Isn't this a bit like saying you just created an nth-order polynomial regresion for any set of points x/y that someone hands you? They are essentially the same

Isn't this a bit like saying you just created an nth-order polynomial regresion for any set of points x/y that someone hands you? They are essentially the same amount of parameters, and both can only approximate one function. So.... no kidding?