aiqbench analysis

This article is squarely about an AI product: Anthropic's Claude Duet, a proactive agent that handles scheduling, forecasting, and app-building for professionals in research preview. For AI readers, the notable points are the shift toward always-on background agents that check in rather than waiting for prompts, and the safety design (an automatic classifier screening for data leaks and confirming high-stakes actions). Claims about usefulness rest on Anthropic's own description and internal use, so independent testing will matter more than the launch framing.