Reinforced random walks are to ordinary random walks what sampling from a Polya urn is to sampling from an infinite population. The first reinforced random walk, introduced by Diaconnis around 1987, can in fact be realized as coupled Polya urns. This will be primarily a survey talk.