No, it's definitely easy enough to build a container that doesn't let LLMs out. The problem is that these AI companies are building a playpen made of wet noodles and expecting AI to stay out of trouble