I'm guessing that if HSI declines in an exponential fashion, then perhaps a large portion of the decline has already happened when the hops get to us. Nevermind that the %AA on the package represents some statistical average from whatever lot of hops the package was taken from. And nevermind any error in the formulas used to calculate isomerization in the boil. In other words, there are lots of sources of error in the IBU numbers that show up on the recipe so I wouldn't sweat it too much.
I brew with homegrown hops using my best (researched) guess of %AA and it works.
I'm actually pretty confident in the number that goes on the package. Here is why: The second stage in the pelletizing process is mixing the hop particles in very large drums in order to homogenize. Also most of the time hops are pelletized from a single lot, not blended. Blending is done when a large brewer requests it. Then each run is sampled throughout the run. The lab gets small ~3 ounce packages of these samples. We test each sample twice. The two need to be within 3% of each other or the result gets tossed. Then, it is true, all of the samples for the run get averaged, but the standard deviation in AA between samples in a given run vary only by ~.02%.
That is not much of a difference. However losing 1-2 percentage points in 3 months off of a 15.0% bittering hop (i.e. CTZ) is a significant drop. If the rate of decline is constant, by July/August 2016 the 15.0% hop is now 7-11%!
I know that not all varieties lose AA at the same rate, and other factors come in to play but...
I did look at the MoreBeer link. To be honest it went over my head, but I am considering making a spreadsheet with all the variables factored in, in order to make a reasonable prediction of AA losses.