I've been building this for a few months, and it's been fun discovering what types of puzzles humans are good at, and terrible at - for example we're pretty good at finding the middle of something, and terrible at comparing geometric areas. It's remarkable sometimes how I'll prototype a new puzzle, and find it absolutely impossible to achieve a good result reliably even with practice, or find I'll be able to get bang-on 100% of the time - discovering the balance between the two has been quite a journey.