Nathan Helm-Burger answers Potential alignment targets for a sovereign superintelligent AI