And there are over 400 million iOS devices. If you think you can get a large enough percentage of iPad 1 users to make it worth you while then develop for it. I'd much rather spend my time ensuring it works well on 300+ million other devices.
I smell some false dichotomy here. I don't know much about iOS development, but it sounds baffling to hear that iOS 6 is so different that it would take "hundred of hours" just to make sure the app is usable on the previous release.
As always, depends what you're doing, but this is not a crazy statement.
There are certainly fewer bombshells in iOS6 than iOS 5, but it's no stretch to say that adopting iOS 5 early on was worth 100s of hours -- it had huge improvements for dev.
Also keep in mind that the fewer OS versions you target the less testing you have to do.
