Two-particle correlations yield detailed space-time information describing how particles are emitted from a heavy ion collision, a method often referred to as Hanbury-Brown Twiss (HBT) interferometry. The characteristic sizes inferred from the analysis of data from the Relativistic Heavy Ion Collider (RHIC) were initially noticeably smaller than predictions from the most sophisticated hydrodynamic models applied to modeling collisions at RHIC. This was coined the HBT puzzle, and remained unresolved for the past seven years. In this talk, I will explain how the puzzle was caused by several shortcomings of the transport models, none of which seemed all that important in isolation, but conspired to push the predictions away from the data in the same direction. The principal culprits were the lack of pre-equilibrium flow, ignoring viscosity, and using an unrealistically soft equation of state. Since the models had been relatively successful in reproducing spectra and elliptic flow variables, I will discuss whether these improvements might destroy previous agreement of hydrodynamic models with other observables.