Exactly what I was thinking. The ability to reliably predict human scores, says more about the terrible way we "standardize" creative writing, than about the q

Exactly what I was thinking. The ability to reliably predict human scores, says more about the terrible way we "standardize" creative writing, than about the quality of the algorithms. If the grading curve of an essay is a checklist how is this finding a surprise?