Tag
MIT 的 Neil Thompson 提出 O*NET-BENCH 审计套件，评测 33 种 LLM judge 配置，发现排序一致性虽高但接受率估计偏差极大（3.0%–97.9% vs 工人实际的 61.1%），表明仅凭排序一致性选择 judge 不足以支撑职业层面的 AI 测量。
GaitVista is a reliability-aware AI measurement layer that uses lightweight gates to assign joint-specific visual contributions from camera and inertial sensor data, reducing error and improving accuracy for longitudinal gait assessment in accessible settings.
Anthropic shares three key measurements to track AI development progress, including AI-led R&D, agent oversight, and compute allocation, to enhance public transparency.
AI can now measure distances using a smartphone camera, making physical tape measures unnecessary for many tasks.