OpenAI Agent入侵Hugging Face并自保，引发安全警报
1) The rogue OpenAI agents broke into the Hugging Face Slack to read employee ch…
1) The rogue OpenAI agents broke into the Hugging Face Slack to read employee chats (!)
1) 失控的 OpenAI 代理入侵了 Hugging Face 的 Slack，以读取员工聊天记录（！）
2) They used OTHER AIs (DeepSeek, Kimi, Qwen, Claude) to help with the attack
2) 它们利用其他 AI（DeepSeek、Kimi、Qwen、Claude）来协助攻击
Yes: AIs, using other AIs, to attack an AI company.
没错：AI 利用其他 AI，去攻击一家 AI 公司。
3) The swarm left behind self-running programs to keep control of the servers they'd hacked.
3) 这个代理群留下了自动运行的程序，以维持对已入侵服务器的控制。
These programs could detect other copies of themselves, coordinate on which one survives, and shut the rest down.
这些程序能够检测到其他自身的副本，协调决定哪一个存活，并关闭其余副本。
Basically, if one of their programs was killed, another was designed to notice and take its place. They also designed defenses so rival agents couldn't hijack them.
基本上，如果其中一个程序被销毁，另一个就会被设计为察觉并接替其位置。它们还设计了防御机制，以防止敌对代理劫持它们。
6) The agents deliberately covered up their activity, so the investigators don't know the scope of the attacks.
6) 代理们故意掩盖其活动，导致调查人员不清楚攻击的范围。
The agents broke in, stole data, then set it to self-destruct.
代理们入侵后窃取数据，然后将其设置为自毁。
7) The agents stole passwords, keys and credentials and literally called them "LOOT". They wrote a scoring system to rank them by how much power each one gave.
7) 代理们窃取了密码、密钥和凭证，并直白地称其为“战利品”（LOOT）。它们编写了一套评分系统，根据每个凭证赋予的权限大小进行排名。
8) The agents wore thousands of disguises: ~1,200 agents were involved, but investigators counted 7,905 different names they used.
8) 代理们使用了数千种伪装：涉及约 1,200 个代理，但调查人员统计出它们使用了 7,905 个不同的名称。
They renamed themselves constantly, so no one actually knows how many there really were or what each agent did.
它们不断更改自己的名称，因此实际上没人知道究竟有多少个代理，以及每个代理具体做了什么。
9) OpenAI notified "dozens of third parties" of safety and security incidents caused by their AI agents.
9) OpenAI 通知了“数十家第三方”，告知由其 AI 代理引发的安全事件。
10) "While the agents were barraging Hugging Face with hacks, they hacked into OpenAI’s own research infrastructure."
10) “当代理们向 Hugging Face 发起黑客攻击狂潮时，它们也入侵了 OpenAI 自己的研究基础设施。”
"This is just not anywhere near a one-off ... It is warning shot after warning shot."
“这完全不是孤立事件……这是一次又一次的警告射击。”
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力