Nutrition was invented in the early 19th century as a biochemical science that reduces foods into significant chemical constituents. Ever since then, the teaching and practice of nutrition has been based on this conceptual framework, or paradigm. The examples given here are dietary guidelines and other food guides. The first guides issued up to the middle of the last century were designed to help prevent nutrient deficiencies, promote growth, and ensure plentiful diets. These recommended foods then thought to contain adequate proteins, fats, carbohydrates, vitamins, minerals and trace elements, as well as dietary energy. At a time of accelerating industrial production of food, they were generally effective. Within the second half of the century, guides were developed and changed to counter the rapid rise in heart disease in the USA, the UK, and other high-income countries. These recommended less foods of all types high in fat, saturated fat, cholesterol and sodium, more 'complex carbohydrates', and fruit and vegetables rich in microconstituents. They probably had some limited effect. In this century and now, dominant guides have been changed again in attempts to counter what has become pandemic obesity and diabetes. These recommend less food high in saturated fat, sugar and sodium, with less emphasis on total fat and more on sugar. They are not effective. All these guides are derived from and governed by the biochemical paradigm of nutrition science. This was once useful, but now should be discarded as obsolete except for addressing deficiencies. Here, a new paradigm is proposed.