Interesting to read this, because I thought I was the only person like this. I typically work nearly constantly with my downtime consisting of reading new AI/machine learning papers, writing code to better understand algorithms, and so on. But I find I can only do this for a few months at a time and I have to fall back to doing something else after my regular work time. I'm currently back to doing woodworking, but at other times it might be playing guitar, ukulele, dulcimer, harmonica, or recorder. Those deadlines are tough though. I've got two presentations coming up and a homework that I needed to create a month ago. If it helps, one of our employee health support staff told me that he was probably the only person at our university who could tell someone to stop working and go do something else that he or she enjoyed. According to him, the benefits on doing this are very high.

BTW, your headline showing "Correlations reported as "science" in the media" is related to part of what I've been spending so much time reading about. The more I learn about Judea Pearl's (and Robbins and Hernan's) Causal Inference work, the more I realize how much we need to take many correlational studies with a grain of salt. Unfortunately, their work also shows that many RCTs suffer from causal inference problems as well due to overadjustment bias.

Proud member of #powerhousecreatives