[giantrobot]

The word "just" is doing some pretty heavy lifting in that statement.

It pre-supposes you can event build a bunch of invincible perfectly programmed micro/nanobots in the first place. Then you can identify a target body in a solar system a long ways away. Then predict that system's position with an accuracy your probe could land on it after a trip of hundreds of light years (or launch a probe with a bunch of fuel to do terminal maneuvering). Then those probes would function after millennia of dormancy.

It's all possible. But the concept shouldn't be treated as inevitable.

It also runs right up against panspermia theories. What's the practical difference between a single celled organism and a self-replicating nanobot?

[roywiggins]

Well, yeah. It's really hard to do, so maybe nobody does it, or it's just impossible. But with several billions of years of galactic history it has to be pretty hard if civilizations are common.

Really it's just one example of a technosignature that is longer-lasting than radio waves, that extends the amount of detectable overlap we might expect to have with an older civilization, and the lack of observation either says that sort of technology is either not feasible or there aren't many civilizations in the galactic past or present.