methyl group wont shift because you'd sacrifice a tertiary carbocation to form a secondary carbocation. hydride shifts so that you trade a tertiary carbocation for a tertiary carbocation. in the previous post, the shift was restricted by cyclic constraint (the carbocation and shifting groups were within a 6-membered ring), and you need overlap of the sigma bond with the empty p-orbital to get shift to occur (same with last week's C-Si bond problem). In this case, free rotation about the C-C single bond allows for both the methyl group or the hydride to have good overlap for shifting.
The reason I didn't have N attack straight away was the disfavored formation of a 4-membered ring, and I liked that I could make a 3- or a 5-membered ring after shifting the carbocation. There are typically 3(/4) things that happen after a prins. Attack by water (with elimination to allylic alcohol sometimes), attack by a nucleophile (like N), or coordination of another aldehyde to form a dioxane.