Some of the stranger things I've come across when non technical people have tried to estimate time-scales to develop something.

* This feature will not take long because <giant technology corp> has already developed something similar.

* This feature will not take long because it will not be used frequently (or because the 100s of edge cases will not be encountered regularly).

* This feature will not take long because it will only be one button click in the UI.

* This feature will not take long because it integrates with a lot of existing systems.

And of course:

* This feature will not take long because I need it tomorrow.