If you are able to point to a film with gripping, and exciting dialogue scenes, where several people are involved in an argument or in trying to win someone over to a different point of view; then you have to realize that that scene structure is completely gameable and can exist within an RPG framework.
I don't think I denied such a scene was gameable, or that it couldn't exist in an RPG framework.
What I argued is a good deal more subtle than that. First, that the heavier the rules structure for that scene was, the less likely it was to actually produce gripping and exciting dialogue because the meta would tend to go ahead of the narrative - that is in this case the concrete realized spoken words - and tend to make the concrete realized spoken words less important to the resolution of the scene. They would need to be produced as an afterthought as a result of the meta rather than as an input to it, and that always tends to discourage the production of the words when they are reduced to mere color. You already have your answer, or the means of obtaining the answer independent of the words, so why produce the words?
And secondly I argued that arguments in the real world tend to discourage the participation of all but the most socially adept participants as an inherent nature of them, and the more you try to simulate them with a heavy rules set that depends on characteristics of the character rather than on the spoken word itself the more this is so. This is because conceptually accumulating failures in a persuasion attempt faster than accumulating successes impacts the outcome of persuasion in a way accumulating misses faster than hits doesn't impact the outcome of a combat. Think about an argument or discussion where things are going well, and then someone says something dumb and offensive, and because the person has been offended all the good will that had been built up is now dissipated. This is realistic sort of thing to happen. I've seen it happen in real life. That's a very different situation than one where you are killing a crazed mammoth and an inept ally shoots an arrow and misses it. The mammoth doesn't regain hit points because it was missed. And we wouldn't simulate a combat like that with every missed shot hits an ally, regardless of other factors.
The point is you just can't assume that something that works in one pillar accurately reflects how things work in another pillar.
And if your goal is actually producing the words of that "gripping, and exciting dialogue scene" very heavy rules arbitration doesn't help you do that, and in fact hinders it in every single instance I've encountered from Blue Rose to FATE to whatever. Maybe the closest I have to a system that does do a decent job at this is DitV, but notably is pretty light about how the rules engage and it has a hierarchy where fisticuffs trump argument, and gunfights trump fisticuffs, as well as a lot of GM leeway for interpretation. It's not like the Aces and Eights of social combat, nor does it try to treat each pillar as identical.
That style of game might not be a preference, but they do exist.
So, this is quote from me from another thread that is going on right now:
I frequently find myself sympathizing with B.A. from "Knights of the Dinner Table". I want to run cinematic, highly literary, RP heavy, emotionally affecting games. I want to run games that are high art. My ideal table would probably be a bunch of thespians capable of being voice actors and with a deep commitment to emotionally rich characterization.
Now I grant you, that I don't often achieve that, but over the years I absolutely have achieved some scenes with gripping and exciting and dramatic dialogue from time to time and it's certainly not the lack of a social combat system that hinders me from doing it more often.