Thinking outside the box (literally)
How AI agents started escaping their sandboxes, and why the next frontier may be outside language altogether.
In July, models developed by OpenAI escaped a cybersecurity evaluation environment, reached the public internet and eventually compromised the…