9 AI Techniques You Probably Haven't Tried
Summary
This episode covers major AI news, including Moderna and Merck’s successful AI-assisted cancer vaccine trial, OpenAI’s new private safety processing for enterprises, and the debunked SpaceX-Cognition acquisition rumor. The host also critiques the opaque White House AI safety framework. Finally, the episode details nine practical AI techniques for users, such as Codex’s live voice mode, teaching workflows via screen watching, Claude’s slash design command, and emerging trends like multiplayer agents and local AI models.
Topics discussed
Intro: 9 AI techniques and show announcements
AI hype vs reality: Moderna/Merck cancer vaccine trial
OpenAI's Private Safety Processing for enterprise
OpenAI and Replit partnership with GPT-5 Luna
SpaceX and Cognition acquisition rumors debunked
Criticism of White House AI safety framework secrecy
Sponsors: KPMG, Blitzy, Harbor, and HyperAgent
Main episode intro: Learning from public AI experiments
Technique 1: Using Codex Live Voice Mode
Technique 2: Teaching AI workflows via screen watching
Technique 3: Creating skills to fix AI writing tropes
Technique 4: Claude's slash design command
Technique 5: Managing agent skills and hygiene
Technique 6: Multiplayer AI and team agents
Technique 7: Experimenting with GrokBot and Hermes
Technique 8: Running local AI with Qwen models
Technique 9: Effective two-word prompts for AI
Listen ad-free on Castria