Claire Vo ranks four personal AI agents across nine tasks in the How I AI Bench

Claire Vo publishes the How I AI Personal Agent Bench, a test in which she runs four personal agents against nine tasks to determine which is best.

In her post on X, she names the ranking as @bot first, by a wide margin, followed by Muse, ChatGPT Dot, and Duet from Claude, in fourth place.

The post links to a full episode on YouTube. The report gives the names and order but not the individual task scores or the criteria behind each task.