... fairly simple repetitive actions. Cranes perform simple repetitive actions, tractors ploughing do to. Pattern recognition, creative thinking, problem solving is still done better by people, and I've grown up being told that we'll have true AI 'in five to ten yers', so I'm not holding my breath. Humans are expensive in space, yes, but robots are not proven to be able to do the same quality of job as a human at all. AI just isn't there, teleoperation is unproven, virtual environments indistiguishable from actual presence (which is whats needed to bring the quality of work up to that of actual presence) are rare and expensive, and the whole operation is limited by lag time. Even a three second delay means hideously slow reactions if something goes wrong.

Robots are cheap, expendable, and very usefull, but a human on mars could have done a lot more than then rovers have, and on the moon we have the option of sending humans there. For science observations robots have long proven their worth. For simple repetitive construction machines have proven their worth. But as far as skilled construction work goes they still lag behind us, and where available a human still does a better job.

The vision you sketch out sounds wonderfull to me, but it's based upon the unproven technology of in space, off world, teleoperated construction. I don't question your thinking, I want to see proof that thinking translates into reality in space.