I think that's probably part of it; but on a really simple level with YAGNI you're not expending effort on something that isn't needed which reduces cost.

What I try to do is think about the classes of functionality that might be needed in the future. How could I build X feature in a years time?

Leave doors open, not closed.