“Brian Zhao predicts that other countries will now operate under the assumption that the US can and will unilaterally revoke access to frontier AI models.”