[syockit]

It's really unintuitive how a mere increase in number of dimensions could totally change the scale of things. This is the crux of the curse of dimensionality in machine learning. If you simply stacked it high, it may almost reach the sun (but no quite), but once you lay it into roughly a square (20000 x 50000), it spans only over a mile per side. To lay it roughly into a cube of 100 dollar stacks, you'd have something like 65 x 150 x 1026, which turns out to be about only 12 yards per side.

EDIT: I screwed up in the 1D calculation. A 10 million-height stack of 100 100$ bills only reaches 67 miles.