#245 - TML-Interaction, Claude For Legal, Sam Altman on Stand
Summary
Hosts Andrei Krenkov and Jeremy Harris discuss OpenAI’s new real-time voice API and Thinking Machines’ low-latency conversational model, highlighting the trade-offs between speed and reasoning. They analyze Anthropic’s expansion into legal workflows via Claude for Legal, debating whether foundation models will cannibalize specialized AI startups. The episode also covers Meta’s integration of AI agents into Threads and touches on broader industry trends in interoperability and policy.
Topics discussed
Intro, podcast reviews, and sponsor message
OpenAI GPT-4o Realtime and GPT-5 updates
Thinking Machines and real-time AI architecture
Anthropic Cloud for Legal and AI business models
xAI Grok ads and Google Gemini Android features
Google Search AI integration and features
OpenAI board drama, Sam Altman, and NVIDIA summit
Anthropic on AWS and the gray market for AI tokens
Isomorphic Labs raises $2.1B for AI drug design
Open source tools: Petri, MCP, and Daybreak
Anthropic's alignment research: Teaching Claude Why
Jack Clark on AI timelines and the 'foom' debate
ChatGPT safety alerts and chain of thought risks
Critique of Anthropic's interpretability and risk reports
Celebrity-backed AI likeness registry
Mythos model evaluation and closing thoughts
Listen ad-free on Castria