This document discusses automated unit test generation and the costs of software bugs. It provides statistics on the annual costs of software bugs in the US ($59.9B) and worldwide ($312B). It then covers different techniques for automated unit test generation including random data generation, search-based approaches, and symbolic/concolic-based approaches. Examples of specific tools that use each technique like EvoSuite, PEX, and jCUTE are also provided. The document concludes with best practices for using automatically generated unit tests and a question about automatically generating GUI/API tests.