What if I told you that, despite what you may have been led to
believe, the entire world was simply a highly detailed simulation. You
may say that sounds preposterous, and you would be right, but it is
still helpful to look at why it's wrong. (No, it is not because we don't all have Tron suits under our skin.)
The
Simulation Argument, by philosopher Nick Bostrom, is a proposition
based on the near certainty of a future where computers are capable of
physics simulations which would allow high-fidelity simulations It
postulates that one of three things MUST be true
- The fraction of human-level civilizations that reach a stage capable of running high-fidelity ancestor simulations is very close to zero
- The fraction of civilizations that are interested in running high-fidelity simulations of their evolutionary history, or variations thereof, is very close to zero
- The fraction of all people with our kind of experiences that are living in a simulation is very close to one.
I
happen to agree with the premise of the argument. There most certainly
will be ample refined and dense computational power in the near future
where we can make artificial worlds that are almost indistinguishable
from reality.While it may never been perfect enough to fool someone who
has actually experienced the real world, it would almost certainly be
possible to fool a human mind which has known nothing else.
If you include the entire past and future of the universe (which Bostrom does), then there is almost zero chance that the first two postulates are true. The fact that human society on Earth has almost reached this level of technology means it is almost certainly "reachable" by some civilization, and the idea that a significant fraction of civilizations who wouldn't want to run simulations in these artificial environments seems also very close to zero.
Unfortunately the third postulate
is not only fatally flawed but there are many other postulations that
can replace it in this trilemma. Let's take a closer look at the third,
and most discussed postulate.
The fraction of all people with our kind of experiences that are living in a simulation is very close to one.
While the first two postulates are rather obvious, the final one hinges on several fatally flawed assumptions:
- That there is some benefit in running these highly detailed simulations, so detailed that we each have (what appears to be at least) detailed consciousness, complex emotions, and also feel very detailed and traumatizing pain.
- That there is some overriding benefit in the vast majority of these simulations to hide the fact that the participants (you and me) are in fact in a simulation.
- The
ethics of deceiving self-aware artificial intelligences into believing
they are humans and that they are really living not only on the real
planet Earth, but in the real universe, must be vastly different from
what we currently think of as acceptable scientific experimentation.
- That
if simulation technology is possible, and allowed, that the number of
simulated humans would be much HIGHER than all the human beings that
have ever existed.
While many of us have been to a play or seen a movie where the actors have went through very emotional states, even putting themselves through physical pain in the quest for an "authentic" performance, it is a quite difficult position to defend to say that a director must actually harm her actors in order to get a meaningful performance from them. If future civilizations are running "simulations" why do they need to include such visceral pain and anguish in our "simulated" existence. If you say, "Well, it's only SIMULATED pain and anguish" I think most people would be hard-pressed to imagine how "simulated pain" would be different from real pain. Even if pain doesn't harm us physically it is still just as real. What possible information can be gained by torturing a bunch of AI in a computer?
Why would the simulation need to include a deception about the reality of the simulation? Many people live their whole lives with the belief that their is some sort of unseen "spirit world" or some other form of life after death. It would be simple enough to avoid much of the ethical conundrum to simply let everyone know that the world we are living in is a simulation and it could be escaped at any time by "death."
What
benefit would be lost if one of us, or all of us, discovered we were in
a simulation? Most of us would be quite terrified to discover we were
not real humans and our world was just a simulation. Why not simply
program the simulated humans to not worry about that? Why not remove the
ability for us to even think about it? While it doesn't prove we are
not in a simulation that we can both think quite rationally about it, as
well as be terrified at the thought of it being real, surely indicates
that the chances of us being in a simulation are far from 1/1.
It
seems an ethical simulation would simply allow all the simulated humans
to privately know they could exit the simulation at any time. The fact
that most governments consider suicide a crime on the level of murder, and
most religions in the world consider suicide to be a sin, indicate that
either the creators of these simulations are monstrously immoral, or
else we can be relatively assured we are not in a simulation. Either
way, it certainly doesn't indicate a simulation is MORE likely given the
circumstances of the human condition we are living with in this world.
Finally, probably the most objectively disqualifying argument against the Simulation Argument comes from how models and simulations usually work. Statistically speaking there is a limit to the sample size that is necessary to accurately model a population. It would be orders of magnitude smaller.
For instance, if some civilization wanted to run a model on how humans thought in the
2020's for instance, the model they would make would be SIGNIFICANTLY
smaller in population than the 7 billion people in the world. For accuracy of accuracy of as high as 99.9% they would actually only need a sample population of less than one million people. In fact, even if they simulated a population a large as the entire population of the human race throughout the 50,000 of our existence (about 110 billion humans) a sample population of one million people would suffice for the same 99.9% accuracy.
(Here is a sample size calculator from Survey Monkey if you want to look into this yourself.)
In other words, it would actually be MUCH less likely that you were a part of such a simulation. The odds would be more like 1 million/110 billion, which is 0.0009%
Even if they decided to run a trillion iterations of history, they would STILL only need to use the same 1 million artificial humans. What would be the point of making more? What would it even mean to add another? Would they just be a copy of another one or would the Simulation Gods create a unique person from scratch even if it wouldn't add significantly to the accuracy of their simulation? Would that mean it was just the same artificial human copied or is the entire million simply copies of some simulated Adam and Eve?
While there will almost certainly be a day when artificial intelligences inhabit simulated worlds, there is no reason to suspect they will be lied to and fooled into thinking they are somewhere they are not. The expense, trouble, and moral transgression are not worth the negligible increase in accuracy that would warrent using the number of simulated humans in ANY future simulation that would even come close to making the odds of us currently being in a simulation anywhere near 1/1.
So rest easy knowing you are definitely in the real world, but cherish the time you have. No resets available!
No comments:
Post a Comment