,
5 tweets,
1 min read
Read on Twitter
1/ I think people worry too much *what* to study instead of *how* to study. This applies to topics like programming languages, machine learning, javascript frameworks, etc. Let's take machine learning as an example.
2/ When reading a paper/book you can go breadth-first or depth-first. You can skim it and skip the parts you don't understand and get the high level idea (breadth). Or you can analyze every sentence and formula, follow the references, and re-implement it in code (depth).
3/ If you follow the breadth-first approach you are likely to stress out about "what to read next". And because you're skipping the parts you don't understand it'll take you a long time to learn.
4/ The counterintuitive thing: If you go depth-first you will by *by necessity* also get the breath. References and question will take you to wildly different material. Eventually, all Deep Learning papers will trace a path back to LSTMs or backprop ;)
5/ That's why it really doesn't matter much *where* you start. It's much more important to be curious, don't be afraid of the things you don't understand, and go deep into something. The breadth will take care of itself over time.
