EHeller comments on AI risk, new executive summary