Evan R. Murphy comments on Emergent Abilities of Large Language Models [Linkpost]