An Argument for Something Like Theism
I was recently challenged my a 12-y.o. graduate of Camp Quest to come up with a novel argument for theism, in the spirit of being able to argue off-brief. Challenge accepted, kid. Of course, there is nothing particularly novel here, but I’m almost sure that it is better than most of the old arguments I’ve heard in favor of classical theism, because I’m toying around with marginally less ridiculous ideas. That said, prepare for a number of counterintuitive (not to say absurd) ideas to be thrown at you all at once, it’s up to you to tell me exactly what I’m getting wrong here:
- There is nothing in the laws of physics that prevents a human mind (or something subjectively indistinguishable therefrom) from being simulated in software.
- There are no practical technical limitations which make it exceedingly difficult to simultaneously simulate large numbers of such minds interactively.
- There are countless possible simulated alternate versions of the years immediately preceding the singularity: A1, A2, …
- The post-singularity AI would be strongly interested in running a large number of such simulations.
- The post-singularity AI would be technically capable of running a large number of such simulations.
- There is in fact only one actual version of events (A0) that leads up to the singularity in the real world.
- You are either living through the events of A0 or one of the alternate simulations A1, A2, etc.
- If all of the above are all true, over the long run the number of simulated minds in alternate scenarios will significantly exceed the number of biological neural networks which were actually walking the Earth in the years immediately preceding the singularity. This makes it more likely that any given mind does not exist in A0 but rather in one of the alternate historical simulations.
And that’s about it. The keys questions are whether such superhuman AI will come into existence, whether our minds are software-simulable (those first two are interrelated) and whether the prevailing AI will have a strong interest in simulating the time period which we subjectively find ourselves experiencing right now, which is ultimately a matter for the programmers. (I’d hate to be in Luke Muehlhauser’s shoes right now.) Then again, that part may be unavoidable. Conscious beings in general seem preoccupied with would’ve/could’ve/should’ve hypothesizing, it’s just that most of us don’t have access to external software to play those scenarios out in arbitrarily fine detail.
But how is this an argument for theism, you ask? Well, if you and everyone you know are in fact part of an ongoing simulation, then there exists a being powerful enough to alter the laws of nature and even reward you with life beyond this life. Consider C. Scott Littleton’s globally broad definition of a deity, “[A] being with powers greater than those of ordinary humans, but who interacts with humans, positively or negatively, in ways that carry humans to new levels of consciousness beyond the grounded preoccupations of ordinary life.” The post-singularity AI may well have other commonly godlike characteristics, such as a lack of human empathy and indifference to human suffering.
I whipped this up in a hurry this morning, so I fully expect that I may have missed something important. Please tell me that I’ve gotten something wrong here.