Despite years of work on robot locomotion, we still do not have robots that can reliably and flexibly move around in homes, workspaces, and natural terrain. For many of these environments, legged robots, as opposed to wheel-based robots, appear to be the most viable option for achieving the desired level of locomotion autonomy. This talk will present recent advances by the Dynamic Robot Laboratory at Oregon State University that brings us much closer to achieving reliable bipedal robot locomotion in natural environments. These advances are a result of combining carefully engineered "animal-like" robot legs with machine learning for teaching the robot to control the legs. This novel approach has allowed for the Cassie robot, developed by Agility Robotics, to reliably stand, walk, run, hop, skip, traverse stairs, and handle novel disturbances in the environment. Most recently this learning approach allowed Cassie to be the first robot to successfully complete a 5K. The talk will cover the journey to these recent results and highlight important general principles that can apply to other engineering control problems.