Beyond Short-Termism: How δ and w Can Realign AI with Our Values