I gave it a try. The difference is that XINPUT2 scroll has inertia: I start scrolling, remove the fingers from the touchpad and it keeps scrolling for a while. If I want to stop when I see something interesting I have to click or it will scroll away. That adds an extra action that must performed at the right time. I don't think I like that because this is not a videogame and Firefox is not supposed to challenge my abilities. I prefer the current behavior: it scrolls until I stop sliding my fingers on the touchpad. Then I have to move them up to the top of the touch pad and start again but I can scroll all the front page of HN in a single motion and it gives me time to check what I'm looking at, especially on longer pages (this comment thread.)
Anyway, it's nice to have different settings that suit different people.
Try it for some longer time. Scrolling on Macbook becomes subconscious. I don't think about it anymore. I don't even notice that I am moving my fingers. The scroll pane just moves where I want it to move. Simulating the behavior of physical objects allows our minds to reuse what it learned about the physical world. That is imho a very powerful user interface metaphor. But beware, this is just an anecdote.
There are two other problems I thought about later:
1) It's too fast. My Android phone also scrolls like XINPUT2 but it's got a smaller acceleration. But maybe it's what you say about getting used to it.
2) That would work in Firefox and every other window would still scroll in the other way. That's really bad and works against getting used to a specific scrolling behavior. I don't get it: the scrolling behavior should be managed at a level below every application. All the scrollbars should behave in only one way, except the random application that implements its own logic because of some important reason. So I would expect a system wide setting for XINPUT2 or traditional scrolling.
Anyway, it's nice to have different settings that suit different people.