This problem is more historical, showing a particular function where a slight change in initial value produces considerably different approximations. This same example was part of my first talk back in 1992.