Neurotechnology promises a way to repair the damaged nervous system that requires a merger of neuroscience, engineering and clinical knowledge. Brain Computer Interfaces can now read out the intention to move through a tiny brain interface connected to a computer. Algorithms, based on our understanding of brain activity can decode patterns of neural activity into useful action commands. These commands can already operate computers for communication, robots to act as a surrogate arm, or even reanimate a paralyzed limb, allowing people to type or reach and grasp with their own arm and hand. Decoding is possible because of years of inquiry into the way non-human primate brains encode complex movements, and sensors can function years in the brain because of neuroengineering advances made in animal research. Despite these advances control by people with paralysis is slow and less dexterous than in a typical person, because we lack an adequate 'theory of brain function' and optimal electrodes for sampling neural activity. BCIs are not yet capable of home use because we need to create novel implantable, miniaturized processors that can communicate large amounts of information at high speed from the brain to the body or machines. But these obstacles are being overcome as well. As these advances emerge, we are moving towards reaching a vision where a person paralyzed from stroke, spinal cord or brain injury, or neurodegenerative diseases like ALS, will be able to resume everyday, independent life.