Tag
The article discusses AI security as an engineering problem, emphasizing the need for security controls across all layers of the agent stack and introducing tools like NVIDIA OpenShell to enforce policies.