Are you living in a computer simulation actually argues for a disjunction that includes “we are almost certainly living in a computer simulation” along with two other statements.
That’s the difference between the Simulation Argument and the Simulation Hypothesis. The Simulation Argument is “you must deny one of these three statements” and the Simulation Hypothesis is “the statement to be denied is ‘I am not in a computer simulation’”.
That’s a really neat link. Thanks. That’s a paper by the director of FHI, Nick Bostrom, also one of the sponsors of LW. Just to summarize and to discuss, it essentially sets up three mutually exclusive possibilities. One, that post-human civilizations aren’t significantly interested in running earth-like simulations, two, that post-human civilizations just don’t make it (e.g., doomsday scenarios), or three, we actually live in a computer simulation ourselves. It doesn’t really argue that the third scenario is so likely, it just (roughly) establishes that these scenarios are mutually exclusive. This all comes under the main (fairly well established) belief that future computing power is capable of these sorts of large-scale simulations.
The argument and the paper is actually pretty reasonable, but the question of whether or not post-human civilizations would want to run earth-like simulations is the sticking point. Sure, it’s possible, but the resources required are huge, the upkeep involved, and so on...
I guess another main criticism you might make of the paper is that it relies pretty heavily on “Drake’s equation” type of reasoning where you don’t really know if you’ve gotten all the dependencies correct. It’s still valid it’s just highly simplistic and so somewhat suspicious on those grounds. And to boot, I think his N_sub(I) variable is actually mis-indicated… but maybe I was just reading a typoed draft or misunderstanding.
Maybe most interestingly, if you decide we’re in a simulation, then you have to wonder if there isn’t a long loop of father/grandfather/great-grand-dad/etc simulations, and the guys that are simulating us are just being simulated themselves. Anyways this is getting long so I’ll just recommend the article and leave it here.
You are right to be confused. The idea that the simulators would necessarily have human-like motives can only be justified on anthropocentric grounds—whatever is out there, it must be like us.
Anything capable of running us as a simulation might exist in any arbitrarily strange physical environment that allowed enough processing power for the job. There is no basis for the assumption that simulators would have humanly comprehensible motives or a similar physical environment.
The simulation problem requires that we think about our entire perceived universe as a single point in possible-universe-space, and it is not possible to extrapolate from this one point.
Are you living in a computer simulation actually argues for a disjunction that includes “we are almost certainly living in a computer simulation” along with two other statements.
That’s the difference between the Simulation Argument and the Simulation Hypothesis. The Simulation Argument is “you must deny one of these three statements” and the Simulation Hypothesis is “the statement to be denied is ‘I am not in a computer simulation’”.
That’s a really neat link. Thanks. That’s a paper by the director of FHI, Nick Bostrom, also one of the sponsors of LW. Just to summarize and to discuss, it essentially sets up three mutually exclusive possibilities. One, that post-human civilizations aren’t significantly interested in running earth-like simulations, two, that post-human civilizations just don’t make it (e.g., doomsday scenarios), or three, we actually live in a computer simulation ourselves. It doesn’t really argue that the third scenario is so likely, it just (roughly) establishes that these scenarios are mutually exclusive. This all comes under the main (fairly well established) belief that future computing power is capable of these sorts of large-scale simulations.
The argument and the paper is actually pretty reasonable, but the question of whether or not post-human civilizations would want to run earth-like simulations is the sticking point. Sure, it’s possible, but the resources required are huge, the upkeep involved, and so on...
I guess another main criticism you might make of the paper is that it relies pretty heavily on “Drake’s equation” type of reasoning where you don’t really know if you’ve gotten all the dependencies correct. It’s still valid it’s just highly simplistic and so somewhat suspicious on those grounds. And to boot, I think his N_sub(I) variable is actually mis-indicated… but maybe I was just reading a typoed draft or misunderstanding.
Maybe most interestingly, if you decide we’re in a simulation, then you have to wonder if there isn’t a long loop of father/grandfather/great-grand-dad/etc simulations, and the guys that are simulating us are just being simulated themselves. Anyways this is getting long so I’ll just recommend the article and leave it here.
It confuses me slightly that, from superficial glances, the discussion there and in threads like this one focuses on “ancestor” simulations, rather than simulations run by five-dimensional cephalopods. Ryan North got it right when he had T-Rex say “and not necessarily our own”, but then he seems to get confused when he says “a 1:1 simulation of a universe wouldn’t work”—why not?
Personally, I like Wei Dai’s conclusion that we both are and aren’t in a simulation.
You are right to be confused. The idea that the simulators would necessarily have human-like motives can only be justified on anthropocentric grounds—whatever is out there, it must be like us.
Anything capable of running us as a simulation might exist in any arbitrarily strange physical environment that allowed enough processing power for the job. There is no basis for the assumption that simulators would have humanly comprehensible motives or a similar physical environment.
The simulation problem requires that we think about our entire perceived universe as a single point in possible-universe-space, and it is not possible to extrapolate from this one point.