We exploit information in a new dataset of monthly patent applications to construct an instrumental variable for the identification of technology news shocks that relaxes all the identifying assumptions traditionally used in the literature. Our sole requirement is that no structural disturbances affect the US economy via our instrument except for contemporaneous technology news. The instrument recovers news shocks that have no appreciable effect on aggregate productivity on impact, but are a significant driver of its trend component. News shocks prompt a sustained business cycle expansion in anticipation of the future increase in TFP, and are responsible for a sizeable share of economic fluctuations at business cycle frequencies. The stock market prices-in news shocks on impact, but consumer expectations take sensibly longer to adjust, consistent with the predictions of models of information frictions.