> Asking me to convert it to `float` myself is annoying.

Well, it's just a call to 'float', right?  On the other side, we're looking at significant extra code to implement Decimal * timedelta, so there needs to be a good reason to add it.  (And there's feature-creep involved, too;  once we've got Decimal * timedelta, there will probably also be requests for timedelta / Decimal, and possibly timedelta / timedelta -> Decimal too).