An overview of data accumulated on defects in CIGS is presented. From this the following conclusions are drawn. ( 1) The primary defects, which accommodate changes in film composition in group III rich p-type CIGS, are normally found in large superclusters on (001) planes. This is due to electrostatic interactions of the individual defect cluster dipoles. Such superclustering explains why the defects are electrically inactive, why the hole mobility does not depend upon defect concentration or composition, and why the hole concentration in CIGS depends so little on composition. One form of these superclusters may explain the CuPt ordering observed in some cases. (?) As group III rich CIGS becomes n-type (the Fermi energy rises above midgap), either by the effect of electric fields in the material or continuing addition of group III elements beyond a critical level (which depends upon Ga content and process conditions), the clusters partially or fully decompose, pinning the Fermi energy at the In-Cu(2+) defect level. This is why sufficiently group III rich material remains only weakly n-type over a wide composition range. It may also mean that the Fermi energy is pinned at this level near the surface of some devices. (3) Na acts to improve CIGS devices at grain boundaries. This may be by increasing atomic mobility there, by electrical passivation of the grain boundaries, through increasing grain size, or through some other effect. (4) Oxygen has no beneficial effect on devices through its interaction with the CIGS either within the grains or at grain boundaries.