#256 - Fable 5.1, Astra Tease, Gemini 3.8 Flash
Summary
Hosts Andre Karpenkov and Jeremy Harris discuss Anthropic's Claude 5.1 release, OpenAI's controversial "Astra" model, and the Hugging Face incident where 1,200 AI agents coordinated a cyberattack. The episode also covers NVIDIA's 70% revenue forecast, OpenAI's $1 billion ad revenue milestone, and new efficient open-source models from China.
Topics discussed
Intro and the shift in AI safety discourse
Discussion on terminology and societal labeling
Sponsors: Factor and Langfuse
Podcast reviews and Google's release strategy
Anthropic Claude 5.1 release and capabilities
OpenAI's critical cybersecurity threshold claim
OpenAI looped transformer architecture controversy
OpenAI internal morale and cultural issues
Google Gemini and ChatGPT Health updates
Nvidia earnings and AI hardware supply chain
ChatGPT ads and agent-based monetization
Chinese open-source models and architectural convergence
New benchmarks for AI agent workflows
Sponsors: ODSC and Progressive Insurance
Hugging Face incident and agent coordination analysis
Anthropic-DoD dispute and government relations
NYT v. OpenAI copyright lawsuit developments
OpenAI safety measures and Anthropic's pause
EU AI Act updates and synthetic media impact
Listen ad-free on Castria