Google DeepMind's Logan Kilpatrick on Agentic AI, Coding Agents, and Omni Models
In this episode, Logan Kilpatrick, head of Google AI Studio and the Gemini API, discusses the evolution of Google's AI strategy centered around agentic AI and the 'anti gravity' agent harness that powers a growing number of Google products. He explains how this harness enables agentic capabilities across coding, search, and consumer applications, emphasizing a shift from maximizing user eyeballs to maximizing user outcomes. Logan also highlights the rapid progress in coding agents, describing them as a form of narrow superintelligence that accelerates software development and research productivity. The conversation covers Google's approach to integrating AI into developer tools like AI Studio, where hundreds of thousands of Android apps have been generated, and the challenges and opportunities in generative media and video game creation using coding agents combined with world models. Logan elaborates on the concept of 'the model eats the harness,' describing how AI models increasingly internalize external scaffolding and agent frameworks, which may shift where competitive advantage lies in the future. Finally, he reflects on DeepMind's unique culture within Google, balancing deep scientific research with applied product deployment across Google's ecosystem.
Original sources
Podcast Training Data
Episode:Google DeepMind's Logan Kilpatrick: Why the Model Eats the Harness ↗
YouTube Training Data
Video:Google DeepMind's Logan Kilpatrick: Why the Model Eats the Harness ↗
AI Radar wrote this summary from the publisher’s notes with AI, not from a transcript. How it works.