My new paper considers function interpolation via infinite-dimensional compressed sensing with weighted l1 minimization. In it I derive new recovery guarantees which are provably sharp in a number of important cases, and explain the practical usefulness of including weights in the optimization.