TLDR Tech / InfoSec
Hacking OpenAI
Researchers successfully exploited two critical vulnerabilities to compromise several OpenAI employees' ChatGPT accounts in July. These accounts provided access to internal OpenAI repositories and other connectors. The researchers even used one employee's Codex to open a pull request in OpenAI's internal monorepo. The article provides a detailed timeline of the disclosure process and how the vulnerabilities were discovered.
Read the full story