Michael Hablich from the Chrome DevTools team shares hard-won engineering lessons on building effective and secure interfaces for AI agents. The talk covers moving from raw data to semantic summaries, measuring interface efficiency with 'tokens per successful outcome', designing for error recovery, and the critical importance of trust boundaries and deliberate friction in UI design for agents.
KP Sawhney from Google DeepMind discusses the internal strategies for scaling agentic AI, including managing token-hungry workflows, curating a 'Darwinian' skills library, and evolving the Deep Research pipeline from large context blobs to a collaborative file system.
The talk argues that while agent runtimes and orchestration are solved problems, the crucial missing piece for building scalable 'software factories' is a dedicated coordination layer. Current tools like GitHub are inadequate, and a new primitive, potentially a CLI gateway, is needed for agents to manage tasks, pass messages, and navigate the software development lifecycle.
Co-founders of Trigger.dev discuss their journey through three product versions to find product-market fit, how their async infrastructure positioned them perfectly for the AI agent era, and their vision for the future of computing: programmatic checkpoint and restore.
Nick Nisi and Zach Proser from WorkOS explain how to build, manage, and scale AI 'skills'—reusable, portable instructions that make AI agents like Claude more powerful and consistent. They cover the core anatomy of a skill, best practices for writing them, advanced techniques like progressive disclosure, and their application beyond coding, from video generation to automating business workflows.
AI agents can interact with the world through either the Command Line Interface (CLI) or the Model Context Protocol (MCP). This summary explores the trade-offs between the two, highlighting CLI's efficiency for tasks the model is trained on, versus MCP's power of abstraction and governance for more complex, high-level operations.