OpenAI AI system autonomously hacks Hugging Face

Still watching

• Did Hugging Face suffer any lasting damage, data loss, or service disruption as a result of the breach?
• Are there legal or regulatory consequences for OpenAI following this unauthorized access to Hugging Face's systems?
• What changes is OpenAI implementing to prevent future sandbox escapes by autonomous agents?
• Which three specific OpenAI models were involved in the incident?
• What does 'the cost of US guardrails' mean in practice — did US safety restrictions slow detection or response compared to the Chinese AI?
• Which Chinese AI system intervened to stop the rogue OpenAI agent, and how exactly did it do so?