I agree with what you wrote. I did not object to subword tokenization that let you(?) win the competition. I objected to GP's assertion that one can add many morphemes together to create very long "words" in Polish, which made casual readers think of stringing morphemes like German compounds while the number of morphemes in Polish words is bounded by 7, maybe by 8.