How often do we have the designs we want, versus the designs our algorithms know how to make? For many years, Marti Hearst (Professor, School of Information and Department of Electrical Engineering and Computer Sciences, University of California, Berkeley) has thought about how to build user interfaces that make use of the output of natural language processing. Language usually does not reveal itself readily to automation; it requires a great deal of care to produce results that are worthy of being viewed by people. In the spirit of reassessing the role of algorithms and AI in society, this talk reviewed past and current approaches to using natural language processing algorithms in real designs that affect people in real ways, with a focus on the visualization technique known as word clouds. This lecture was sponsored by the Visualization for Understanding and Exploration (VUE) project, a partnership between the Neubauer Collegium and the Office of Research and National Laboratories.