Without plasma turbulence, we would already have self-sustaining magnetic fusion machines. The turbulent convective loss of heat dominates plasma confinement and thus dictates the required scale of fusion devices. Calculating this multi-scale turbulence, which spans large ranges of scale in five dimensions of phase space, is a monumental achievement that has only recently become possible—just within the last decade. I will describe the multi-scale approach that enabled this progress and outline the critical issues that still need to be resolved.