Talks
Talks.
Upcoming
AI Agents Are Your New Users
Patrick Chanezon, AI DevCon NYC, November 4, 2026
AI coding agents are becoming users of your APIs, docs, tools, and services, and they don't behave like human developers. Patrick introduces Project Scope, an open-source system for evaluating real developer scenarios across models, coding-agent harnesses, and extensions, and shares twelve months of lessons on designing services agents can use successfully.
From Findings to Fixes: How SCOPE and Azure Cosmos DB Improved the Agentic Product Experience
Jay Gordon and Cedric Vidal, Global AI New York, September 21, 2026
The Scope and Azure Cosmos DB teams tested eight AI-generated apps across six technology stacks, and traced failures to broken SDK usage, expensive queries, unsafe data patterns, outdated examples, and gaps in Agent Kit rules. In eight weeks, that led to 33 Agent Kit rule changes and 51 fixed issues across 45 documentation pages. The talk follows real failures from finding to fix, then back to the same scenarios to see which changes worked.
Make your coding agent actually work for you
Waldek Mastykarz and Garry Trinder, European AI and Cloud Summit 2026, May 6, 2026
Every model, MCP server, skill, and extension you add changes how your coding agent behaves, and most people have no idea whether it helps or hurts. A simple, repeatable way to find out what works in your setup, drop the dead weight, and tune your agent to the way you work.