I would posit that this is just a vision problem. Or at least a large part vision. If you presented these questions in something like JSON format, or maybe eve

I would posit that this is just a vision problem. Or at least a large part vision. If you presented these questions in something like JSON format, or maybe even ascii art I wonder how it would do? Driverless cars is the same thing. Actually driving with good info is easy. Getting good info from an imprecise world is hard