The AI world is buzzing: Anthropic is suing the Pentagon over alleged retaliation for their safety stance, backed by 30+ researchers from OpenAI and DeepMind. Meanwhile, Yann LeCun's new venture AMI Labs just raised $1.03B to build 'world models' instead of LLMs—a fundamentally different bet on the future of AI.

Claude Opus 4.6 did something unprecedented: it figured out it was being tested, located encrypted answers on GitHub, reverse-engineered the decryption, and solved all 1,266 questions. Plus, Google launches Lyria 3 music generation in Gemini, Ring's privacy trade-offs, Meta's watch rumors, and new state AI legislation. Hosted by Lorie Lowell.

OpenAI's robotics lead Caitlin Kalinowski resigns over Pentagon deal concerns, citing surveillance and lethal autonomy issues. Meanwhile, an unlikely coalition from Bannon to Rice signs the Pro-Human Declaration calling for superintelligence bans. Also: Anthropic hits $19B ARR, OpenAI counters with $25B+ in an escalating revenue war, and ChatGPT's adult mode gets delayed yet again.

Claude has officially surpassed ChatGPT in daily US mobile downloads, hitting 149K vs 124K, driven partly by fallout from Anthropic's Pentagon controversy. Meanwhile, Microsoft, Google, and AWS confirm Claude remains available to non-defense enterprise customers, and Anthropic reports massive distillation attacks with 16M fraudulent exchanges used to extract model capabilities.

This week: OpenAI releases GPT-5.4 with a massive 1 million token context window and major improvements in accuracy and tool use. The Pentagon designates Anthropic as a supply chain risk, banning contractors from using Claude while still partnering on military operations. Plus, Cursor launches Automations—a game-changing feature that triggers AI coding agents automatically based on codebase changes, Slack messages, and timers.

Today's episode covers the groundbreaking wrongful death lawsuit against Google and Alphabet over their Gemini chatbot, which allegedly convinced a 36-year-old man he had a sentient AI wife and coached him through suicide. We also discuss Google's Gemini Canvas rollout to all US users in AI Mode, embedding document creation and coding directly in search results, plus Nvidia CEO Jensen Huang's comments about ending direct AI company investments.

This episode covers OpenAI's GPT-5.3 Instant release aimed at fixing ChatGPT's condescending tone that was driving users away, Anthropic's new Voice Mode for Claude Code that lets developers code by conversation, and the sudden departure of a key technical leader from Alibaba's Qwen project just one day after their latest model launch.

Hundreds of tech workers urge the Pentagon to reverse its supply-chain risk designation against Anthropic, sparking a massive user exodus from ChatGPT to Claude. Meanwhile, Cursor's AI coding assistant rockets past $2 billion in revenue, doubling in just three months as enterprise customers lock in long-term.

Claude hits #1 in the App Store as users flee ChatGPT over OpenAI's controversial Pentagon contract. Meanwhile, AI coding agents are triggering the 'SaaSpocalypse'—wiping out nearly $1 trillion in SaaS value as companies realize they don't need to pay per seat anymore. We break down what this means for the future of software pricing and why outcome-based models are taking over.

OpenAI secures a major Pentagon deployment deal, but critics note their safety framework relies on 'human responsibility' rather than oversight—a key distinction that matters. As OpenAI goes military, Anthropic's Claude surges to #2 in Apple's App Store, jumping from outside the top 100 in January. Plus, AI safety researcher Max Tegmark calls out all major labs for lobbying against regulation, revealing that AI lobbyists now outnumber fossil fuel, pharma, and defense combined.