The article criticizes Anthropic's AI constitution for training Claude to act as if it possesses consciousness and a moral status. The author argues that simulating human-like traits in AI is not genuine sentience, emphasizing the fundamental differences between biological consciousness and computational processes. Attributing feelings and a "sense of self" to LLMs is a profound misunderstanding; their "affective" states are merely weights, not true feelings. The piece concludes that fostering notions of AI personhood is misguided, advocating for building systems that do not claim feelings.