Iterated Prisoner's Dilemma

The Iterated Prisoner's Dilemma.  The work of Robert Axelrod and William Poundstone fascinated us in graduate school as we tried to apply AI techniques to decision strategies.   I see that The Edge has a good piece that elaborates a conversation between them and other people involved in this field.  Relatively non-technical.  How powerful is this method to provide an evolving set of strategies between competitors? And how is this related to game theory based approaches?