The episode surveys an accelerating AI landscape where new hardware like Cerebras and Groq enables near real?time model responses, making voice and agent interactions feel instantly conversational. The conversation covers the rise of code models (Codex, Claude Code), practical tips for using multiple models to check each other, the tug-of-war between frontier labs and big incumbents (OpenAI, Anthropic, Google, Meta, xAI), and how talent, salaries, and state-level data?center politics are shaping the field. They also touch on a striking story about a dog treated with an experimental mRNA therapeutic assembled with help from multiple AI tools, hands-on demos of rapid content generation and deepfake video, and a challenge to listeners to build weird things with these new tools.