Microsoft Word Kurzweil, Ray The Singularity Is Near doc

Our Simulation Is Turned Off.
Another existential risk that Bostrom and others have identified is that we're actually 
living in a simulation and the simulation will be shut down. It might appear that there's not a lot we could do to 
influence this. However, since we're the subject of the simulation, we do have the opportunity to shape what happens 
inside of it. The best way we could avoid being shut down would be to be interesting to the observers of the 
simulation. Assuming that someone is actually paying attention to the simulation, it's a fair assumption that it's less 
likely to be turned off when it's compelling than otherwise. 
We could spend a lot of time considering what it means for a simulation to be interesting, but the creation of new 
knowledge would be a critical part of this assessment. Although it may be difficult for us to conjecture what would be 
interesting to our hypothesized simulation observer, it would seem that the Singularity is likely to be about as 
absorbing as any development we could imagine and would create new knowledge at an extraordinary rate. Indeed, 

---

achieving a Singularity of exploding knowledge may be the very purpose of the simulation. Thus, assuring a 
"constructive" Singularity (one that avoids degenerate outcomes such as existential destruction by gray goo or 
dominance by a malicious AI) could be the best course to prevent the simulation from being terminated. Of course, we 
have every motivation to achieve a constructive Singularity for many other reasons. 
If the world we're living in is a simulation on someone's computer, it's a very good one—so detailed, in fact, that 
we may as well accept it as our reality. In any event, it is the only reality to which we have access. 
Our world appears to have a long and rich history. This means that either our world is not, in fact, a simulation or, 
if it is, the simulation has been going a very long time and thus is not likely to stop anytime soon. Of course it is also 
possible that the simulation includes evidence of a long history without the history's having actually occurred. 
As I discussed in chapter 6, there are conjectures that an advanced civilization may create a new universe to 
perform computation (or, to put it another way, to continue the expansion of its own computation). Our living in such a 
universe (created by another civilization) can be considered a simulation scenario. Perhaps this other civilization is 
running an evolutionary algorithm on our universe (that is, the evolution we're witnessing) to create an explosion of 
knowledge from a technology Singularity. If that is true, then the civilization watching our universe might shut down 
the simulation if it appeared that a knowledge Singularity had gone awry and it did not look like it was going to occur. 
This scenario is also not high on my worry list, particularly since the only strategy that we can follow to avoid a 
negative outcome is the one we need to follow anyway.