I don't envy you at all. I'd ship a working prototype and say 'like this one' :) My ability to spec this would surely become an exercise in finding all of the ways in which I could mess it up.

Learning at industrial scale is sometimes temporally cheaper because it allows you to conclusively exhaust a design path through efficient and spectacularly memorable failure. ;)

HN should have a 'comment of the day' section. I'd nominate this one.