Richard_Ngo comments on Deep learning—deeper flaws?