NYU and UT Austin researchers identify linguistic patterns causing AI text fatigue. Frontier models overuse complex structures due to code and math training. Smaller models produce clearer text, suggesting a trade-off between size and readability.
Opening Kapyn…