When starting to develop deep learning modules a lot of time will be wasted on managing resources. You need to setup a machine with the proper frameworks and the proper hardware, wasting many patient cycles when you just want to kickstart and check and idea as soon as you can. Today we have several ways to focus on your idea without wasting your time in server setup and management, Google AutoML allows you to take your data and get results immediately, Amazon Sagemaker (and Google CloudML) get take your module and run it without worrying about server and platforms.