The AI lab watch
Tracking OpenAI, Anthropic, Google DeepMind + 2 more
- OpenAI — Frontier lab; track model, pricing and org changes as they are discussed.
- Anthropic — Frontier lab; track safety positioning, enterprise strategy and funding.
- Google DeepMind — Frontier lab inside Google; track research direction and product integration.
- Meta AI — Open-weights strategy and how rivals respond to it.
- Mistral — European frontier lab; track open-weights releases and enterprise deals.
OpenAI security breach and Anthropic revenue concerns dominate
~84 hrs of podcasts, listened to for you. This is the 2-minute version.
329 search results → 117 episodes scanned → 8 made it in · 7 analyzed in full
The OpenAI security breach has evolved from a singular incident into a broader crisis, with reports confirming that physical surveillance was used to target executive strategy sessions. Meanwhile, Anthropic’s financial performance is facing investor scrutiny after leaked Q3 data showed growth slowing to 46%, while Meta AI continues to push its open-weights strategy with the release of the Muse Wonder model series.
Oct 15 · 17 min · Top 10 news
OpenAI confirms boardroom surveillance by staff engineer
- An OpenAI engineer, Abdul Rahman Al-Fozan, was arrested for allegedly planting a secret recording device in a company conference room.
- The device was discovered during a routine security sweep, leading the company to work with the FBI to monitor the suspect.
- OpenAI stated there is no evidence that sensitive data, source code, or model weights were accessed or exfiltrated by the engineer.
Why this matters to you
Internal security breaches involving high-level access to leadership spaces pose significant risks to proprietary research and long-term strategic development at frontier labs.
OpenAI says there was evidence he planned to install more of these devices throughout OpenAI offices in San Francisco.”
— Emily Drushinsky
Listen Deep dive
OpenAI
4 episodes · 4 of 21 mentions
Revolutionizing Cybersecurity and AI-Driven Science with OpenAI President Greg BrockmanOpenAI shifts budget toward defensive cybersecurity infrastructure · No Priors: Artificial Intelligence | Technology | Startups · Listen · Watch
Greg Brockman announced a $1 billion commitment to cybersecurity, signaling a major strategic pivot toward defensive AI infrastructure and mass-market model integration.
OpenAI's Office Was BuggedMicrosoft leak reveals OpenAI's use of loop transformers · Tech Brew Ride Home · Listen
A Microsoft disclosure suggests OpenAI's GPT-6.1 Sol model utilizes loop transformers, an architecture that complicates the monitoring of model reasoning.
The Strip Mining of Mathematical Proofs - with Hugo Duminil-CopinMathematicians criticize OpenAI's automated proof release · The Numberphile Podcast · Listen
Academic researchers are pushing back against OpenAI's rapid release of mathematical proofs, arguing it undermines human intuition and the traditional academic pipeline.
Anthropic
2 episodes · 2 of 8 mentions
Oct 16 · 23 min · Top 25 technology
Anthropic Q3 growth misses investor expectations
- Anthropic's Q3 revenue reached $16.8 billion, representing a 46% quarter-over-quarter growth rate that fell below investor expectations.
- The company reported an 11% adjusted operating margin and 85% adjusted gross margin, excluding costs like stock-based compensation.
- Distillation services for competitors accounted for 5-7% of revenue, though Anthropic claims to have reduced these attacks by 90%.
Why this matters to you
The reliance on distillation revenue from competitors and slowing growth rates challenge the sustainability of Anthropic's current valuation and enterprise strategy.
“Anthropic's Q3 numbers dropped through leaks from meetings with investors in preparation for its IPO... sales only grew by 46% quarter over quarter.”
— Ed Zittron
Listen Deep dive
AI Chat: AI News & Artificial Intelligence
Oct 16 · 14 min · Top 100 technology
Anthropic formalizes third-party safety oversight with METR
- Anthropic has provided METR researchers with employee-level access to internal tools and the latest model versions.
- METR staff now hold Anthropic-issued laptops and office space to facilitate independent safety and risk assessments.
- Anthropic employees are encouraged to share internal information regarding alignment risks directly with the METR researchers.
Why this matters to you
This partnership signals a shift toward third-party oversight as a primary mechanism for building public trust and avoiding heavy-handed government regulation of models.
Google DeepMind
1 episode · 1 of 10 mentions
Spy Rhymes with AI - DTNS 5373Google DeepMind expands scientific model access to researchers · Daily Tech News Show · Listen
DeepMind is providing 200 European researchers with access to specialized science models and $3 million in computing credits via the RAISE platform.
Make it yours
Start from these trackers
Copy its trackers, shows and tone into a page of your own. Change anything — this page isn’t touched.
Make it mine