If algorithms reward outrage and punish nuance, how many minds are being trained to stay shallow?