@Samuel:
It is like Nathan wrote: In a closed system entropy can only increase. It cannot shrink.
That is what Geoff means: If you have not input from outside (closed system), the working ability of the energy in your system declines (or at most keeps the same) with every energy transformation happening.
But it is practically impossible to really have a closed system.
Even the space-capsule from my example above will radiate and absorb electromagnetic waves.
If it would be perfectly isolated, all the energy from the rubber balls would turn to heat (via friction) over the time. The temperature of the capsule would slightly increase until the balls have stopped moving. As all of the capsule would have the same temperature, the heat energy stored in it would not have any working ability within the system anymore. You could call this a boundary value of entropy.
In reality you still have stored a giant amount of usable energy in your system, in the matter of the system itself. Since Einstein we know: Energy is equivalent to mass (E=mc2). The explosion of the Hiroshima-bomb only converted about 1 g of matter into energy.
You could say the universe has a certain mass. Therefore it contains a limited amount of energy. Therefore entropy of the whole thing should come to some boundary value in a very far future, when all mass is transformed and/or the system cooled down to absolute zero (=head death theory
http://en.wikipedia.org/wiki/Heat_death_of_the_universe )
But I do not think we (I) know enough about our universe to give a certain answer to such questions.
If we open our little space-capsule and add the heat contained in it to some colder system, we of course gain working ability again. But this cascade is limited by absolute zero (Temperature of 0 K = -273,15°C).