I recall writing "programs" that performed some of the calcs I needed in the physics curriculum. I also recall being amused that I could calculate 1/3 (or 1/7) and add it to itself until I got (nearly) 1. From there this led me to study what fixed and floating point were, and think about the implications for the code I was writing (precision, truncation/roundoff error, etc.) to do longer calcs.
This eventually led me to skipping a postdoc, and working in the high performance computing industry. Where I am today.
They were/are awesome tools.