Ever since William Gilbert's 16th century treatise "De Magnete," it has been known that magnetic fields can be used to manipulate the magnetic orientation of ferromagnets. This has been the foundation for electronic applications of magnetic materials for the past 100 years, in the form of, for example, inductors, microwave isolators, and magnetic disk drives. But in 1996, our understanding of how to control magnetization was fundamentally altered when John Slonczewski1 and Luc Berger2 independently predicted that spin-polarized current in magnetic conductors can affect their magnetization, due to a purely quantum mechanical transfer of angular momentum from the charge carriers to the magnetization. To observe this effect, current densities on the order of 108 amperes per square centimeter are required to generated sufficient torque to overcome the intrinsic viscous damping (the magnetic analogue of mechanical friction.) It was not until researchers started to investigate electron transport properties of conducting magnetic heterostructures in the deep nanometer regime that there was a chance of seeing such a phenomenon. This effect, alternately known in the literature as "spin torque" and "spin transfer," has been the subject of feverish study since 1998-1999, when the first experimental confirmations of the theories of Slonczewski and Berger appeared3,4. Now, two new works in this issue, by Pribiag and colleagues5 (page 498) and by Boulle and colleagues6 (page 492), show promising approaches whereby spin torque may find its way into commercial applications.