This book, originally published in 1950, is a compilation of a bunch of short stories written by Asimov in the 1940's. They are loosely tied together with the simple tool of one character interviewing another who then tells us the stories. The character being interviewed is a long-time robopsychologist who has worked for the company that created robots for fifty years. She tells stories of the development of robots from the first very basic robots that couldn't even talk to the "latest" super robots that were designed by other robots that were designed by other robots and so on so that now no human can even begin to understand how they work. These stories explore ideas of what could possibly go wrong with the robots. What if you made a robot and then the robot refused to believe that you made it? What if, in the heat of anger, you told a robot to "go lose itself," and it did? What if you needed a robot to help you in a dangerous situation when the robot has been programmed to not let any harm come to a person so won't let you near that dangerous situation?