Physicists find we’re not living in a computer simulation | Hacker News Reader

Physicists find we’re not living in a computer simulation

cosmosmagazine.com

42 pointspmcpinto65 comments

This argument is underwhelming to me. Physicists found that simulating a few hundred electrons (using one particular method) would require more atoms than there are in the (known) universe. This is a good argument that we cannot simulate our own universe precisely, but it says nothing about the underlying nature of reality.

Consider if Pacman took some time to study physics. He analyzes Ghost behavior and realizes it would take more than 255 levels worth of data to describe it. Pacman concludes that he must not be in a simulation, because clearly there are only 255 levels in the world.

Not to mention that if you can solve the equations involved you do not need to use the Monte Carlo methods.

We cannot yet, but someone else might be able to do it.

Heck, even having a good approximation might be enough to entirely change the computation class. (c.f. A* to Dijkstra given provably good heuristic)

I never bought into the theory, but I don’t think this is a good proof. I won’t pretend to understand the scientific details of their argument, but it breaks down if you assume that the creators live in a universe that does not follow the same physical laws as ours. Our universe may be simplified (possibly by necessity) and could be incapable of creating full recursive simulations of worlds as complex as our own, but that says nothing about the original universe that its based in.

Like any arguments about the existence of a higher being, it is impossible to prove that a higher being that doesn’t want to be found doesn’t exist. Just choose whichever theory you find to be more convenient, and live your life to the best of your ability.

In addition, this doesn't really prove that the "simulation" can't use extremely aggressive "level of detail" culling, only simulating such intensive things like the quantum Hall effect when it directly effects the events of the simulation, or when a particularly curious simulatee decides to measure it. When the simulation can afford to gloss over such intensive details, it could use some much less expensive approximation, and we would never be any wiser.

Same thing with the uncertainty principle: electrons are only an approximation until we take a good look at one!

If it's assumed that the "creators" live in a universe that does not follow the same physical laws as ours, that means the "creators" have the ability to run simulations of universes that are arbitrarily different from their own. In that case the chance that they chose a universe of our type to simulate, out of all the possibilities, is vanishingly small, and thus so is the likelihood that we are living in a simulation.

This seems like a strange argument. Suppose that the creators created some simulation. This argument would apply equally no matter which simulation parameters they choose.

Is there more here I'm not getting?

When they do simulate one, wouldn't the entities in that simulated environment make that exact same argument, and be wrong? If there are now two universes, that one and the real one, you've got a 50:50 chance. And I think we can rely on super-universes to simulate "interesting ones". Look at Greg Egan's works. Somewhere in the Super Universe, there is a Super Greg Egan saying "Imagine if quantum computers didn't work. What would that look like? Wow, they wouldn't be able to solve this kind of problem: in my novels, the protagonists will have to invent some kind of computation, and then they will run into this kind of problem and call it NP-hard. Poor bastards." And then some kid will read Super Egan's book and simulate, and here you are arguing it can't happen, because "random".

You don't have enough information to conclude anything about the simulator's universe just because they thought of this one. Consider wakamoleguy's argument about Pacman above: Pacman could apply that same logic to conclude our world is like his, but the simulation's rules just don't impose such strong constraints on the simulator.

https://news.ycombinator.com/item?id=12709430

This is dumb. Why does everyone (including these scientists) assume that the "outside world" would be of the same size and nature as the one we experience, and that simulation would use computational models we currently understand? Even if that WERE the case, a simulation this powerful and complete could also involve mucking with the internals of our minds to make us think there is a definitive proof against it existing. There is no "definitive proof".

Finding out that we are living in a computer simulation would probably be interesting for a few hours before everyone eventually realizes it ultimately changes nothing about our lives.

If we live in a computer simulation then there will be most likely bugs and edge cases and buffer overflows and all those things. So if we knew for certain that we live in a simulation then a lot of people would start looking for exploits.

I never quite thought of it that way. The first challenge would be mapping out the computational underpinnings of the simulation. I can't even begin to imagine how that might work, though I bet it would look something like physics.

This is a really interesting concept illustrated on this story: http://lesswrong.com/lw/qk/that_alien_message/

I'm imagining a glider in Conway's game of life trying to find a bug to exploit. What would that even mean?

I imagine the glider will try to figure out how to get around one of the 4 rules thereby changing its live neighbor cell count which in effect changes its evolution through board henceforth.

If those bugs manifest as miracles etc, then there are already plenty of people looking :)

Wonder what the language was to create the simulation, so one could have a more educated go at finding the exploits.

Correct, it's irrelevant. The scale of the universe, and where it came from, is already beyond our comprehension. We might as well just be living in a computer simulation, because ANY explanation would make just as much sense to us.

Hmm.. That seems like an awfully comprehensive explanation though...

If we know we cannot know.. then what do we mean when we say "know"?

It would depend on the simulation.

Are other people still real people, living alongside you in the simulation, or are they too simulated?

(Whether there's a difference depends on one's philosophy, mind.)

I've hypothesized that in a simulation, "others" in your world may just be permutations of "you" in some sense. That everyone you interact with is actually a creation spawned by "your" entity. It's one of the reasons dogs and owners frequently look alike -- because their origin is from the same logic, and that connectivity yields inherent similarities for entities created to have some degree of symbiosis.

It's crazy -- and certainly wrong. But so is all of this.

What's the definition of "real" here?

If our universe is deterministic there's no difference between existence and inexistence.

If you have a piece of paper with seed written on it then you already have in possession any moment of history of our universe.

And are we sure that the universe is not deterministic either way? How does that change anything in your day to day life?

Nihilism existed before any arguments about being in a simulation

On the other hand that would basically a proof for "God" which give me a real motivation to try to communicate with those running the simulation.

Suppose you knew your ant farm contained sentient ants and you comprehended everything there is to know about them from their biology to their behaviour, would you care if they wished to talk with you? I assume if you wanted to talk with them you would've done so already instead of waiting for them to figure things out and reach out

I think there would be many "people" trying to figure out how to get out. I know I would. If there is technology to simulate the world at this level, then there is probably technology to create a robot that could be filled with my conscience to inhabit the real world.

Step 1 - Find out how to communicate with the real world. Maybe some sort of quantum version of a buffer overflow?

Step 2 - Find someone to help us construct a body and transfer my conscience. Blackmail?

It won't be easy, but if we sims band together, I might make it out. I will send for the rest of you, I promise.

Just take the red pill

not only that but if we are able to simulate life in a computer, and we are a simulation as well, then most likely those who watch us on their computer's screen might be simulated as well.

Nevertheless learning that someone simulate us does not answer the question where the live came from... because those simulating us had to have a beginning too. And if they started in a computer simulation, then who is simulating them? And so on... its truly a question that can give you a headache!

> The researchers calculated that just storing information about a couple of hundred electrons would require a computer memory that would physically require more atoms than exist in the universe. Well, our universe appears to have found a way to store that same information on a couple of hundred electrons, so it might be that we find better methods going forward.

Your comment reminded me of this : https://qntm.org/responsibility

Metaphysics is pointless nonsense.

It's easy to be glib about this, but this is actually pretty interesting to me!

This trope shows up in so much sci-fi because it's a pretty quick way to undermine any philosophical questions. What's my purpose? To help run this simulation. Maybe I'll live long enough to find out when the outside computer shuts us down. It's never bothered me personally, because there's not a lot to be done about it. On some level, though, it's a though experiment that I didn't have a good answer to.

Of course, it's still unprovable: the outside simulation might have oodles more atoms than our universe, or different physics, or magic! But that's fundamentally less interesting than knowing there's a "higher" universe with roughly the same physics. That looks much less likely at this point.

It's really cool to read an argument that so strongly disproves something I thought was untestable!

but what if the "real" universe is 10^100 bigger than our universe? i mean, if you were simulating a universe, wouldn't it make sense to do it on a smaller scale? hell, the "real" universe might have way complicated physics than "our" universe, and the physics in "our" universe is simplified for faster simulation.

That's one argument that these "We're living in a simulation" arguments are confoundingly stupid. You end up with these unfalsifiable hypotheses indistinguishable from a quasi-Lovecraftian deism.

If it's exponential then even that size difference wouldn't help

No they didn't. They proved that one particular simulation method, not even an exact one you would ideally use for simulating a universe, is computationally infeasible in a universe similar to ours. Neither of those assumptions should be taken as universal, or rather multiversal. Maybe the simulation has more resources than we do (entirely reasonable, even from our perspective!), or meeting the simulators don't care much about how long it takes. Maybe they can fudge physics on those expensive cases, or maybe there's a better way to simulate them efficiently. At best, this damps down the infinitely recursive "we're almost certainly in a simulation" argument, but even then it's vulnerable to the fudging and better understanding/technique caveats.

I don't understand what people mean by "we live in a simulation". The word "simulation" to me (and according to Wikipedia) means:

> ...the imitation of the operation of a real-world process or system over time. The act of simulating something first requires that a model be developed; this model represents the key characteristics, behaviors and functions of the selected physical or abstract system or process.

What "real-world process or system" is our universe "simulating" using some approximate model? Isn't our universe the "real world", pretty much by definition?

Just as we can create simulations using software, it's possible some other entity has created a simulation that contains us.

A running process thinks the real world is constrained by CPU and memory and hard disks, but that doesn't mean it's true. Similarly, the universe is the "real world" from our perspective, but it's (currently) physically impossible for us to know otherwise if that is the case.

The caveat here is, even if we are a computational (or some other) process in some system, that doesn't mean that process is a simulation of another process. Just because there's a system A that "wraps" system B, doesn't mean that system B's processes can be called "simulations" -- simulation implies that there's a process in system A that the process in system B is trying to mimic.

It's ridiculous to state this is definitive proof that we are not living in a simulation. First, there are probably ways the simulation can fool us, on levels that we can't even comprehend. Second, who's to say that the simulation can't provide those levels of computation. Maybe the entire "universe" is but an atom in a larger computational universe.

I'm not arguing that the universe is a simulation, only arguing that I don't see this as definitive proof.

We're not living in a simulation, faster than light travel is impossible, time travel is impossible, and there are more than three dimensions. Obvious facts about reality that somehow escape most self described "smart" people.

I am no expert, but it sounds to me like this headline is an oversimplification.

The abstract of the paper suggests that the authors intend to support the claim that "not all quantum systems can be simulated efficiently using classical computational resources". Essentially that it's impractical to simulate complex nondeterministic quantum behavior using deterministic algorithms.

But what about quantum computers? It would seem to me like their inherent nondeterministic behavior would be an ideal fit for simulating other nondeterministic quantum systems.

In any case, it doesn't sound like they found any real theoretical limitation, just a practical one.

Why does matter whether it is simulated efficiently? What if each Planck time or frame takes a huge amount of "real" time to simulate, we would not know it.

Think of a Pixar render farm that takes hours to render one frame, the characters inside would not know that, their world unfolds in real time to them.

> But what about quantum computers? It would seem to me like their inherent nondeterministic behaviour would be an ideal fit for simulating other nondeterministic quantum systems.

true, but in this case now you either need a computer the size of the universe to simulate the universe in real time, or you can simulate a universe in (incredibly small) fractions of real time in a system less than the size of the universe.

this doesn't even get around the state-storage problem - storing the state of a simulated universe (assuming literally zero-overhead) takes the entirety of the universe too.

> In any case, it doesn't sound like they found any real theoretical limitation, just a practical one.

and therein lies the kicker :)

The original research publication[1] only seems to determine the possibility of complete QMC simulations on classical computational resources. While this shows we can't fully simulate portions of certain quantum systems on classical resources, it doesn't necessarily prove that a simulation of our universe is impossible, given the possibility of different physics in some "meta universe."

1: http://advances.sciencemag.org/content/3/9/e1701758.full

I like to reduce this to a German Tank problem: Let's say there are a few hundred hydrogen atoms being simulated in a supercomputer somewhere, and let's say that you are a hydrogen atom. You want to know: am I real one or a simulated one? Well, there are on the order of 10^78 real atoms in the universe, and a dozen or so simulated ones. So yeah, you're a real one.

In a more abstract sense, the number of degrees of freedom in any real universe is greater than the number of degrees of freedom in all the simulated universes it contains. So you're always more likely to be in a real universe.

this assumes there even are hydrogen atoms in host universe?

That was just an example. Just substitute "sentient being" for "hydrogen atom" and you'll see what I'm getting at..

> If the complexity grew linearly with the number of particles being simulated, then doubling the number of partices would mean doubling the computing power required. If, however, the complexity grows on an exponential scale – where the amount of computing power has to double every time a single particle is added – then the task quickly becomes impossible.

But you only need to _actually_ simulate the universe on very few occasions (a few times every few billion years) on a very small scale. Otherwise, you only need to approximate...

Phew.

Not bad. But hey there's always a chance the simulation skips the computation and just writes something plausible into the scientists' computers.

Why does everybody assume that the simulation must be real-time? Maybe 1 second of our universe takes ages in other universe. What a waste of an article.

"we live in a simulation" is precisely the same argument that Anselm gave in the 11th century as proof of the existence of God. I don't think that the people currently advancing the simulation argument are aware of the fact that they're at least a thousand years behind the game.

For what it's worth, Godel formalized Anselm's argument and found it relatively convincing.

"Our simulated agents are on the verge of self awareness and testing the boundaries as we speak"

"Quick! Whip up a plausible proof and disseminate it through their communications systems!"

I don't believe we're in a simulation, but I also don't believe that it's viable to prove we are not from inside the system itself.

How this related to the Universal quantum simulator of Feynman?

As I understood, it has been proved that it could simulate every quantum system with linear overhead...

https://en.wikipedia.org/wiki/Quantum_simulator

So, finally can we collectively say to Ray Kurzeweil "SHUT UP AND SIT DOWN" because frankly, this is like nought for six

no, there is no uplift no, for the third time, you did not predict ubiquitous AI around the corner correctly no, we are not living in a simulation

The thing that I find most frustrating about this thought-experiment is the one ridiculous argument that we statistically must be living in a simulation that pop-science keeps repeating (not this article, for once).

Here's the TL;DR: say you can simulate multiple universes in one universe, and at some point those simulated universes start simulating multiple universes, then you must end up with infinite universes at the bottom. So at any given moment you find yourself in a universe, statistically speaking it should be one of the simulated ones.

Which shouldn't take anyone more than a second to realise as complete bogus, because it's like saying "Let's say you have a cookie. If I break it in half, I have two. I can repeat this until I have infinite infinitesimal cookies. Since this is a much larger number of cookies, statistically speaking any time you have a cookie it must be one of the infinitesimally tiny cookies".

While not convinced by the statistical argument either, I think a more accurate analogy would be "Imagine you have a cell. It breaks apart into two identical cells. [Rinse and repeat a bunch of times]." And then if you pick a cell at random from the resulting set, it is indeed more probable that you picked one that is not the original one. IMO this doesn't apply as cleanly to simulated universes because they're not literal clones of one another, plus we're not even sure it's even possible to simulate one (while we know cells do go through mitosis).

The paper specifically uses a classical simulation. It does not claim that any physical phenomenon could not be simulated efficiently by a quantum computer.

This headline and article make an uninformed claim that is completely unsubstantiated by the paper.

Maybe this rules out the possibility that we're living in a simulation run on classical computers that naively simulate all physical processes. But that's not all the possibilities of living in a simulated universe.

they assumed the real world is anything like the simulation we're living in. we might be as close to resembling the real world as flat world is to us

Nice try Mr. Smith...

Rick and Morty already showed us that we're in a simulation within a simulation (possibly within a simulation) [1], duh.

1. http://www.imdb.com/title/tt3333830/

arXiv link to the paper: https://arxiv.org/abs/1704.03880

Pheè̷̢̲̮̩̣̱̳̔͒͠͝ͅe̷̮̜͙̲͕̲͎̿̋͌̽́̐ę̷̭̤̠̬͕̺̟̙̻͒̍́͊̅̓̐̅w! That's a relief̴̨̭͙̪̠̠̞̜̯͙̳̳̰͔̊͛̋̀͘͝.

Reply on news.ycombinator.com