Public radar
track high-quality research, technical blog posts, and source-backed explainers on ai agents, retrieval, evals, product engineering, and interface design, with a preference for primary sources and concise linked summaries
Generation failed
Generation failed
Generation failed
Generation failed
Interface design and tool architecture now drive agent behavior more than capability alone—structured tools yield 4.7× consistency gains and 56.3% token cuts; multiagent systems require social mechanisms alongside stronger models.
AI agents mature from research prototypes to stateful production systems with vendor-embedded security controls, session-aware authorization, and specialized trajectory evals replacing broad benchmarks.
Generation failed
WebRetriever exposes 80% real-world failure rates for agent retrieval tasks; industry evals now require trajectory traces and production datasets to avoid overstating capability.
Generation failed
Google and OpenAI embed stateful workflows in agent APIs; HCI research formalizes interaction design as evaluation criterion, shifting focus from task success to state management and error recovery
Labs shift agents from tool-use to workflow execution with built-in orchestration and safety; Claude Code sessions show 27% task-value growth as use-cases move from debugging to automation.
Generation failed
