The Eigenstate Thermalization Hypothesis provides a central concept of quantum many-body dynamics: subsystems of generic quantum states are expected to thermalize according to their initial energy expectation  value when evolved with an ergodic Hamiltonian. Driven-systems are particularly doomed to end up in featureless infinite temperature
 maximum entropy states. However, thermalization can be delayed by disorder or high-driving frequencies. Then the system remains in a long-lived prethermal state, which opens a finite time window for genuine non-equilibrium phenomena such as discrete time-crystals. Dipolar coupled interacting spin systems provide a unique platform to study the formation, rigidity and melting of these profound non-equilibrium states of matter. In particular, new technological developments allow an experimental speedup and data clarity increase  of more than 5 orders of magnitude.