Choose the right sized steps. Like for a new programming language: Tutorial > first example project > first meaningful addition to something existing > a small

Choose the right sized steps. Like for a new programming language: Tutorial > first example project > first meaningful addition to something existing > a small valuable first project > a project where you initially are a bit out of your depth but figure it out after reading the docs and asking around Choose the steepest slope you can manage without giving up. This means sometimes giving up a project because you misjudged the slope, but that should be the exception. I'm currently diving in to deep learning, I'm not sure yet if I'll make it. But I got an Idea and I already found some baby-steps that will be a lot of work as a noob, but should be doable for me. My biggest challenge is the width of the topic here, ranging from data acquisition, over math, actual deep learning, deployment to a production environment, web stuff for a front-end and so on.