SITUATION EXPLAINED: ChatGPT Work found 13 vulnerabilities on Greg Brockman's personal site in 15 minutes. • In the Hugging Face incident, an agentic swarm autonomously penetrated not just OpenAI research infrastructure but another company's production systems • ChatGPT Work found 13 issues on http://gregbrockman.com in 15 minutes, then fixed them over the next hour • Almost all of OpenAI's initial security alerts are now triaged by models before humans are looped in • OpenAI has started…