Anthropic Mythos 泄露与运行时治理的崛起——Agent 安全范式转移

2026 年 3 月 27 日,Anthropic 因 CMS 配置错误泄露了约 3,000 份未发布资产。其中一份草稿描述了代号 Mythos(内部称 Capybara)的下一代模型,声称在 coding、reasoning 和 cybersecurity 上有显著进展,且在网络安全能力上"far

Administrator Administrator Published on 2026-04-01

The Anthropic Mythos Leak and the Rise of Runtime Governance — A Paradigm Shift in Agent Security

On March 27, 2026, Anthropic leaked approximately 3,000 unpublished assets due to a CMS configuration error. Among them was a draft describing the nex

Administrator Administrator Published on 2026-04-01

Learning Curves 背后:AI 使用能力比 AI 本身更重要

Anthropic 在 3 月 24 日发布了第五期经济指数报告《Learning Curves》,基于 2026 年 2 月的 Claude 使用数据。这份报告最有价值的地方不是告诉我们 AI 能做什么,而是揭示了**谁在有效使用 AI,以及为什么**。 核心发现很简单,但影响深远:**使用 Cl

Administrator Administrator Published on 2026-03-28

Behind Learning Curves: AI Usage Skills Matter More Than AI Itself

Original Report: Anthropic Economic Index report: Learning curves On March 24, Anthropic released its fifth Economic Index report, "Learning Curves,"

Administrator Administrator Published on 2026-03-28