Software people are comparatively terrible at schedules, schedule fidelity, long term planning (in terms of what needs to be considered now to satisfy anticipated requirements later), cost management, cost over time modeling, and day one quality, among other things.

Well, having worked at a hardware company I think you over estimate how accurate hardware schedules are. And day one quality is more about how you define day one I think. (Some types of hardware) you can release A01 to "internal customers" but the broader public's experience is over a year and several spins later.

No disagreement on the long term planning aspect though.

With hardware the lead times are longer and there are no updates after delivery. Thus, a waterfall approach is more suitable. Customers can actually provide useful and comprehensive requirements.