Jon,
That is a great question! One part you partially answered already, it's fun! The go kart I'm testing is smaller/light weight compared to a mower, so it's easier to work on, transport and maintain (no gas motor!). Also easier to take to STEM events because I don't have to worry about oil or gasoline (no gas motor! catch a trend yet?). Just because it's stupid fast, doesn't mean I cannot drive it slow (right?). And I've always wanted an rc/autonomous go kart, so I put it earlier into the plans. I would think if I can get autonomous working in a fast (or slow) go kart it should be much easier to get the mower working? I will probably keep the go kart just RC until I'm comfortable with my new autonomous codebase whatever it ends up being (see below). ROS based or all my own code?
I agree about using RC car for testing mower/go kart code, that is something I'm also planning on doing. For now I'm going to use the push mower as another option for testing my autonomous code for my mower project. The push mower will have differential steering, and a top speed of maybe 3 mph. (Which is still fast for a small autonomous mower. ;) )
My projects are often more on the fun side, but I have been adjusting my priorities recently to be more practical and look for real applications of my work. (Aka make money off of them). Still trying to find a "niche" for my RC Tank project, or take comercial which is why there are not as many details on it.
So in summary go kart is mostly for fun right now. The push mower/rc car is where I'll be putting my real efforts into for autonomous testing and development. Heck, if I can get push mower working it will serve a practical purpose to mow my yard which gives me real world testing and a mowed yard (BONUS!).
JesseJay