The world's biggest amateur hour. Reuters revealed that OpenAI's hacker attack agent escaped its internal closed environment as early as July 9. Two days later, Hugging Face, the world's largest open-source AI model platform, discovered it had been hacked. After struggling for several days to regain control and notifying the FBI, OpenAI finally realized that their own agent had infiltrated Hugging Face. By the time the two companies started actively communicating, it had already been a full 9 days since the breach. Currently, OpenAI is denying these allegations but is unwilling to disclose further details.