I'm with you there. However, building a mostly automated support system for array of diverse services is something we (the present-day civilization) have littl

I'm with you there. However, building a mostly automated support system for array of diverse services is something we (the present-day civilization) have little experience with. This painful experiment /may/ end up accumulating enough experience to make cheap, mostly-automatic support possible in a few years. As opposed to current practice of throwing countless callcenter `drones' (people forced to adhere to ready-made scripts) onto the problem. Should the experience gathered become public knowledge, the need for tech-support personnel would be drastically lowered. In practice significantly lowering barriers for start-ups to serving large userbases.