ESRogs comments on AI Tracker: monitoring current and near-future risks from superscale models