You shouldn't retire a system until you have something ready to replace it...
...unless the said system was mostly useless in the first place.
Remote controlled vehicles have done a lot of research in the recent years. I don't see the point of sending a man to pick up stones from the surface of the Mars when a robot can do the same with a fraction of the cost and risk. If there were a way to colonize Mars, that would be different, but at the moment the remote controlled missions seem to be a lot more effective.
Upper atmosphere missions then, I don't know. Perhaps fixing satellites could be something that cannot be done with a robot, and that might be a reason to develop a new shuttle. But my understanding of this is that it is a lot more cheaper to build a new satellite and let the old one burn instead of fixing it - and there is only one case where this has been done, that being the Hubble telescope. The telescope was defective from the start though, and that should have been noticed in the testing
before it was on the orbit.