In high school I studied ancient Greek and Latin, and I was told I was "good". So when later on I transitioned to quantitative subjects, I adopted the same study method (I was reading math books the way I would read a novel: going through stuff from top to bottom, and only once). But this time I struggled! Was I suddenly a "bad" student? Probably not, however, I was using the wrong approach. It took me a while to figure out what was happening, and how to improve. Luckily, you can adapt your study right away just by reading Chapter 2 of Kevin Houston's "How to think like a Mathematician".

A 1 h video lecture on the functioning of LLMs from one of the greatest programmers of our time.

Coursera is an online platform where you can find several courses (like EdX), and audit them for free (NOTE: you only need to pay if you want to get a certificate, so look for the "audit" button in the fineprint). I recommend the Machine Learning courses 1/2/3 by Andrew NG (founder of Coursera). In 2021, the ML course was updated and all the material is now in Python (previously in Matlab) and exercises come on Jupiter Notebooks.

Courses and resources for deep learning, AI, and neural networks from Andrew Ng's team. Visit DeepLearning.AI.

Gentzkow & Shapiro 2014, Stanford

Essential guide on best practices for organizing code and data in empirical research projects. Read the full paper.

The Markup

Guide on legal considerations and methods for web scraping in the EU context. Read the full article.