This is good (though old) work. One interesting aspect is that even though they tout their models ability to produce grammatical text, their abstract has a numb

This is good (though old) work. One interesting aspect is that even though they tout their models ability to produce grammatical text, their abstract has a number of grammatical errors. On a dark night in a melancholic mood, that might seem to bear on the question of what it is to be truly human and bar that we set for (linguistically) anthropomorphic automatons like LLMs.