Robust AI Security and Alignment: A Sisyphean Endeavor?