I don't know enough about string theory, or for that matter cosmology in general, to comment about that part of parent post.
I do know enough about astrology to know that the scientific method could now be applied to its core concepts. It just won't be, because it is not real easy to do, and because the prejudice against astrology in the scientific community is very strong.
One approach that could be done now that was not possible until the last few decades would be to find some birth center that recorded times of birth in a consistent manner for several decades, take all the birth records for a few decades and reduce them to only date and time (thoroughly anonymous). Then compose a set of random dates and times of similar size with the same start and end points. Use software like Astrolog (that can be automated to work on large data sets) to generate horoscopes for each date/time in both data sets (using the birth center location for longitude and latitude). The random set is now representative of all possible births at the birth center, while the other is the subset of births that actually occurred. As anyone involved in obstetrics will tell you, births are not uniformly distributed over time: there will be periods of many births in a day or a week and times of many days with few if any births.
If there is anything to astrology, there will be significant statistical differences between the two sets of horoscopes. However these probably will not be blatantly obvious or they would have been noticed before now. There might, for instance, be more of a particular type of relationship between Sun, Moon, and Ascendant in the subset of birth charts than in the random charts. We don't care what the astrological significance of any findings are-- that might be the subject of further research. We are only interested in whether we find a significant difference between the moments when a real birth occurred and what the random sample of potential births shows.