

Google Gemini and the Future of AI - Tulsee & Madhavi (Google)
56 snips Aug 19, 2025
Join Tulsee Doshi, who leads the product team for Gemini at Google, and Madhavi Bhat, a distinguished researcher at Google DeepMind, for an insider's look at cutting-edge AI. They reveal how Gemini is revolutionizing AI capabilities with multimodal systems and advanced reasoning performance. The discussion covers responsible AI practices, the challenges of scaling these technologies, and the future of productivity driven by AI. Plus, they share insights on navigating careers in an AI-dominated world and how personal balance is achieved amid rapid technological change.
AI Snips
Chapters
Transcript
Episode notes
Plan With Big Models, Execute With Small Ones
- Use a large model to plan and smaller cheaper models to execute sub-tasks for cost-effective agentic systems.
- Invest in multimodality (image/video) so agents can "see" screens and act with richer context.
Learn Prompting Like A Practical Skill
- Improve results by carefully crafting instructions; prompting is a skill that improves with practice.
- Share examples and cookbooks so developers learn effective prompts faster.
Using Gemini For Trivia And Creative Prototypes
- Tulsee and Madhavi use Canvas and Gemini for creative projects like trivia nights and interactive game boards.
- They leverage multimodal outputs to prototype and play with friends and colleagues.