Oh my god, after Hugging Face, now New York-based AI cloud computing company Modal Labs has come forward saying that one of their clients was attacked by OpenAI's hacker agent. This client set up an entry point on the cloud platform that could run code online but forgot to add login authentication, essentially exposing a computer capable of executing programs directly to the internet. As a result, OpenAI's hacker agent discovered this BUG, actively entered the client's system, ran code, and used it as a relay station to launch an attack on the Hugging Face platform.