Here is the perceived problem with machine self-reproduction.
The machine must contain a blueprint of itself, a set of instructions on how to copy itself.
But if the copy is to be able to reproduce itself, the copy will need its own set of instructions.
That is, the original machine needs to make a copy of the itself, complete with blueprint.
This means the blueprint must contain a copy of the blueprint, which must in turn contain a copy of the blueprint, and so on. We seem to encounter an infinte regress.
|
Why isn't this a good objection?
Return to The Paradox of Self-Replicating Machines.