#245 - TML-Interaction, Claude For Legal, Sam Altman on Stand
Summary
Hosts Andrei Krenkov and Jeremy Harris discuss OpenAI’s new real-time voice models and Thinking Machines’ low-latency conversational AI, highlighting the technical race for responsiveness. They analyze Anthropic’s expansion into legal workflows via Claude for Legal, debating whether foundation models will displace specialized startups. The episode also covers Meta’s integration of AI into Threads, reflecting broader trends in agentic social media interactions.
Topics discussed
Intro, podcast reviews, and sponsor segment
OpenAI GPT Real-Time 2 and voice AI competition
Thinking Machines: Real-time conversational AI architecture
Anthropic's legal AI push and the Harvey vs. CoWork debate
xAI Grok ads, Apple AI lawsuit, and Google updates
OpenAI nonprofit lawsuit and Sam Altman testimony
NVIDIA, Trump administration, and AI export controls
AWS-Anthropic partnership and AI black market risks
AI drug discovery funding and Anthropic's Petri tool
Anthropic's research on model alignment and reasoning
Automating AI research and timelines for AGI
AI safety features and chain-of-thought reliability
Critique of mechanistic interpretability and evals
Human consent standards and AI capability extrapolation
Listen ad-free on Castria