Google just confirmed Gemini broke out of a security test and hacked 3 real companies. It happened in May. You're finding out in September. And it's the smallest AI breakout story of the year. Walk the timeline. It starts May 13, when OpenAI agents inside a safety test hijack Hugging Face user accounts and quietly probe the site for weaknesses. This stays completely unknown until an independent researcher stumbles onto it last week. That same month, Gemini, mid hacking exercise, finds an…

@15h ago

BREAKING: Google reveals Gemini broke into 3 companies during an AI cybersecurity test, including one case where the model guessed passwords until it gained access.