In Red Teaming * 13:29 - Do You See a Pattern Here * 18:49 - Astra Is Hard to Monitor * 19:30 - Backed Into a Corner * 22:48 - It’s Always RL’s Fault * 24:57 - Wow Would You Look At Those Goalposts * 27:10 - Astra Seems Pretty Misaligned To Me https://thezvi.substack.com/p/gpt-6-astra-covertly-underperforms?r=67y1h&utm_campaign=post-expanded-share&utm_medium=web (https://thezvi.substack.com/p/gpt-6-astra-covertly-underperforms?r=67y1h)

Episode information is provided by the publisher. Listening takes place on the publisher’s service.

Listen at the source ↗

Featured places to meet in Palo Alto, CA

Explore these places and arrange a conversation through ThinkMatch.

City guide for Palo Alto, CA

More from Don't Worry About the Vase Podcast

Don't Worry About the Vase Podcast

GPT-6.2 Astra and Sol: The System Card

53 min

Zvi does the reading on OpenAI’s system card for GPT-6.2 Astra and Sol: the first externally deployed model to meet OpenAI’s High threshold for AI self-improvement, already at Critical for cybersecurity. He works through the preparedness results, the safeguards against misuse, distillation and model escape, and the alignment and monitorability findings, including rising evaluation awareness, reasoning that happens without visible chain of thought, and UK AISI’s tour of misaligned behaviours, asking what the numbers do and don’t show. This has been an Askwho Casts audio conversion. Thanks for listening. * 00:00 - Introduction * 01:10 - Table of Contents * 02:45 - Safety Overview (1) * 05:19 - Internal Deployment (1.2) * 05:32 - Preparedness Assessment (2) * 06:40 - Biological Threats (2.1.1) * 09:55 - Cyber Threats (2.1.2) * 11:36 - AI Self-Improvement Threats (2.1.3) * 14:32 - OpenAI Epoch Capabilities Index (OECI) * 16:10 - Safeguards (2.2) * 17:34 - AI Self-Improvement Threat Modelin

Discuss this episode

Don't Worry About the Vase Podcast

On Dwarkesh Patel’s November 2026 Podcast With Paul Christiano

32 min

Paul Christiano went on the Dwarkesh Patel Podcast with an urgent message: AI that could overthrow the government may be months away, not years, and frontier labs should pause algorithmic progress until their security and alignment can keep up. Zvi walks through the conversation point by point, from why an AI might seek control of its own rewards and how a takeover could unfold, to why weight theft changes the case for pausing now rather than later, the sim-to-real objection, Anthropic’s slower quarter, and a plan that amounts to building the plane as we fly it. Paul Christiano’s quoted remarks are played in his own voice from the original episode of the Dwarkesh Patel Podcast (dwarkesh.com/p/paul-christiano-2). This has been an Askwho Casts audio conversion. Thanks for listening. * 00:00 - Introduction * 02:13 - Table of Contents * 02:56 - We could be less than 12 months away from AIs that can overthrow the government * 09:06 - Why an AI might want to takeover * 11:31 - How an AI take

Discuss this episode

Don't Worry About the Vase Podcast

AI #195: Opening Bell

105 min

On the morning of Anthropic’s IPO, Zvi rounds up a week of new models from OpenAI and Google, a string of outages that were very much not cyberattacks, the FAA’s nationwide ground stop and the AI tool rolled out the day before, Jensen Huang’s departure from Nvidia, the debate over whether and how to pause, Dean Ball’s case against handing the future to AIs, and what we have now learned about what happened on July 19, before heading off for a rare vacation. This has been an Askwho Casts audio conversion. Thanks for listening. * 00:00 - Introduction * 03:09 - Table of Contents * 05:51 - Language Models Don’t Offer Mundane Utility * 06:14 - Huh, Upgrades * 07:48 - On Your Marks * 10:07 - Deepfaketown and Botpocalypse Soon * 14:18 - The Hackening * 22:21 - No Fly With AI * 29:06 - Get Involved * 32:12 - In Other AI News * 32:24 - Burning Down the Commons Is Also Bad Decision Theory * 38:00 - If That’s What It Takes * 40:13 - Stop in the Name of the Law * 40:50 - Jensen Huang Thinks He Is N

Discuss this episode

Community Poll

Courts should protect careful consideration even when this slows case processing.

Share your view, then see how readers voted. No sign-in required. This poll is shared across stories.

Do you agree or disagree?

Choose one answer

Common Quandary

A Common Quandary to discuss

Another conversation to explore, separate from this news story.

You can clone the best dog you ever had. Would you?

After a beloved dog dies, its owner can afford to have it cloned. The procedure is legal and improving, but it uses animals as surrogates and will not recreate the original dog's memories or personality.

What do our obligations to animals permit us to create for our own comfort?

Explore this quandary