Abstract: Relativistic dissipative hydrodynamics has been quite successful in describing experimental results from ultrarelativistic AA, pA, and even high-multiplicity pp collisions.  This is despite the fact that there can be large deviations from local thermal equilibrium generated in such collisions (large Knudsen and inverse Reynolds numbers).  I will discuss this unreasonable success of hydrodynamics in the context of dynamical attractors and, using QCD kinetic theory, present evidence that an early-time dynamical attractor exists that connects smoothly onto the hydrodynamic late-time attractor.  I will also discuss exact solutions obtained in the relaxation time approximation that can be used to assess the accuracy of the different non-equilibrium hydrodynamic frameworks on the market.