Generative AI is transforming countless aspects of society, from education to healthcare to shopping to travel. Much of this impact has come from chatbot-like applications. While the progress in these applications, as measured by standardized benchmarks, is impressive, this is just the beginning. We are on the cusp of a new era with more advanced AI models, capable of native multimodality — understanding and generating different types of content like text, images, audio, and even code — and exhibiting more autonomous, agentic behavior. These developments are opening up dramatically new avenues for impact and creating exciting opportunities for developers.

However, more powerful models alone are not enough. As developers, we need to integrate them into user-friendly applications. As these models become more sophisticated, we need to carefully consider how they are designed, deployed, and monitored. This introduces new considerations for developers in areas like evaluation, testing, and debugging, especially for multimodal and agentic systems.

In this talk, I will review recent advances in GenAI models and applications, with a particular focus on the implications for developers. I will also dive into how GenAI has revolutionized product development, focusing on novel aspects such as evaluation, red-teaming, prompt engineering, and orchestration that are now essential for building AI-first product experiences. Attendees will gain insights into the future of AI beyond chatbots, learn about the new challenges and opportunities that multimodal, agentic frontier models like Gemini present for developers, and acquire practical knowledge to navigate the evolving landscape of AI-driven product development.