The idea of reducibility is really just a matter of using a solution to one problem as a subroutine in solving another.
We say problem Y is (polynomial) reducible to X if we can solve Y using a polynomial number of calls to an algorithm for X.
This implies that Y is as easy as X (X is at least as hard as Y).
We used this idea when we talked about least squares and solving linear equations: we can reduce least squares to solving linear equations very directly.
