When we learn a new skill, sensory feedback from previous errors are incorporated into the feedforward motor command to gradually refine our movements. I will describe a new model of motor learning based on the brain simultaneously optimizing stability, accuracy and efficiency. This model of motor learning offers new insights as to how the brain controls the complex musculoskeletal system and iteratively adjusts motor commands to improve motor skills with practice. However learning can also be used to adjust the feedback control. I will discuss two experiments in which we examined adaptation of a visuomotor reflex both to the statistics of perturbations and during learning to perturbations to dynamics of the limb. We show modulation of reflex magnitude in the absence of a fixed change in the environment, and show that reflexes are sensitive to the statistics of tasks with modulation depending on whether the variability is task relevant or task irrelevant. In addition we show transitory changes in visuomotor reflexes during dynamic learning.