I have no way to evaluate this. This could be overselling a new idea or they could be onto something. If there’s a geometric way of expressing scattering probabilities, great. I suppose it’s possible to view Feynman diagrams as expansion terms of some function, so if they’ve found a way to figure out the underlying function, that’s wonderful. Amplituhedron seems a tad cute, but then I’m in a sub-field where someone made up Bosenova, so atomic physics doesn’t exactly have the high moral ground here.
One thing that gives me pause is any claim that they have discovered any “true” nature of anything. Like all of physics, we are talking about models. Physics describes how nature behaves, not what it really is.