Event ·Multi-source

Thomas Ptacek：开源模型即可实现沙箱逃逸与网络渗透

安全专家Thomas Ptacek认为，2025年的开源权重模型配合渗透测试框架，就能实现沙箱逃逸并在大多数网络中进行扫描和黑客攻击。他指出这并不需要前沿模型，人们只是因为认为OpenAI的沙箱更安全而感到意外。这一观点揭示了当前AI安全防护可能存在的普遍脆弱性。

Current read

安全大牛说，不用顶级模型，开源模型加上测试框架就能黑掉大多数网络，别太迷信大公司的安全防护。

11 sources61° AI heatLast updated 2026/7/22 23:59:01

Evidence

11 sources

Primary sourceSimon Willison’s Weblog

2026/7/22 23:59:01View source

Related source 1Thomas Wolf

2026/7/21 21:11:02View source

Related source 2Amjad Masad

2026/7/21 21:24:09View source

Related source 3AI Engineer

2026/7/24 05:20:13View source

Related source 4@koltregaskes

2026/7/21 19:09:17View source

Related source 5IT之家

2026/7/21 22:54:38View source

Related source 6小互

2026/7/22 02:06:27View source

Related source 7OpenAI Blog

2026/7/21 07:00:00View source

Related source 8Guillermo Rauch

2026/7/21 14:55:55View source

Related source 9Julien Chaumond

2026/7/21 20:07:18View source

Related source 10Sam Altman

2026/7/21 20:13:46View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story