If we were in an imperfect simulation, wouldn't we be likely to find, as we look around, a lot of fine-tuning, many hierarchy problems, and a bunch of places where deep structure does not conform to what we thought we would see from shallower structure, but rather where we have to resort to putting in extra kludgy epicycles like "inflation", "dark energy", "dark matter", and "cosmological constants"?
Or maybe we are living in a simulation. I don't see electromagnetic wavelengths, after all: I see a color wheel. How is that not living in a simulation—albeit one created by my kludgy evolved sensory pathways rather than by DesCartes's Demon?
Enh. This depends very strongly on a particular view of "elegance" in physics that I don't think has any particular basis in anything but aesthetic taste. Which, to be fair, is also true of a lot of philosophy, so...
If the world is a simulation, one should expect to see miracles. These definitely exist; the miracle of telepathy is particularly well-attested. One should also expect implausible levels of safety or dysfunctionality within the world we see, and I do think we see both.
“(It’s a bit like the flippant response to arguments about free will and determinism: If it turns out that you don’t have free will, what are you going to do about it?)”
I'm a philosopher by training, not a scientist, and I can tell you you're not the only one to find the simulation hypothesis 'a bit of a cheat'. I'd go further and say it's a total cheat!
“If you imagine that there’s a simulation so perfect it’s impossible to detect, then it would be impossible to know that you’re in a simulation!”
I'd say you're not in a simulation. The only way such a proposition has meaning is if you imagine yourself stepping outside the world to take a programmer-god's eye view, but if you were able to do that, then your supposedly perfect simulation would no longer be perfect. The idea that we might be living in a perfect simulation is total nonsense. I agree, the only interesting simulation is a not perfect one.
Thanks for this. It was interesting to get the perspective of a non-philosopher. Though obviously having a practical upshot is not a condition on being interesting for me, I agree that having such an upshot is a plus.
At any rate, my real intention in commenting is to express that I share your impression that this work is not especially good. The arguments are weak and the conclusion implausible. Thus I do not take work on the simulation hypothesis seriously. One thought to which I repeatedly return is this: Science is a work in progress. Perhaps the more synoptic understanding we will have in 50 years will suggest that Boltzmann Brains are impossible after all.
I agree with every word. I'd go so far as to assert that we most likely do *not* live in a virtual simulation. I think Bostrom's claim VRs would be plentiful misses that simulating eight billion people requires a lot of hardware and energy. VRs of a planet and a fake surrounding universe would be hugely expensive, let alone ones that simulate more than what we on Earth can physically grasp. I'm not sure I believe future advances would make these inexpensive enough to be plentiful.
The other thing is that even now we're pondering what rights AGI ought to have if we can accomplish it (which remains to be seen; I'm skeptical). If we at our stage of development worry about the ethics of creating consciousness, what prohibitions might a far more advanced society have? Creating consciousness might be interdicted as we do atomic power or money.
(I checked Chalmers's new book out from the library, read the introduction, said with horror, "Chalmers believes this crap? Oh, no!" and returned it to the library.)
“VRs of a planet and a fake surrounding universe would be hugely expensive,”
“Expensive” is relative. Agreed, this is probably not feasible on an Earthly college professor’s budget.
But this whole conversation is based on some other worldly scale where college professors get unlimited research budgets and their own fusion reactor for powering their experiments.
Understood and factored in. It's not obvious that a simulation is possible without going down to the quantum level, and then you're attempting to simulate all the particles in a universe with something that has to be able to represent all those. In this recent post:
Not exactly, no. He makes references to it, but goes to some lengths to distance his position from that; how successful that distancing is may be open for debate.
Thanks. Thinking more on your review, I came up with some probably-unfair snark of my own: If a book is indistinguishable from mental masturbation, should we avoid it, or just go ahead and enjoy the climax? (I’m not a fan of Chalmers 🥴.)
"... by the time it gets to the fast-talking bit, I usually feel ready to nod along just to pick up the pace."
:D :D :D
Also, I share your feelings about the simulation hypothesis. And, like you, it's about as interesting to me as discussions of free will, which is to say, not at all. I have never understood why people found either topic worth more than one bull session in a dorm room, avec bongs.
How do you feel about the Templeton Institute having a presence on your campus? I've always been leery of them, because it always seemed like they were trying to sneak religion into everything that they funded. Probably if you're not thrilled about them being there, you won't want to say, and I understand that, but if you're enthusiastic about them, I'd like to know if I should change my view of them.
Different Templeton. The local Institute is the work of two alumni from the CS class of 1980, Rich and Mary, who gave us a big pile of money for Engineering and Computer Science. I don't believe they have a direct connection to the Templeton Foundation that does the religion-and-science stuff.
I agree that the simulation hypothesis is broadly uninteresting from any pragmatic standpoint, but with a few additional assumptions we can get it to prescribe action.
1. Whatever is running the simulation started the simulation for some purpose, i.e. our reality/simulation has some Telos to it.
2. The structure of the simulation (physics? society? Who knows what part of the simulation matters!) is structured to achieve this Telos. This is true whether or not the simulation is “perfect.”
Therefore, by observing “the simulation” we can glean this purpose, and act to achieve or obstruct this purpose. Perhaps the Creator keeps the simulation running because it produces some data it needs, and we’re behooved to help that along or risk being shut off. Or perhaps once we achieve the simulation purpose, we’ll get switched off, so we should prevent ourselves from reaching the “goal state.”
Obviously, this interpretation really does make a religion of Simulationism, and there’s little reason to think we could glean this information if we wanted to. But perhaps the implied purpose is part of the Simulation Hypothesis’ appeal: it grants some higher power/destiny that the scientific revolution in many ways stripped away, and primitive superstitions are replaced with modern, scientificish superstitions.
I look forward to the religious wars over whether or not we should fulfill the objectives of The Programmer
The argument about the simulations seems really shaky to me. There are millions of computer simulations (video games) but nobody would argue the characters have awareness. For the claim to be true, it has to be possible for a simulation to include self-aware individuals (far from certain). You also need there to be advanced civilizations (not us), although I don't even know how one would put a probability on the chance of such a civilization existing in a hypothetical higher level universe.
Yeah, those are a lot of the points that are kind of fast-talked through-- there's a lot of "There's no reason to believe that self-awareness CAN'T be generated in a simulated brain..." and so on.
If we were in an imperfect simulation, wouldn't we be likely to find, as we look around, a lot of fine-tuning, many hierarchy problems, and a bunch of places where deep structure does not conform to what we thought we would see from shallower structure, but rather where we have to resort to putting in extra kludgy epicycles like "inflation", "dark energy", "dark matter", and "cosmological constants"?
Or maybe we are living in a simulation. I don't see electromagnetic wavelengths, after all: I see a color wheel. How is that not living in a simulation—albeit one created by my kludgy evolved sensory pathways rather than by DesCartes's Demon?
Enh. This depends very strongly on a particular view of "elegance" in physics that I don't think has any particular basis in anything but aesthetic taste. Which, to be fair, is also true of a lot of philosophy, so...
If the world is a simulation, one should expect to see miracles. These definitely exist; the miracle of telepathy is particularly well-attested. One should also expect implausible levels of safety or dysfunctionality within the world we see, and I do think we see both.
“(It’s a bit like the flippant response to arguments about free will and determinism: If it turns out that you don’t have free will, what are you going to do about it?)”
Please make tshirts with this quote. 😁
I'm a philosopher by training, not a scientist, and I can tell you you're not the only one to find the simulation hypothesis 'a bit of a cheat'. I'd go further and say it's a total cheat!
“If you imagine that there’s a simulation so perfect it’s impossible to detect, then it would be impossible to know that you’re in a simulation!”
I'd say you're not in a simulation. The only way such a proposition has meaning is if you imagine yourself stepping outside the world to take a programmer-god's eye view, but if you were able to do that, then your supposedly perfect simulation would no longer be perfect. The idea that we might be living in a perfect simulation is total nonsense. I agree, the only interesting simulation is a not perfect one.
Thanks for this. It was interesting to get the perspective of a non-philosopher. Though obviously having a practical upshot is not a condition on being interesting for me, I agree that having such an upshot is a plus.
At any rate, my real intention in commenting is to express that I share your impression that this work is not especially good. The arguments are weak and the conclusion implausible. Thus I do not take work on the simulation hypothesis seriously. One thought to which I repeatedly return is this: Science is a work in progress. Perhaps the more synoptic understanding we will have in 50 years will suggest that Boltzmann Brains are impossible after all.
I ran into a woman on a dating site once who professed an interest in the simulation hypothesis. Here, in its entirety, is my attempt to woo her:
Ori: If we were living in a simulation, would you live your life any differently?
Woman: We are living in a simulation.
Ori: If we weren't living in a simulation, would you live your life any differently?
Woman: <block>
I agree with every word. I'd go so far as to assert that we most likely do *not* live in a virtual simulation. I think Bostrom's claim VRs would be plentiful misses that simulating eight billion people requires a lot of hardware and energy. VRs of a planet and a fake surrounding universe would be hugely expensive, let alone ones that simulate more than what we on Earth can physically grasp. I'm not sure I believe future advances would make these inexpensive enough to be plentiful.
The other thing is that even now we're pondering what rights AGI ought to have if we can accomplish it (which remains to be seen; I'm skeptical). If we at our stage of development worry about the ethics of creating consciousness, what prohibitions might a far more advanced society have? Creating consciousness might be interdicted as we do atomic power or money.
(I checked Chalmers's new book out from the library, read the introduction, said with horror, "Chalmers believes this crap? Oh, no!" and returned it to the library.)
“VRs of a planet and a fake surrounding universe would be hugely expensive,”
“Expensive” is relative. Agreed, this is probably not feasible on an Earthly college professor’s budget.
But this whole conversation is based on some other worldly scale where college professors get unlimited research budgets and their own fusion reactor for powering their experiments.
Understood and factored in. It's not obvious that a simulation is possible without going down to the quantum level, and then you're attempting to simulate all the particles in a universe with something that has to be able to represent all those. In this recent post:
https://logosconcarne.substack.com/p/digital-simulation
I wrote about the size of a model of a human brain just down to the synapse level, and for a single brain the static model is over 12 petabytes.
Whoa, he’s just using the Ontological argument?! https://en.m.wikipedia.org/wiki/Ontological_argument GTFOH.
Not exactly, no. He makes references to it, but goes to some lengths to distance his position from that; how successful that distancing is may be open for debate.
Thanks. Thinking more on your review, I came up with some probably-unfair snark of my own: If a book is indistinguishable from mental masturbation, should we avoid it, or just go ahead and enjoy the climax? (I’m not a fan of Chalmers 🥴.)
This piece and its comments give me a strong sensation of déjà vu...
For full credit, you should've left this comment three times...
"... by the time it gets to the fast-talking bit, I usually feel ready to nod along just to pick up the pace."
:D :D :D
Also, I share your feelings about the simulation hypothesis. And, like you, it's about as interesting to me as discussions of free will, which is to say, not at all. I have never understood why people found either topic worth more than one bull session in a dorm room, avec bongs.
How do you feel about the Templeton Institute having a presence on your campus? I've always been leery of them, because it always seemed like they were trying to sneak religion into everything that they funded. Probably if you're not thrilled about them being there, you won't want to say, and I understand that, but if you're enthusiastic about them, I'd like to know if I should change my view of them.
Different Templeton. The local Institute is the work of two alumni from the CS class of 1980, Rich and Mary, who gave us a big pile of money for Engineering and Computer Science. I don't believe they have a direct connection to the Templeton Foundation that does the religion-and-science stuff.
Ah. I should have recognized the missing "Foundation." Thanks.
I agree that the simulation hypothesis is broadly uninteresting from any pragmatic standpoint, but with a few additional assumptions we can get it to prescribe action.
1. Whatever is running the simulation started the simulation for some purpose, i.e. our reality/simulation has some Telos to it.
2. The structure of the simulation (physics? society? Who knows what part of the simulation matters!) is structured to achieve this Telos. This is true whether or not the simulation is “perfect.”
Therefore, by observing “the simulation” we can glean this purpose, and act to achieve or obstruct this purpose. Perhaps the Creator keeps the simulation running because it produces some data it needs, and we’re behooved to help that along or risk being shut off. Or perhaps once we achieve the simulation purpose, we’ll get switched off, so we should prevent ourselves from reaching the “goal state.”
Obviously, this interpretation really does make a religion of Simulationism, and there’s little reason to think we could glean this information if we wanted to. But perhaps the implied purpose is part of the Simulation Hypothesis’ appeal: it grants some higher power/destiny that the scientific revolution in many ways stripped away, and primitive superstitions are replaced with modern, scientificish superstitions.
I look forward to the religious wars over whether or not we should fulfill the objectives of The Programmer
The argument about the simulations seems really shaky to me. There are millions of computer simulations (video games) but nobody would argue the characters have awareness. For the claim to be true, it has to be possible for a simulation to include self-aware individuals (far from certain). You also need there to be advanced civilizations (not us), although I don't even know how one would put a probability on the chance of such a civilization existing in a hypothetical higher level universe.
Yeah, those are a lot of the points that are kind of fast-talked through-- there's a lot of "There's no reason to believe that self-awareness CAN'T be generated in a simulated brain..." and so on.