Add a note about the inability to model FP -> int conversions which