It is well known that the plastic, or permanent, deformation of a metal is caused by the movement of curve-like defects in its crystal lattice. These defects are called dislocations. What is not yet clear is how to use this microscale information to make theoretical predictions at the continuum scale. Motivated by this, M.G. Mora, M. Peletier and I considered a system of interacting dislocations and studied the convergence of the evolution of the corresponding empirical measures in the limit of many dislocations. In the continuum limit we obtained an evolution law for the dislocation density. In this talk I will present this result and discuss its limitations and further extensions towards more realistic and complex systems.