Well done to the people behind the Alpha Magnetic Spectrometer (AMS) on the Space Station that has just reported some results. It is always good to collect data in new regimes. Whatever they find, they will find something new. They do not seem to have anything significant to say about where their extra positrons are coming from though.
The AMS was designed to search for dark matter: a terrible theory in my view, because it is not falsifiable and it looks to me like an attempt to support a theory (general relativity), that was devised before we knew about galaxy rotations, by adding invisible matter ad hoc to galaxies to force the galactic mass and rotation data to fit general relativity (dark matter is so flexible that you can fit many models to the data). There was a similar argument in Galileo's time. Aristotle had said heavier objects fall faster, but Galileo noticed in a hailstorm that big and small hailstones fell together and later on he showed, using balls and inclined planes, why this must be. His contemporaries, to play safe and support Aristotle, said "Aha! The big ones must have fallen from a greater height". This could have been right, but the clue was in the ad hoc way they had to set up the initial height of the various sizes of hail: there was no reason for it. There is also no reason why dark matter should be in a diffuse halo around a galaxy.
If dark matter is too vague to be falsified, all one can do is suggest less arbitrary alternatives.
Milgrom's empirical MoND (Modified Newtonian Dynamics) is a little less arbitrary. MoNDians suggest that we tweak either gravity or inertia at low accelerations (so there are two varieties). They do not have a specific mechanism in mind, but at least they have less wriggle room than dark matter. MoND has a formula (well, a few choices), but it does still have one fitting parameter: a0. By varying a0 MoND can be made to fit galaxy rotation curves very well, but is a bit off with clusters.
Then there is MiHsC, which has a very specific physical model of inertia. MiHsC predicts galaxy rotations and galaxy cluster dynamics within the uncertainty in the observations, with comparable accuracy to MoND, and without any adjustable parameters at all, see: http://arxiv.org/abs/1207.7007 (Journal: Astro. & Space Sci., 342, 2, 575-578). See also the figure in my blog entry here.
An analogy: You go into a shoe shop. The owner proudly produces his revered Einstein Shoe (nice curves, no sock needed). He measures your foot, concludes with regret that your foot is too big for the shoe and suggests chopping most of your foot off. That's the dark matter way. A MoND shoe would fit with some adjustment at the cobblers. A MiHsC shoe would fit with no adjustments, but it's an unknown brand..