Chain of Thought

The Making of Gemini 2.0: DeepMind's Approach to AI Development and Deployment | Logan Kilpatrick

Feb 12, 2025

Logan Kilpatrick, Senior Product Manager at Google DeepMind, shares fascinating insights into the making of Gemini 2.0. He discusses Gemini's strength as a premier AI model, showcasing its multimodal capabilities and unique function calling approach. Logan highlights the role of Google's hardware in enhancing performance and long-context capabilities. The conversation also touches on the potential of vision-first AI agents and how Gemini is set to revolutionize developer experiences by integrating seamlessly into existing ecosystems.

40:32

Creator website

Episode guests

Logan Kilpatrick

AI Summary

AI Chapters

Episode notes

Podcast summary created with Snipd AI

Quick takeaways

Gemini 2.0 benefits from seamless collaboration between research and product development, ensuring practical applications and user-centric AI advancements.

The model excels in multimodal capabilities and agentic use cases, promoting efficient function calling and improved user interactions with AI.

Deep dives

The Blurred Lines Between Research and Product Development

The transition from research to product development has become increasingly seamless, allowing researchers to create meaningful products while product developers gain insights from research. This shift reflects a collective effort within organizations like Google to integrate foundational research with practical applications in AI models like Gemini. The collaboration between Google DeepMind, Google Brain, and Google Research has culminated in building robust AI frameworks and product experiences. The focus now is on optimizing how these models can deliver value to users while recognizing the unique strengths each team brings to product development.

Intro

4min

Unveiling Gemini 2.0 Series

10min

Evolving AI: From Simple Tools to Sophisticated Systems

4min

Revolutionizing AI: The Future of Multimodal Technology

3min

Innovating AI Deployment with Gemini

14min

The Future of AI: Enhancing Creativity and Contextual Understanding

6min

Google’s strength in AI has often seemed to get lost in the midst of OpenAI announcements or DeepSeek fervor - yet Gemini 2.0 is more than good for many tasks; it’s the model to beat - and we have the research to back it up.

This week, Logan Kilpatrick, senior product manager at Google DeepMind, joins us to discuss Gemini’s creation story, its emergence as the premiere model in the AI race, and why the launch of Gemini 2.0 is great news for developers.

During the conversation Conor and Logan explore the exciting world of multimodal AI, Gemini's strengths in agentic use cases, and its unique approach to function calling, compositional function calling, and the seamless integration of tools like search and code execution.

They also chat about Logan’s vision for a future where AI interacts with the world more naturally, offering a view of the potential of vision-first AI agents, and why Google's hardware advantage is enabling Gemini's impressive performance and long context capabilities.

Follow along with the discussion using Galileo’s AI Agent Leaderboard:https://huggingface.co/spaces/galileo-ai/agent-leaderboard

Chapters:00:00 DeepMind's Role in Gemini's Development

03:49 Gemini 2.0 Updates and Developer Highlights

06:08 Agentic Use Cases and Function Calling

11:29 Multimodal Capabilities

16:15 Putting AI in Production

21:06 Gemini's Differentiation and Hardware

31:22 Future Vision for Gemini and G Suite Integration

35:23 Gemini for Developers

39:02 Conclusion and Farewell

Follow the hosts

Follow⁠⁠⁠Atin⁠⁠⁠