Heramb comments on AI pause/​governance advocacy might be net-negative, especially without a focus on explaining x-risk