Technical debt refers to suboptimal choices during software development that achieve short-term goals at the expense of long-term quality. Although often discussed informally, technical debt has not crystalized into a consistently applied type and remains invisible when describing issues in issue trackers that are used for coordinating work among software developers. In this study, we manually labeled references to technical debt for 1,934 tickets in the Chromium issue tracker. We used these labels to train a classifier to estimate labels for an additional 475,000 tickets. Our classifier significantly outperforms key phrase search, and we conclude that discussion of technical debt appears in about 16% of the tracked Chromium issues. This application of machine learning to locate technical debt can not only make it visible but also help in its timely resolution.