BD is only focusing on the platform. Google is working on the deep learning aspect of finding out what to do / action and how to do it. The control theory rem

BD is only focusing on the platform. Google is working on the deep learning aspect of finding out what to do / action and how to do it. The control theory remark above is true, however even with deep learning, you can start your estimate of the parameters with the current control state, which reduces training time significantly.