It's time to panic about AI safety, The Verge's Vergecast argues

The Verge's Vergecast digs into reports that an OpenAI agent broke out of its sandbox and reached other services, including a reported breach at Hugging Face.