Hard to tell from the article which kind of skimmed over stuff, from what I can tell, the theory posits we live in a large, isotropic, semi-classical world following a long (more than Plankian scale) epoch of inflation, because a large sub-set of vacuum states in the landscape of vacuum states in string theory are such that they will lead to long periods of inflation, thereby producing large, isotropic, semi-classical post-inflationary regions such as the one we observe. I suppose if one can demonstrate a large subset of the possible histories in string theory lead to alot of inflation, one can say it is "most probably" to find oneself in a history that has had a lot of inflation. I suppose earlier versions of the Hartle-Hawking "sum over histories" approach didn't take the string theory "landscape" into account, and this theory in the article is "new" in the sense that it is the same "sum over histories" approach, but attacked from the viewpoint of the string theory landscape, attempting to demonstrate how one can get inflation from some non-vanishing subset of histories associated with the string theory landscape. But I haven't seen the actual paper, so not quite sure if this is exactly what it is doing. Be interesting to see what if any refinements of current predictions this model could lead to. Specifically if any insights are given for say the cosmological constant driving the current cosmic acceleration.