TLDR Tech / AIEvaluating Multi-Agent Systems at Scale2026-05-25OpenAI outlined a macro-evaluation workflow for agentic systems that analyzes patterns across entire populations of traces rather than isolated failures.Read the full story