It was a different era, DL field was a lot simpler. Until about 2015 I could read all important papers in all subfields, as soon as they were published. On the other hand, today there are so many great learning resources. For example, back in 2013 I could not find a simple explanation of how RNNs worked. I emailed Schmidhuber and asked him "where can I learn the basics of RNNs". He pointed me to Alex Graves PhD thesis. Today you have lots of code to play with (all important ideas have multiple implementations), libraries are mature and user friendly, everything just works.
My goal from the start was to learn enough to do research, to advance the state of the art, and that requires a lot of effort. What's your goal?