If we lived as ems in a simulated universe, literally any claim of any religion of any religion could be true. Any other claim, too.
The point is, when people seriously ponder if we lived in a simulated universe but have nothing but scorn for religions, it is not so much rational as cultural. In a sci-fi geek subculture, simulation is cool, religions not.
Culture, in this sense, means the following. If theory A and B makes the same predictions but the formulation, the wording of A seems vastly preferable than B—that is culture.
IMHO this is one of the most important kind of bias in me I need to control for. I need to ask myself “if this was reworded in a way that is culturally compatible with me, would I still reject it so strongly?” or the opposite “is it something entirely without merit, but merely worded in a way that presses my ‘cool’ buttons?”
If we live in a non-simulated universe, (almost) any claim of any religion can still be true. It’s just so astronomically unlikely that we shouldn’t spend any time considering it. (See: Pascal’s Wager.)
How are ems different? A uniform prior over all possible religions isn’t useful. What evidence would they have, from the mere fact of being in a simulation, that any particular set of religious claims is likely, aside from the bare claim “there are Simulators but we don’t know anything about them except that they want our universe to be simulated”?
I think a step from a lawful universe to an arbitrarily programmable universe would be fairly big. We exclude miracles in principle, for ems, miracles would be possible. If ems would agree the universe has a Programmer who is allmighty (can simulate what he wants), omniscient (can look at any part of the source code) it would be a big step.
So it would be relevant to the meta step “are miracles possible?” and not to the more object level step “is this X miracle report for reals?”
Also a uniform prior would be useful for the more syncretic approaches—it would not be so useful for primitive kind of religious approaches like “thing X is written in my holy book while some other holy books writes something totally different” but it would be more useful for the kind of “unified theology/philosophy” that the best minds in Islam (Avicenna) Judaism (Maimonides) Catholicism (Aquinas) were developing.
My point is not the veracity of each religious claim but the veracity of living in a kind of universe where religious claims are possible, even likely that something like those happens (the Programmer plays around with stuff, tests ideas, throws a bunch of tablets on a desert tribe to experimentally test some sociology theories of ems) and then all we had to decide is which ones.
Essentially, you assume here that the distribution of “universes someone bothers to simulate” is different from the Solomonoff priors in a way that makes universes where religion is true more likely.
May be. If the simulators are humans, some of them would enjoy playing gods; and even if it is a small minority, it would still be a larger fraction than universes where gods “naturally” exist as complicated laws of physics. If the simulators are alien intelligences… well, I would be less certain about those, but still seems like the fraction of situations where the simulator decides to personally interact with the simulation should be larger than universes where an intelligence is “hardcoded” into laws of physics.
But there would still be many simulated universes where the simulator is not interested in interaction with sentient beings, and all religions arise naturally for reasons unrelated to their correctness. Or the simulator would interact with the world, but in a manner totally different from what religions talk about; imagine for example that our world is just a computer game played by a bored teenager who once in a few millions of years clicks a mouse button to drop a huge meteor and change ecosystem; the individual humans are too small and short-lived for him to even notice. Maybe the simulator completely ignores humans, and is only using this universe as an incubator for an AGI that he will later ask to compute some mathematical problem.
So, the probability of some religion being true could be greater, but still relatively small. Not sure if large enough to excuse Pascal’s wager.
Across the multiverse, some universes support life, and some of those features civs which create lots of sims. Given that our physics seems to support enormous numbers of sim universes per basement universe, most of the copies of ourselves are probably in sims.
Realizing that, there is a general convergent strategy that we should precommit to: in the future when we have the power we should ‘win the sims arms race’ so to speak and create many sims of our past. In this way we ensure that our sim is benevolent.
These kinds of game-theoretic convergent considerations deeply effect the distribution over sim scenarios, and they do suggest what kinds of religious/metaphysical theories are more or less likely to be true.
General resurrection—as in the book religions (zoraster/judaic/christian/islam), is a high likelihood cluster, as it naturally falls out of the whole use future sim power to defeat death strategy mentioned earlier. Basically, we want the god above to be aligned to our values, and we can best achieve this by expending sim power in the future.
The point is, when people seriously ponder if we lived in a simulated universe but have nothing but scorn for religions, it is not so much rational as cultural. In a sci-fi geek subculture, simulation is cool, religions not.
I hesitate to mention this, but I believe there was a period where a crankish LWer or two was advocating religious belief on simulator-god grounds. I think it had more to do with intellectual hipsterism than anything.
Your culture point is also discussed in this Slate Star Codex post.
See here for an example of the crankish intellectual hipsterism being referred to. Also, we should be careful to distinguish theistic beliefs from religious ones.
If we lived as ems in a simulated universe, literally any claim of any religion of any religion could be true. Any other claim, too.
The point is, when people seriously ponder if we lived in a simulated universe but have nothing but scorn for religions, it is not so much rational as cultural. In a sci-fi geek subculture, simulation is cool, religions not.
Culture, in this sense, means the following. If theory A and B makes the same predictions but the formulation, the wording of A seems vastly preferable than B—that is culture.
IMHO this is one of the most important kind of bias in me I need to control for. I need to ask myself “if this was reworded in a way that is culturally compatible with me, would I still reject it so strongly?” or the opposite “is it something entirely without merit, but merely worded in a way that presses my ‘cool’ buttons?”
If we live in a non-simulated universe, (almost) any claim of any religion can still be true. It’s just so astronomically unlikely that we shouldn’t spend any time considering it. (See: Pascal’s Wager.)
How are ems different? A uniform prior over all possible religions isn’t useful. What evidence would they have, from the mere fact of being in a simulation, that any particular set of religious claims is likely, aside from the bare claim “there are Simulators but we don’t know anything about them except that they want our universe to be simulated”?
I think a step from a lawful universe to an arbitrarily programmable universe would be fairly big. We exclude miracles in principle, for ems, miracles would be possible. If ems would agree the universe has a Programmer who is allmighty (can simulate what he wants), omniscient (can look at any part of the source code) it would be a big step.
So it would be relevant to the meta step “are miracles possible?” and not to the more object level step “is this X miracle report for reals?”
Also a uniform prior would be useful for the more syncretic approaches—it would not be so useful for primitive kind of religious approaches like “thing X is written in my holy book while some other holy books writes something totally different” but it would be more useful for the kind of “unified theology/philosophy” that the best minds in Islam (Avicenna) Judaism (Maimonides) Catholicism (Aquinas) were developing.
My point is not the veracity of each religious claim but the veracity of living in a kind of universe where religious claims are possible, even likely that something like those happens (the Programmer plays around with stuff, tests ideas, throws a bunch of tablets on a desert tribe to experimentally test some sociology theories of ems) and then all we had to decide is which ones.
Essentially, you assume here that the distribution of “universes someone bothers to simulate” is different from the Solomonoff priors in a way that makes universes where religion is true more likely.
May be. If the simulators are humans, some of them would enjoy playing gods; and even if it is a small minority, it would still be a larger fraction than universes where gods “naturally” exist as complicated laws of physics. If the simulators are alien intelligences… well, I would be less certain about those, but still seems like the fraction of situations where the simulator decides to personally interact with the simulation should be larger than universes where an intelligence is “hardcoded” into laws of physics.
But there would still be many simulated universes where the simulator is not interested in interaction with sentient beings, and all religions arise naturally for reasons unrelated to their correctness. Or the simulator would interact with the world, but in a manner totally different from what religions talk about; imagine for example that our world is just a computer game played by a bored teenager who once in a few millions of years clicks a mouse button to drop a huge meteor and change ecosystem; the individual humans are too small and short-lived for him to even notice. Maybe the simulator completely ignores humans, and is only using this universe as an incubator for an AGI that he will later ask to compute some mathematical problem.
So, the probability of some religion being true could be greater, but still relatively small. Not sure if large enough to excuse Pascal’s wager.
Across the multiverse, some universes support life, and some of those features civs which create lots of sims. Given that our physics seems to support enormous numbers of sim universes per basement universe, most of the copies of ourselves are probably in sims.
Realizing that, there is a general convergent strategy that we should precommit to: in the future when we have the power we should ‘win the sims arms race’ so to speak and create many sims of our past. In this way we ensure that our sim is benevolent.
These kinds of game-theoretic convergent considerations deeply effect the distribution over sim scenarios, and they do suggest what kinds of religious/metaphysical theories are more or less likely to be true.
General resurrection—as in the book religions (zoraster/judaic/christian/islam), is a high likelihood cluster, as it naturally falls out of the whole use future sim power to defeat death strategy mentioned earlier. Basically, we want the god above to be aligned to our values, and we can best achieve this by expending sim power in the future.
As I have repeatedly stated (without much response), first at the bottom of my baseline post.
I hesitate to mention this, but I believe there was a period where a crankish LWer or two was advocating religious belief on simulator-god grounds. I think it had more to do with intellectual hipsterism than anything.
Your culture point is also discussed in this Slate Star Codex post.
See here for an example of the crankish intellectual hipsterism being referred to. Also, we should be careful to distinguish theistic beliefs from religious ones.
Theistic beliefs from religious practices then.