1. Computing and engineering in general is such a discipine where one nitpicky detail can make all the difference. One bit flip, one keystone, one badly selected material. However, the conduct of negotiation, could be improved as always, but it's a general listening problem most of the time.
2. From the dickheads point or view, everything can be solved by spending more money on HW. Each unexpected spend like this is a risk to open the floodgates and lose both financial and engineering discipline by setting a bad precedent. Not sure you want easy handed decisions there as well.