If you asked all the LLM to find flaws in the arguments presented, and to come up with counter-arguments, I doubt many current models would be so bad as to come up with that, and even the ones that did would fold when asked "how is that even an argument?".