I would suggest building something that you yourself find useful / solving a problem that you regularly face that doesn’t have a satisfactory off the shelf solution.

Note that this implies spending time in a different problem domain than just programming all day (alternately read as: have a hobby or two).

I finally have a side project with traction & paying users and it’s because sheer frustration drove me to improve an inefficient workflow in my language learning process (hobby), which turned out to not be such a unique problem at the end of the day.

(Not to shill - I was a frustrated Anki user and built this to improve my own life: https://ankiweb.net/shared/info/1531888719 )