Shmi comments on Acknowledging Human Preference Types to Support Value Learning