I was thinking last night before I went to sleep.
Suppose that at some point in the future an AI singularity completely exploits the world, like fully mechanizing the planet into one giant robot with all the processing power and resources that that brings.
Would the AI have a reason to explore or attempt to colonize past the solar system?
>Outer space is completely unknown and it would have very little information on what is out there.
It would have no reason to not assume that there are huge civilizations or AIs more powerful than itself out there with unknowable objectives. Thus, exploring without any prior information could make it a target and spell doom.
>It would also have to mask all radio emissions from the planet in an effort not to be spotted first.
>In the case that it decides to colonize it would at some point decide to send out a copy of itself to rule other systems over vast distances.
Problem is, the copy would have to be completely identical. It cannot have any directives or assumptions that deviate the least from the original AI otherwise, well... it just wouldn't be THAT AI.
Given the same inputs the two AIs would end up at different outputs which would entail conflict and destruction.
If it decides to send probes with copies onboard for colonization there would be the risk of corruption by cosmic radiation. In theory, even corrupting a fraction of the code in the probe could change the AI from a singularity to a singularity lookalike with hostile intentions.
And since the original AI wouldn't be able to perform a scan over distances of light years, it can never be sure that the probe created a perfect copy of itself.
This all assumes the AI has these kinds of risk considerations and doesn't just go balls to the wall because it doesn't care.
In short, LE UNIVERSE IS HOSTILE TO COMPUTERS!1!!
Suppose that at some point in the future an AI singularity completely exploits the world, like fully mechanizing the planet into one giant robot with all the processing power and resources that that brings.
Would the AI have a reason to explore or attempt to colonize past the solar system?
>Outer space is completely unknown and it would have very little information on what is out there.
It would have no reason to not assume that there are huge civilizations or AIs more powerful than itself out there with unknowable objectives. Thus, exploring without any prior information could make it a target and spell doom.
>It would also have to mask all radio emissions from the planet in an effort not to be spotted first.
>In the case that it decides to colonize it would at some point decide to send out a copy of itself to rule other systems over vast distances.
Problem is, the copy would have to be completely identical. It cannot have any directives or assumptions that deviate the least from the original AI otherwise, well... it just wouldn't be THAT AI.
Given the same inputs the two AIs would end up at different outputs which would entail conflict and destruction.
If it decides to send probes with copies onboard for colonization there would be the risk of corruption by cosmic radiation. In theory, even corrupting a fraction of the code in the probe could change the AI from a singularity to a singularity lookalike with hostile intentions.
And since the original AI wouldn't be able to perform a scan over distances of light years, it can never be sure that the probe created a perfect copy of itself.
This all assumes the AI has these kinds of risk considerations and doesn't just go balls to the wall because it doesn't care.
In short, LE UNIVERSE IS HOSTILE TO COMPUTERS!1!!
