Reading papers is so crucial. It is a great way to get ideas and apply them. I was able to set up some really good tree based forecasting models based on a PhD dissertation I had read once.
Yeah I didn't get it until I read this. I've read a few papers before and some of them have been too abstract / unfamiliar for me to understand. I've had to read them slowly 3-4 times just to get it to 'click'. I think it takes practice to read papers, but once you have the habit, maybe you won't be able to stop.
After reading a paper, you might want to try getting a pen and paper then recall by writing what have you understood as if you are teaching to it others. If you can't recall what you've read then reread the paper to fill in the gap and test your understanding again. This method is called Feynman Technique.
Well...I somewhat cheat here. I ignore papers without any code or coded examples. I don't have the time or patience to implement something just from raw equations.
Off topic, but does it strike anyone else as strange that Quora is a platform to ask these types of questions (how come X is so good at Y)? After reading the headline, I knew immediately this was going to be on Quora. The questions often have some good content in response but they seem like they have some subjectivity baked in.