Agent Design
Observability Gaps in Agentic Pipelines
Traditional monitoring can't detect when agents make confident but wrong decisions.
Marcus Delacroix-Webb · October 23, 2026 · 9 min read
Section
Agent Design
3 stories in Agent Design
Agent Design
Goal Stability in Long-Running Autonomous Agents
Autonomous agents lose track of original goals during long tasks, not just when context fills up.
Priya Anantharaman · October 22, 2026 · 9 min read
Agent Design
Reliability Measurement for Autonomous Agent Systems
Classical assumptions break when agents produce distributions, not fixed outputs.
Jin-Ho Baek · October 10, 2026 · 10 min read