The pacemaker example doesn't resonate with me. Its a very well-posed problem - don't let the patient die. You don't have to be user to understand how to make things better. For things like personal computing its a bit different. Problems are very ill-posed and often the question is what to make. In the second category its better to start with a map of the future and refine it with user feedback. However often users don't know what they want and then you need folks like Jobs.