Re: Why are robots always immortal?
It makes perfect sense to assume that the lifespan of any given machine will be different from the lifespan of a human being. Of course in almost all cases that lifespan is shorter than a human's, but then we haven't yet built machines with the capability of self-preservation. Many machines of the last century have, with proper TLC, remained quite functional using entirely non-rubber technology. If one posits a milieu in which robots have the will and resources to repair themselves for hundreds of years, I can't see a pressing reason to assume they won't do so.
I grant you, though, that assuming a robot can lie dormant for thousands of years and then work fine is not a very reasonable default assumption. Then again robot longevity is generally used as a plot device, which is also not unreasonable, as it's simple enough to extrapolate current machine lifespans into future technologies.
As for transferring the mind between hosts, I can readily imagine the technology for transplanting a living human brain into another body. With concerted effort, I expect we could develop the first steps of that technology within a decade. How much more straightforward, then, to imagine transplanting a machine's brain?