Event Notes · English · Five Parts
Build With Google
Day 1 2026
Five English field notes connecting the event’s labs to the engineering structures required to delegate real work to AI agents.
Once AI became intelligent, the next thing we needed was a system that would let us entrust it with work.
Five-Part Series
0102030405
Beyond AI That Can Write Code—What Is Agentic Engineering?
The foundation for turning capabilities into work that can be delegated continuously.
What Does Antigravity Manage About AI?—Understanding It as a Workplace, Not a “Chat”
Understanding Projects, Conversations, permissions, MCP, and Artifacts as a single workplace.
From Long Prompts to Operating Manuals—Designing Agent Skills
Designing reusable Skills that read the necessary procedures only when needed.
Separating Ambiguous Judgment from Precise Control—Agents CLI and ADK Graph
Separating the LLM’s semantic judgment from route control using Schema, Python, and Graph.
How Can We Trust an AI Agent’s Claim That It Is “Done”?—Evaluating Output and Trajectory
Inspecting not only the final output, but also the path taken to produce it.