Gunnar_Zarncke comments on The Value Learning Problem