It's about topological quantum computers involving non-Abelian pseudoparticles, if they exist you could build a quantum computer that produced FAR fewer errors than a conventional quantum computer. This paper does not claim to have proven they exist in our real physical universe, however from the point of view of an engineer trying to make an improved quantum computer that might not matter very much because they did prove that they can make non-Abelianpseudoparticles in a simulated virtual world and perform low error quantum calculations in that world. And there is no difference between a simulated calculation and a "real" calculation.
Of course it would be preferable if real physical non-Abelian pseudoparticles did exist in the real world because it takes some computational horsepower to create that virtual world, and maybe they do exist, but if not then this is an alternative. There is every reason to believe that algorithms will improve allowing conventional computer hardware to efficiently emulate that virtual world. The big question is, how much conventional hardware is needed to create and maintain that virtual world. If the overhead can be made modest (and the algorithms are improving) then whether non-Abelian pseudoparticles are "natural" or not becomes unimportant, at least to an engineer trying to build a quantum computer, although a physicist interested in discovering the nature of physical reality might feel differently.