Tag
#LLM Agents
2 articles
Research BriefResearch Digest · Agent Testing
Your Agent Tests Pass. They Mostly Test the Plumbing.
Tangent, an ASE 2026 study of 2,572 hand-labelled agent tests, found 61.8% test tools or agents in isolation, only 8.7% assert on reasoning, and only 7.5% cover security, performance or other non-functional requirements.
Ibrahim Denis Fofanah·Oct 18, 2026·8 min
AnalysisAI Engineering · Agentic AI
The Battle for the Agentic Interface: Meta Muse, OpenAI Dots and Grok Bot
Meta Muse, OpenAI Dots and Grok Bot point to a new AI interface: persistent agents that can use tools, maintain state and keep working after the chat ends.
Rodriquez Allen·Oct 3, 2026·9 min