𝓵𝓮𝔃𝓲𝓼𝓱𝓮𝓷 (@lezishen) 在 “过于先进不便展示”的最强 Ai 安全模型 Mythos，上线当天即被未授权访问 中发帖
https://www.bloomberg.com/news/articles/2026-04-21/anthropic-s-mythos-model-is-being-accessed-by-unauthorized-users
[image]
[image]
[image]
[image]
[image]
[image]