Vulture comments on The Value Learning Problem