Text this: Towards General Evaluation of Intelligent Systems: Lessons Learned from Reproducing AIQ Test Results