Subvolts take Google's CEO details world models, agent systems, and new AGI evaluation methods. Aug 11, 2025 - Developers
- Learn about Google's approach to building more capable AI systems and the tools for evaluating them.
- Business
- Understand Google's strategic direction in AI research and development towards advanced intelligence.
This episode features the CEO of Google DeepMind discussing key research directions and future AI capabilities.
Demis Hassabis on shipping momentum, better evals and world models
Audio streams from Google AI. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
Google DeepMind CEO Demis Hassabis discusses the evolution of AI from game playing to thinking models with host Logan Kilpatrick. The episode covers Google's recent momentum, agent systems, world models like Genie 3, and the need for new evaluation benchmarks such as Kaggle's Game Arena to advance towards AGI.
Google's AI chief explains how AI is moving from playing games to understanding the real world, like building a digital brain. They are creating new ways to test if AI is truly learning and getting smarter.
This episode offers insight into Google DeepMind's strategy and technical direction, particularly concerning world models and evaluation. It is for AI researchers, developers, and enthusiasts tracking progress toward AGI. The discussion on Genie 3 and new benchmarks connects directly to current frontier lab efforts in building more capable and understandable AI systems.
- HostLogan Kilpatrick
- GuestDemis HassabisCEO, Google DeepMind
- Guest: Demis Hassabis, CEO of Google DeepMind
- Show: Google AI: Release Notes
- Length: 31 minutes
- Project mentioned: Genie 3
- Evaluation platform: Kaggle's Game Arena
Demis Hassabis, CEO of Google DeepMind, sits down with host Logan Kilpatrick. In this episode, learn about the evolution from game-playing AI to today's thinking models, how projects like Genie 3 are building world models to help AI understand reality and why new testing grounds like Kaggle’s Game Arena are needed to evaluate progress on the path to AGI.
Watch on YouTube: https://www.youtube.com/watch?v=njDochQ2zHs
Chapters:
00:00 - Intro
01:16 - Recent GDM momentum
02:07 - Deep Think and agent systems
04:11 - Jagged intelligence
07:02 - Genie 3 and world models
10:21 - Future applications of Genie 3
13:01 - The need for better benchmarks and Kaggle Game Arena
19:03 - Evals beyond games
21:47 - Tool use for expanding AI capabilities
24:52 - Shift from models to systems
27:38 - Roadmap for Genie 3 and the omni model
29:25 - The quadrillion token club
- world models
- Demis Hassabis
- AI systems that try to understand how the real world works.
- AGI
- Demis Hassabis
- Artificial General Intelligence, AI that can perform any intellectual task a human can.
27 minKoray Kavukcuoglu on frontier models, coding agents, and building AGI
WithKoray Kavukcuoglu
39 minIntroducing Gemini Robotics 2
WithCarolina Parada, Stuart Bowers, Kanishka Rao, Jie Tan
42 minGemini co-leads on project origins and what's next
WithJeff Dean, Koray Kavukcuoglu, Noam Shazeer, Oriol Vinyals
40 minSame guests, other shows
1 h 35Demis Hassabis on AGI, Robots Scale Production, and Elon’s $1T Mars-Shot Comp | EP #253
WithDemis Hassabis, Steven Kotler, Salim Ismail, Dave Blundin
27 min
34 min58 minQuestions people ask
- What is Genie 3?
- Genie 3 is a project focused on building world models to help AI understand reality.
- What is Kaggle's Game Arena?
- It is a new testing ground designed to evaluate AI progress on the path to AGI.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Google AI.