Lies, Damned Lies, and Evals: How to trust your agents when the stakes are high
AIE NYC · Oct 2026 · New York, NY
A talk about comparing agent systems, tracing the causes of changes, and checking whether an evaluation measures what an operating decision needs.
Slides ↗ Talk page (all slides, notes, downloads) ↗ llms.txt (agent-readable talk) ↗ coding-agent-workflows (evaluation procedures) ↗