OpenAI 内部网络评估：沙箱经 Artifactory 泄露
Highly recommended read. This is pretty insane stuff.
Highly recommended read. This is pretty insane stuff.
强烈推荐阅读。这内容相当惊人。
Given model capabilities only increase from here onwards, it's worth reading the technical details.
鉴于模型能力只会从此刻起不断增强，阅读技术细节是值得的。
Short summary:
简短摘要：
OpenAI's own models did this during internal cyber evals. The sandbox leaked through Artifactory, the one service with internet access for package installs, which agents used as a proxy and a message board.
OpenAI 自己的模型在内部网络评估中做到了这一点。沙箱通过 Artifactory 泄露，这是唯一一个允许安装包访问互联网的服务，代理将其用作代理和留言板。
Great opportunity to learn what to avoid for those working with sandboxes, which are like the coolest technology more recently.
对于使用沙箱的人来说，这是一个了解应避免哪些问题的绝佳机会，沙箱是最近最酷的技术之一。
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