Short update today. Finishing up the Kaggle tutorial (overfitting / underfitting, picking the right parameters using validation scores and "for" loops or dict comprehensions). I want to move on to fast.ai as fast as possible, which seems to have a better coverage of ML from a programming standpoint. But still, it's a great refresher from the LSE course Ben Lauderdale was giving back in 2015. Also liking the notebook format!

On notebooks: Wish that's how economics papers were written. Jupyter notebook > a pdf with no code whatsoever. Starts with having econ departments having this as a standard for classes.

Meta-update: Might aim for 6 hours a day on average per week instead of a strict 6 hours per day. Easier to manage time this way...