IBM isn't investing billions of dollars every year into research and development - and winning more patents than our top 10 competitors combined for more than a decade - as an academic exercise. But research is now being driven much more by what people need rather than just by what is possible