That's a great description of the boundary between logical deduction NLP and bullshitting NLP.

I still have hope for the former. In fact, I think I might have figured out how to make it happen. Of course, if it works, the result won't be stubborn and monotone..