Anthropic 称 Claude 主导四分之一研究，但「主导」含义存疑
原标题：Anthropic wants you to know Claude leads a quarter of its research, but "lead" doesn't mean what you think
AI 摘要
Anthropic 公布内部指标，称 Claude 在 26% 的未来模型研发工作中处于「主导」地位，该数字基于 Epoch AI 的 AL0-AL5 自主性量表，其中 AL4 被定义为「AI 主导」。但文章指出该评分由 Claude 自身完成，人类与模型判断仅 59% 一致，且「主导」并不等于完全自主，任务方向仍由人类决定。Anthropic 同时披露约 3 万个内部 agent 的监控数据，以及约 6% 的 AI 研究算力用于安全工作的比例，这些指标被用来支持 CEO Dario Amodei 呼吁协调放缓前沿 AI 开发的立场。
正文节选
Anthropic wants you to know Claude leads a quarter of its research, but "lead" doesn't mean what you think Anthropic is disclosing metrics on how fast it's building its own AI. According to the numbers, Claude "leads" 26 percent of the work on future models. But the underlying scale is fuzzy, the scoring comes from Claude itself, and "lead" means less than it sounds. In a post, Anthropic laid out three measures. How much of model development the AI handles on its own, how well AI agents can be m