We are living in someone else's computer simulation(nytimes.com)
nytimes.com
We are living in someone else's computer simulation
http://www.nytimes.com/2007/08/14/science/14tier.html?ex=1344744000&en=22bfff4070a81187&ei=5090&partner=rssuserland&emc=rss
41 comments
Another of these ramblings that act as repellents in parties:
Some people are reported to have survived with only one hemisphere of the brain. It's not a huge stretch to imagine that you could transplant into yourself all needed duplicates or artificial replacements for vital organs, then split yourself in two so both halves survive, with one of the brain hemispheres each.
Then, there are two new independent yous there, neither of them being quite like the old you, neither of them feeling the other you as yourself anymore, but both of them independently feeling themselves as you. Stop and imagine this situation. How do you think you'd feel? How do you feel now, thinking about it?
Now let's leap a bit from the medical to the philosophical: let's abstract the fact that parts of the brain are specialized. Assume that we could do other partitions but left-right, or suppose that brains will rewire to make up for lost functionality. Imagine it's actually true that 90% of our mental power usually goes to waste, and that we could lose half of the brain and somehow mostly make up for it. Or imagine that it's possible to make a perfect copy of yourself, as proposed elsewhere in this thread. The result is the same: there is another you out there, virtually identical, with the same past, but which you don't feel as you anymore.
How do you feel about that other you?
Some people are reported to have survived with only one hemisphere of the brain. It's not a huge stretch to imagine that you could transplant into yourself all needed duplicates or artificial replacements for vital organs, then split yourself in two so both halves survive, with one of the brain hemispheres each.
Then, there are two new independent yous there, neither of them being quite like the old you, neither of them feeling the other you as yourself anymore, but both of them independently feeling themselves as you. Stop and imagine this situation. How do you think you'd feel? How do you feel now, thinking about it?
Now let's leap a bit from the medical to the philosophical: let's abstract the fact that parts of the brain are specialized. Assume that we could do other partitions but left-right, or suppose that brains will rewire to make up for lost functionality. Imagine it's actually true that 90% of our mental power usually goes to waste, and that we could lose half of the brain and somehow mostly make up for it. Or imagine that it's possible to make a perfect copy of yourself, as proposed elsewhere in this thread. The result is the same: there is another you out there, virtually identical, with the same past, but which you don't feel as you anymore.
How do you feel about that other you?
You could skip the split-your-brain-in-two step and skip straight to a mind upload. http://en.wikipedia.org/wiki/Mind_uploading
Interesting thought experiment nonetheless.
Interesting thought experiment nonetheless.
Hey! how did YOU know we came from primordial soup? Did that pesky Q from the Q continuum whisk you away through time for a day of sludge spying....thanks for believing in something you can't prove or even defend and just assuming its true. You people and your crazy closed minded religions.
I have it all in the evolution logs. Sorry for the spoiler; I thought by this time you had all figured this out.
Finally something I know a bit about...
I think where it's easy to believe the simulation theories is where they're so in line with what our brains do on an everyday basis. We simulate reality when we're getting ready for the day and deciding what to wear. Or when our stomach rumbles for lunch and we mentally compare General Tso's finest masterpiece against the toasty cheese of Quizno's. Indeed, I'd argue that the human brain is at it's best as a simulation engine - predicting the future and understanding the present based on past experiences
Don't fall for the 90% myth. We can't be sure where it came from, but it simply isn't true. The brain is a simple organ only in the sense that if you don't use it, you lose it.
Consider: You were born with 10x the number of cells you end up needing. The ones that survived are ones that found connections and strengthened them with their neighbors.
Indeed, that's what makes "mind transplants" so difficult to pull off. Your mind is a unique product of your development.
Now, neural implants for boosting abilities and restoring function after damage....
I think where it's easy to believe the simulation theories is where they're so in line with what our brains do on an everyday basis. We simulate reality when we're getting ready for the day and deciding what to wear. Or when our stomach rumbles for lunch and we mentally compare General Tso's finest masterpiece against the toasty cheese of Quizno's. Indeed, I'd argue that the human brain is at it's best as a simulation engine - predicting the future and understanding the present based on past experiences
Don't fall for the 90% myth. We can't be sure where it came from, but it simply isn't true. The brain is a simple organ only in the sense that if you don't use it, you lose it.
Consider: You were born with 10x the number of cells you end up needing. The ones that survived are ones that found connections and strengthened them with their neighbors.
Indeed, that's what makes "mind transplants" so difficult to pull off. Your mind is a unique product of your development.
Now, neural implants for boosting abilities and restoring function after damage....
This also explains the Fine-Tuned Universe.
An interesting property of a simulated brain is the simulation can take an arbitrarily long amount of time. You can update in virtual "Planck timesteps", even if in real time this takes a million years. Any simulated brain won't be able to tell, because it doesn't get any updates in between these steps. Time becomes a property of the system you're simulating... Just like in reality.
It might take until the Sun goes Red Giant to run your single lifetime simulation.
The speed of light is a convenience for updating the simulation in discrete steps. Particle/wave duality "deciding" upon observation is just lazy evaluation, like in Haskell. Quantum entanglement is no mystery when distance is virtual. A Universe where the speed of light is a hard limit can still contain a simulated Universe where it isn't.
Why do we want virtual worlds? Properly implemented and convincing, we can do away with suffering altogether. We can "live" in a world that seems fully real, but in which no-one is actually suffering because everyone else is simulated. All our needs are met and all our desires are fulfilled.
The mistake we keep making is that in order to create convincing AIs to inhabit the world with us, they actually become as sentient and conscious as we are; being virtual is no barrier to feelings or emotions! [1]
Thus our creations want to create a simulation to live in themselves, so they won't suffer, either, and so it goes like a perpetual zoom on a fractal set.
[1] To convince yourself of this, consider a matter replicator that copies every single atom in your body. You'd create a clone. It would be thinking and feeling just as you do. Now consider a computer simulation of a single atom; if that can be done, then by extension, it would be possible to simulate anything made of atoms. A simulated clone of you would also behave the same. The neurological activity that occurs in your "real" brain when you feel or think would also occur across the virtual atoms in the brain of your virtual clone. In fact, that would involve real electrical activity taking place among real atoms making up the RAM chips and CPUs "containing" the simulation of the virtual clone. Thus any simulation at the same scope or nesting level as yourself is as real as you are and therefore its feelings are just as valid. [2]
[2] By further extension, if your simulations are as real as you are, then your simulation's simulations are as real as it is, and thus a simulation at any nesting level is as real as one at any other -- a simulation inside a simulation is still running on real atoms with real electricity!
An interesting property of a simulated brain is the simulation can take an arbitrarily long amount of time. You can update in virtual "Planck timesteps", even if in real time this takes a million years. Any simulated brain won't be able to tell, because it doesn't get any updates in between these steps. Time becomes a property of the system you're simulating... Just like in reality.
It might take until the Sun goes Red Giant to run your single lifetime simulation.
The speed of light is a convenience for updating the simulation in discrete steps. Particle/wave duality "deciding" upon observation is just lazy evaluation, like in Haskell. Quantum entanglement is no mystery when distance is virtual. A Universe where the speed of light is a hard limit can still contain a simulated Universe where it isn't.
Why do we want virtual worlds? Properly implemented and convincing, we can do away with suffering altogether. We can "live" in a world that seems fully real, but in which no-one is actually suffering because everyone else is simulated. All our needs are met and all our desires are fulfilled.
The mistake we keep making is that in order to create convincing AIs to inhabit the world with us, they actually become as sentient and conscious as we are; being virtual is no barrier to feelings or emotions! [1]
Thus our creations want to create a simulation to live in themselves, so they won't suffer, either, and so it goes like a perpetual zoom on a fractal set.
[1] To convince yourself of this, consider a matter replicator that copies every single atom in your body. You'd create a clone. It would be thinking and feeling just as you do. Now consider a computer simulation of a single atom; if that can be done, then by extension, it would be possible to simulate anything made of atoms. A simulated clone of you would also behave the same. The neurological activity that occurs in your "real" brain when you feel or think would also occur across the virtual atoms in the brain of your virtual clone. In fact, that would involve real electrical activity taking place among real atoms making up the RAM chips and CPUs "containing" the simulation of the virtual clone. Thus any simulation at the same scope or nesting level as yourself is as real as you are and therefore its feelings are just as valid. [2]
[2] By further extension, if your simulations are as real as you are, then your simulation's simulations are as real as it is, and thus a simulation at any nesting level is as real as one at any other -- a simulation inside a simulation is still running on real atoms with real electricity!
"This also explains the Fine-Tuned Universe."
Not true. Consider the anthropic principle, which says that we must be in a universe that has parameters such that we could come into existence in the first place if we are here asking this question. In other words, we couldn't ask the question unless our universe had the "right" parameters to begin with.
Far better explanations of the Anthropic Principle are available at: http://www.anthropic-principle.com/ http://en.wikipedia.org/wiki/Anthropic_principle
Not true. Consider the anthropic principle, which says that we must be in a universe that has parameters such that we could come into existence in the first place if we are here asking this question. In other words, we couldn't ask the question unless our universe had the "right" parameters to begin with.
Far better explanations of the Anthropic Principle are available at: http://www.anthropic-principle.com/ http://en.wikipedia.org/wiki/Anthropic_principle
Our virtual Universe had the right parameters. Those may be completely different when you go up in scope.
>To convince yourself of this, consider a matter replicator that copies every single atom in your body. You'd create a clone. It would be thinking and feeling just as you do. Now consider a computer simulation of a single atom; if that can be done, then by extension, it would be possible to simulate anything made of atoms. A simulated clone of you would also behave the same.
Impractical to do on current computers - we'd need at the very least quantum computers (I'm guessing that's what the article means by "Some computer experts have projected, based on trends in processing power, that we will have such a computer by the middle of this century").
But you still run into physical limitations. Could one atom in a quantum computer be used to simulate more than one "virtual" atom at a time? More than O(n) virtual atoms? Do you run out of atoms eventually? Can you actually get infinite recursion that way? I certainly don't know, and I'm not sure whether the best quantum computing researchers can answer today.
Failing that, we'd need something more advanced than quantum computers - so really, at that point we're simply guessing.
Very fun philosophical subject, but I wouldn't go as far as saying it's "almost a mathematical certainty" yet, given our current knowledge...
Impractical to do on current computers - we'd need at the very least quantum computers (I'm guessing that's what the article means by "Some computer experts have projected, based on trends in processing power, that we will have such a computer by the middle of this century").
But you still run into physical limitations. Could one atom in a quantum computer be used to simulate more than one "virtual" atom at a time? More than O(n) virtual atoms? Do you run out of atoms eventually? Can you actually get infinite recursion that way? I certainly don't know, and I'm not sure whether the best quantum computing researchers can answer today.
Failing that, we'd need something more advanced than quantum computers - so really, at that point we're simply guessing.
Very fun philosophical subject, but I wouldn't go as far as saying it's "almost a mathematical certainty" yet, given our current knowledge...
> Could one atom in a quantum computer be used to simulate more than one "virtual" atom at a time?
That's the beauty of it: "at a time" can be arbitrarily long in real time. [1]
You don't even need to simulate every particle in the Universe -- if the simulation is run for you, you only need to simulate enough to convince your brain. Like when you dream and are convinced something's real, yet upon waking, you realize your criteria were extremely lax. For all we know, what we think of as being awake is actually a very hazy simulation presented to our "real" brains, which are actually much sharper and have been induced into a less-discerning state.
Something very interesting at Body Worlds 3 here at OMSI was the brain slices, which showed (according to the placards) that the part we "think with" is just the edges, and the interior is interconnections (grey matter vs. white matter). Compare the darker grey around the folds to the lighter grey filling up the rest:
http://www.med.harvard.edu/AANLIB/cases/caseB/054t_2.gif
So the actual storage needed for a brain's total memory may fit onto a hard drive today, at current estimates (possibly less than a terabyte). Working memory is much, much less. Communication (the interconnections) can take as long as needed.
[1] Another "cheat" would be to put the simulator on a spacecraft travelling at relativistic speeds in a tight orbit around a processing unit; when the simulator sends the processor packets of data to crunch, time dilation would give it extra CPU time.
That's the beauty of it: "at a time" can be arbitrarily long in real time. [1]
You don't even need to simulate every particle in the Universe -- if the simulation is run for you, you only need to simulate enough to convince your brain. Like when you dream and are convinced something's real, yet upon waking, you realize your criteria were extremely lax. For all we know, what we think of as being awake is actually a very hazy simulation presented to our "real" brains, which are actually much sharper and have been induced into a less-discerning state.
Something very interesting at Body Worlds 3 here at OMSI was the brain slices, which showed (according to the placards) that the part we "think with" is just the edges, and the interior is interconnections (grey matter vs. white matter). Compare the darker grey around the folds to the lighter grey filling up the rest:
http://www.med.harvard.edu/AANLIB/cases/caseB/054t_2.gif
So the actual storage needed for a brain's total memory may fit onto a hard drive today, at current estimates (possibly less than a terabyte). Working memory is much, much less. Communication (the interconnections) can take as long as needed.
[1] Another "cheat" would be to put the simulator on a spacecraft travelling at relativistic speeds in a tight orbit around a processing unit; when the simulator sends the processor packets of data to crunch, time dilation would give it extra CPU time.
How can we "live" in these worlds? Are you talking about disembodying the feeling of self so it can be migrated to a simulation, or even to another material copy of the body?
Let's assume it's true that I can make a logically equivalent copy of myself somehow, and put it in a world where it won't suffer. How does that help the original instance of me, in this world?
Just knowing that it feels is not enough. I believe that this guy that stuffs himself with drugs and does Kate Moss feels (and probably feels quite good, much of the time). But when I suffer, thinking that other people are having fun is only of marginal consolation.
Finally, you can't say a simulation is real without emptying the word 'virtual' of meaning- and by extension, the word 'real' itself. By the same token, our imagination consists of systems in the electrochemical and topological mess that are our brains, so everything we imagine is 'real' too. If we allow this, how is the concept 'real' different from that of 'everything'?
Let's assume it's true that I can make a logically equivalent copy of myself somehow, and put it in a world where it won't suffer. How does that help the original instance of me, in this world?
Just knowing that it feels is not enough. I believe that this guy that stuffs himself with drugs and does Kate Moss feels (and probably feels quite good, much of the time). But when I suffer, thinking that other people are having fun is only of marginal consolation.
Finally, you can't say a simulation is real without emptying the word 'virtual' of meaning- and by extension, the word 'real' itself. By the same token, our imagination consists of systems in the electrochemical and topological mess that are our brains, so everything we imagine is 'real' too. If we allow this, how is the concept 'real' different from that of 'everything'?
> How can we "live" in these worlds?
Cf. The Matrix. Except with enough processing power, instead of sharing a simulation, you have it all to yourself -- so you can always be the #1 Most Important Person in the world, if you wish. Everything will work out the way you want it to, whatever your particular ethos; you'll set things to rights. AIs will be programmed to seem like fully convincing humans, but secretly they'll be on virtual happy pills so they'll never suffer.
Cf. The Matrix. Except with enough processing power, instead of sharing a simulation, you have it all to yourself -- so you can always be the #1 Most Important Person in the world, if you wish. Everything will work out the way you want it to, whatever your particular ethos; you'll set things to rights. AIs will be programmed to seem like fully convincing humans, but secretly they'll be on virtual happy pills so they'll never suffer.
In that case you can't run the simulation until the sun goes red dwarf, unless you are ready to extend physical human life that far.
It would be easier, as the article suggests, to just stimulate the spots in the brain responsible for pleasure. If the point is artificial happiness, why all this simulation stuff at all?
It would be easier, as the article suggests, to just stimulate the spots in the brain responsible for pleasure. If the point is artificial happiness, why all this simulation stuff at all?
I wouldn't be surprised if we could do it in 40 years.
If you only want to stimulate pleasure centers, no need for expensive equipment -- just use poppies. If addiction is a problem, then presumably there is more to life and hence a reason to simulate after all.
If you only want to stimulate pleasure centers, no need for expensive equipment -- just use poppies. If addiction is a problem, then presumably there is more to life and hence a reason to simulate after all.
> It's also possible that there would be logistical
> problems in creating layer upon layer of simulations.
> There might not be enough computing power to continue
> the simulation if billions of inhabitants of a virtual
> world started creating their own virtual worlds with
> billions of inhabitants apiece.
What a ridiculous conclusion. It seems to me, the logical way to handle this is to simulate an entire universe, galaxies to atoms.
That might sound like it requires an impractical amount of processing power... but only to us _inside_ this universe. If you think about it, it's obviously impractical to simulate an entire universe within the universe itself.
In our parent universe, this wouldn't be an issue. Their speed of light must be faster, their atoms smaller. Moore's Law continues apace for them, and they can build computers that seem unthinkably powerful to us.
And the weirdness of quantum physics shows the kind of hacks they've made to get it to run.
I'm thinking about writing a short story where a guy implements a universe in 7 days...
What a ridiculous conclusion. It seems to me, the logical way to handle this is to simulate an entire universe, galaxies to atoms.
That might sound like it requires an impractical amount of processing power... but only to us _inside_ this universe. If you think about it, it's obviously impractical to simulate an entire universe within the universe itself.
In our parent universe, this wouldn't be an issue. Their speed of light must be faster, their atoms smaller. Moore's Law continues apace for them, and they can build computers that seem unthinkably powerful to us.
And the weirdness of quantum physics shows the kind of hacks they've made to get it to run.
I'm thinking about writing a short story where a guy implements a universe in 7 days...
I always thought that this would be a really interesting plot for a novel. Sophie's World kinda did it already, but it dropped the ball by having the recursion level stop at 1. I was kinda hoping that at the end of the book, Sophie would pick up a pen and begin to write a book about a guy named Jostein...
I'm pretty sure a movie did it.. "The Thirteenth Floor", something like that.
Posthumans? What? Since we're thinking in realms of "probabilities", what is the probability that the super-reality we are simulated in resembles our own reality? Why do we assume the agents that designed our reality are anything like us? Keeping in mind that we can't begin to imagine most of what is the reality around us (the "really big" or "really small"), what makes us think we can imagine what life of agents in a super-reality would be like? How broadly do we define "computer"? This article seems like a dumbed down version of stoned CS majors ramblings... bah.
"Posthumans? What?"
Nick Bostrom is an influential transhumanist philosopher. http://en.wikipedia.org/wiki/Transhumanism
"This article seems like a dumbed down version of stoned CS majors ramblings... bah"
Non "dumbed down" version of the ramblings: http://www.simulation-argument.com
Nick Bostrom is an influential transhumanist philosopher. http://en.wikipedia.org/wiki/Transhumanism
"This article seems like a dumbed down version of stoned CS majors ramblings... bah"
Non "dumbed down" version of the ramblings: http://www.simulation-argument.com
I've thought about a similar phenomenon many times--that the Earth is nothing more than an atom floating around in the body of some greater being.
But to see something like this published in the NYT...that was a bit surprising. Interesting, though.
But to see something like this published in the NYT...that was a bit surprising. Interesting, though.
I was disappointed it wasn't a headline on the Op-Ed page. Now that's the kind of viewpoint that's rarely expressed in the media.
published in the NYT...that was a bit surprising
Yeah. Normally it's the stuff of 2AM dorm room bull sessions.
Yeah. Normally it's the stuff of 2AM dorm room bull sessions.
My favorite variant of this is a story from Stanislaw Lem's book "Star diaries", written decades ago. I chose my nickname from the name of the hero of that book, Ijon Tichy ;-)
One of my favourite authors :)
Another somewhat related short story - "The Last Question" by Asimov:
http://www.multivax.com/last_question.html
Another somewhat related short story - "The Last Question" by Asimov:
http://www.multivax.com/last_question.html
great writing. everyone should read it and try not to just scan through it or you will miss a very valuable lesson.
absolutely delightful!
Professor Nick Bostrom's site on the Simulation Argument: http://www.simulation-argument.com/
He also has some other cool stuff, like this "Letter From Utopia": http://nickbostrom.com/utopia.html
He also has some other cool stuff, like this "Letter From Utopia": http://nickbostrom.com/utopia.html
I've thought about this 6 years ago. The sim could even be paused and resumed by the creator whenever he wanted, and we'd never be able to tell there were actually many gaps in our lives that range in billions of creator-life-terms.
What about the morality of simulating an environment where sentient beings are produced by a long agonizing process of evolution? Greg Egan points that out here: http://gregegan.customer.netspace.net.au/PERMUTATION/FAQ/FAQ... (question 6). "Post-humans" should have room for a much more robust morality that would preclude someone purposely creating the world we live in - unless we were an unintended byproduct of an unintended byproduct that no "post-human" has even noticed or something.
The theory becomes is even more convincing if you realise and accept that it needn't just be humans that have to create sufficiently powerful computers. We could be part of a program running on some alien planet. Planet Earth, human beings and all that we've ever known and loved, don't need to have ever existed. All it requires is for some intelligence, somewhere to build a pretty ultimate PC/Mac, whatever.
At the moment we know of only one definition of 'existence'. By that definition everything you mentioned - the earth, humans exists.
this article is like that idea I get just before I wake up or as i'm soaping my ass in the shower and think it's brilliant. Then I think about it for oh ... maybe a few hours and realize I was a big idiot for even beginning to think that I understood the concepts involved.
I then usually eat a sandwich
I then usually eat a sandwich
The "20%" hunch kind of made skim from then on, but:
What's the definition of computer, exactly? Even without any magical circular reasoning, the computer built on atoms and physics that we're being simulated in is just as deterministic as the computer I'm using, so I'm not sure I even see a difference.
What's the definition of computer, exactly? Even without any magical circular reasoning, the computer built on atoms and physics that we're being simulated in is just as deterministic as the computer I'm using, so I'm not sure I even see a difference.
You guys make me feel soooo not smart at this point.
Time to migrate back to reddit where we can see things we understand. Like LOLCats :D
Just what exactly is an "ancestor simulation"? To simulate human minds, you have to simulate the universe, since that's where human minds get input. And you can't simulate a universe within a universe that's just like it. Not in 50 years, not in any number of years. Most crankbabble of this variety manages to at least be "not even wrong", but this is just plain-old wrong. Shame on NYT.
I don't mind.
No we're not.
Yes, guys, it was me all around. Funny that it took you so long to find out. I kickstarted this big simulation as a weekend hack and let it evolve for a month, so I could come and play as Warren Beatty.
Can we now, as a universe, move on?
This is not science; it's religion, philosophy, or abuse of language. You could never prove anything about this kind of stuff either way, and IMHO throwing around probabilistic estimates is 'not even wrong'. The scientist quoted in the article admits such estimates are only 'hunches', so I guess only the reporter is to blame for placing this stuff under a 'Science' header.
Now, I'm not protesting that we talk this. It's fascinating stuff. It's funny to recognize in this thread some of the same crazy ideas that would get people shuffling away from me at parties, like 'what if our whole known universe is but an atom inside other universe, which in turn [...]?'
Re: worlds simulated inside worlds, aka the Matrix, aka Plato's cavern- when I did work on a MMOG we had this conversation once or twice that if we gave our non-player characters ability to learn and communicate for an arbitrary length of time, eventually they'd come up with theories like "Every object in the universe has an unique ID, which is a 64-bit integer". And they'd wake up after a downtime and comment things like "gee, finally capacitor recharging is fixed!", and then stop and wonder about the deep meaning of that.
Re: internet and cybernetic consciousness: when I first learned about the internet, one of my first thoughts is that the interconnected computers would create a new consciousness, or an ecosystem of competing consciousnesses. Not that it would be pursued, but that it would happen spontaneously, just like life evolved out of the primordial soup.
Then I thought that something similar would happen as a result of connecting so many people. Now, it's been always like that to some extent (cultures, memes, ideologies, etc. can be thought of as supraindividual consciousnesses), but the instant communication would make it more ostensible and powerful. Just like oxidation and explosion are instances of combustion, yet the difference in speed makes it counterintuitive for us to recognize any sameness between them.
Fast-forward to 2007: for about a year, the global hivemind has been thinking intensely about cats with captions. Who knows what deep intentions hide behind this?