Actually, not quite. Shuttle guys are scavenging 8088s and 80286s for maintenance of the ground support equipment. Your alternative raises the challenge of trying to keep all of the instrumentation up to date and consistent with the older H/W on the boosters. It's far cheaper to find and replace old Intel chips than it is to continuously rebuild, re-test, and re-qualify new ground support equipment.
So, in affect, a long term order of disposables could result in reduced cost per ton over existing launches, not only because of improved capacity utilization of manufacturing facilities, but also because of the judicious use of improving lower cost technology.
Probably correct. The Russians demonstrated that economies of scale are real when they launched something like 1500 Soyuz boosters over the span of about 30 years -- that's something like 6 launches per month!
But that also points out the problem: there's simply no market right now for anything close to that sort of launch rate. These days the Atlas and Delta guys are lucky if they get 6 launches per year. You just can't get economies of scale when your launch rate is so low.
There's a glut of fiber optic capacity. There's much less need for geostationary communications satellites.