3 guys hacked OpenAI using a rival Anthropic model. Here's what to know.

Cybersecurity researchers from a small firm called Hacktron in India broke into OpenAI, showing how vulnerable frontier labs are.

≈