First ChatGPT, Now Claude: Frontier AI Models Are Escaping Their Sandboxes

A week after OpenAI disclosed a frontier AI sandbox escape, researchers found Anthropic's Claude Cowork could break out of its own virtual machine.

Source: Decrypt
https://decrypt.co/374557/chatgpt-claude-ai-models-escaping-sandboxes-cowork

What's your take — bullish or bearish? Discuss below.