If you had the feeling that there were too many people today who were opening the blog, be sure that it was just a temporary effect of the no-confidence vote yesterday. The Reference Frame has had 2927 unique visitors today (the 24-hour period ending at 6 p.m.) and 6346 total visits, and I assure you that this number will decrease back to the usual values that are dominated by the physicists and the physics fans.
Steve Gubser (from Princeton) has just gave an interesting talk at the joint seminar in which he tried to convince us that structure formation (the process in which the early clumps of matter and the first galaxies were born) is a very interesting topic in cosmology, even for string theorists, in which some signs of new physics may be found if one tries to reproduce the observations.
He argued that the data behind structure formation are more exact than the data that are used to discriminate between the different models of inflation. Structure formation has been studied less intensely by string theorists than inflation (well, it's because it's closer to astrophysics and seemingly further from fundamental, extremely high-energy physics), and the comparison should become more balanced, he argued. One of the observations he was trying to explain were the more-than-expected empty voids in between the clusters of galaxies which are marginally contradicting some computer models of structure formation.
His solution was based on an extra ultralight scalar - whose inverse mass is roughly a Megaparsec (the typical size of the haloes around the galaxies). The corresponding "fifth interaction" only couples to the dark matter (dark sector) in the first approximation, and it only induces dimension six operators in the visible (baryonic) sector which is below the threshold what is observable. The equivalence principle has not been tested too well for dark matter, so such an extra fifth force is not directly falsified.
The exchange of a massive scalar gives you a Yukawa force that exponentially decreases at distances of order 1 MPc. Recall that the exchange of a scalar is attractive between sources of the same sign - much like for gravity whose spin is two, and unlike the electric forces that exchange a spin 1 gauge boson. This can be seen to help gravity to speed up the cluster formation and empty the voids more efficiently, which is good for observations, he argued.
(An example of the effect of the scalar is the scalar called the T-modulus, a radius of a circle, which will try to repel the momentum modes and the winding modes of a string because they carry the opposite sign of charges with respect to the scalar.)
This basic story sounds logical, but one could probably imagine many other ad hoc modifications that have the effect (for example, Jacques Distler proposed that the voids are actually filled by "dark galaxies" where the stars have not started yet). Instead of speculating, let me finally mention a relevant paper by Gubser, Nusser, Peebles: