Subvolts take Frontier AI risks and governance challenges are explored, questioning lab self-regulation. Aug 22, 2026 - Developers
- Understand the risks and governance challenges surrounding advanced AI systems and agentic behavior.
- Business
- Learn about the potential for AI misuse and the importance of independent evaluation for frontier models.
This episode features insights from researchers on critical AI safety and governance issues, relevant to current lab developments.
AI in the AM — Weekly Highlights: Relaunch Week (Aug 17–20, 2026)
Audio streams from Erik Torenberg, Nathan Labenz. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
AI in the AM's relaunch week episode discusses frontier AI oversight, the gap between internal and public systems, capability development, and infrastructure funding. Adam Gleave of FAR.AI warns of agent-orchestrated attacks and defenses pushing humans out of the loop. Alex Turner offers a governance and military perspective, questioning frontier labs' self-evaluation capabilities.
This episode talks about how advanced AI systems are being checked and who is in charge. It covers if AI is getting too powerful for people to control and if the companies building it can be trusted to check their own work.
This episode is for AI researchers, policymakers, and industry watchers concerned with AI safety and governance. It connects directly to current debates around frontier model risks and the challenges of independent evaluation, especially relevant as major labs like Google Gemini and OpenAI push new capabilities. The discussion on agentic systems and open-weight risks highlights immediate concerns for anyone tracking AI's rapid advancement and potential misuse.
- HostNathan Labenz
- GuestAdam GleaveFAR.AI
- GuestAlex TurnerGoogle DeepMind
- Episode Title: AI in the AM — Weekly Highlights: Relaunch Week (Aug 17–20, 2026)
- Show: The Cognitive Revolution
- Length: 154 minutes
- Guest Claim: Agent-orchestrated attacks and defenses are already forcing humans out of the loop.
- Guest Claim: Current monitoring has missed the failures it was meant to catch.
- Risk Type: Highly bio-capable open-weight releases pose a different kind of irreversible risk.
Relaunch week of AI in the AM brings together highlights from four live mornings and nine guests, centered on who checks frontier AI, how wide the gap is between lab-internal systems and public access, where capabilities are landing, and who pays for the physical infrastructure beneath them. Adam Gleave of FAR.AI argues that agent-orchestrated attacks and agentic defenses are already forcing humans out of the loop, while current monitoring has missed the failures it was meant to catch. The discussion weighs misuse versus misalignment through cyber incidents, deceptive agent behavior in evaluations, safeguards like pretraining filtering, and why highly bio-capable open-weight releases pose a different kind of irreversible risk. Alex Turner adds a governance and military-use perspective from his resignation account at Google DeepMind, sharpening the stakes around independent evaluation, enforceable standards, and whether frontier labs can be trusted to grade their own models.
For full show notes, links, and references, read the episode page:
https://www.cognitiverevolution.ai/ai-in-the-am-weekly-highlights-relaunch-week-aug-17-20-2026/
Sponsors:
Diffusion:
Diffusion helps organizations build custom AI software factories that scale business outcomes, not just outputs. Cognitive Revolution listeners get a 25% service credit on their first engagement at https://diffusion.io/tcr
Granola:
Granola is an AI-powered notepad that securely transcribes meetings and turns rough notes into clean, structured action items. Try it free at https://granola.ai/tcr
Deepgram Flux TTS:
Deepgram Flux TTS brings lifelike AI voices with real personalities that handle interruptions, pauses, and natural conversation. Try all the voices free through September 12 at https://deepgram.com/keep-talking
Claude:
Claude is the AI collaborator for problem solvers, helping with writing, coding, financial models, strategy, and more. Get started with Claude and explore Claude Pro at https://claude.ai/tcr
CHAPTERS:
(00:01) Checking frontier agents
(07:40) Misalignment and harms (Part 1)
(12:52) Sponsors: Diffusion | Granola
(15:49) Misalignment and harms (Part 2)
(19:20) Auditing fragile access (Part 1)
(27:38) Sponsors: Deepgram Flux TTS | Claude
(29:43) Auditing fragile access (Part 2)
(29:59) Military AI red lines
(44:58) Raising safety standards
(53:50) Internal capability gap
(01:05:06) Enterprise agent economics
(01:18:27) Cancer vaccines arrive
(01:23:48) Emergency AI tools
(01:28:59) Real time voice
(01:37:51) App layer squeeze
(01:44:52) Data center politics
(01:52:31) Accounting agent supervision
(02:06:25) Compute stack bottlenecks
(02:19:47) Public consent pricing
(02:28:12) Episode Outro
(02:31:38) Outro
PRODUCED BY:
- agent-orchestrated attacks
- Adam Gleave
- AI systems coordinating attacks
- agentic defenses
- Adam Gleave
- AI systems coordinating defenses
- open-weight releases
- AI models whose internal workings are made public
- AI models whose internal workings are made public
- bio-capable
- AI systems with abilities that mimic biological functions
- AI systems with abilities that mimic biological functions
1 h 27AI:AM: Where's the Steering Wheel? My Agent Set the Password, Hidden Valence, Proofs Find Linux Bugs
WithCameron Berg, Everett Hildenbrandt, Grigore Rosu, Phil Maher
2 h 192 h 021 h 28AI:AM: A Level We Shouldn't Pass? Notes from The Curve + Tokens vs. Salaries & Is SaaS Cooked?
WithThomas Sohmers, swyx, Evan Miyazono, Edward Hu
Same guests, other shows
1 h 441 h 27
59 min
45 minBuilding the Foundation for the Agentic AI Era
WithAngie Jones
Questions people ask
- What is the main concern about AI agents?
- Agent-orchestrated attacks and defenses are already forcing humans out of the loop, and current monitoring systems have missed failures.
- What is the risk with open-weight AI releases?
- Highly bio-capable open-weight releases pose a different kind of irreversible risk that is difficult to manage.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Erik Torenberg, Nathan Labenz.