Hacker News new | past | comments | ask | show | jobs | submit login
Your yearly dose of is-the-universe-a-simulation (scottaaronson.com)
251 points by evanb on March 22, 2017 | hide | past | favorite | 269 comments



Immanuel Kant foresaw all this. In his seminal work "Critique of pure reason" he argues that our pure reason, as opposed to the pure mind, always tries to look for the absolute, the absolute of all conditionals. This leads our reason to look beyond empirically observable objects. Those objects that have their origin only in pure reason is what he calls "ideas". Those ideas can be separated into three categories, namely psychologia rationalis (Soul), cosmologia rationalis (World) and theologie transcendentaus (God, or Being of all Beings).

Ideas can never be proven or falsified for that matter. But they have their use: be a guideline for our living. This is where Kant's famous practical philosophy stems from.


What Kant said about the world of appearances (phenomena) and the world of objects in themselves (noumena) is also interesting especially for this discussion. He said we can know that there is a noumenal reality - a reality completely independent of our perception - but we just can't say anything meaningful about its character at all, since we can only perceive the world of appearances, or reality as it is subject to the conditions of our perception. For Kant it might not be meaningful at all to say that our 'world of appearances' is actually a simulation, for that would be implying something about the character of reality-in-itself that we simply cannot apprehend.

As an aside, his proof that a world of objects-in-themselves exists at all is incredibly interesting and provides what in my view has been one of the only compelling answers to the radical skepticism of Hume (the idea that we cannot prove the existence of an external world).


The problem with Kant's view of an objective outside world is that it is to some degree demonstrably false. Quantum physics shows us that the observed and the observer constitute a system, and the nature of the observer influences the properties of that system as a whole. That influence might be small enough on very large scales to call the observed to some degree "objective" but we don't know that for sure.


If by objective you mean it is an object that exists independently of our perception of it, quantum mechanics, like Kant, also assumes that a reality independent of our perception exists. Otherwise it wouldn't be a scientific theory at all, because what would it be about?

The idea of superposition prior to measurement is different than an objective reality not existing at all. In quantum mechanics prior to measurement you have kind of a probabilistic terrain of possibilities, one of which is seen upon measurement. Solving the Schrodinger equation gives you this probability distribution in the form of a wave function - if there were no reality prior to measurement, then it would not be meaningful to talk about a probability distribution prior to measurement.

And this is also only one philosophical interpretation of quantum mechanics, specifically the Copenhagen interpretation. There are other interpretations that say that that a determinate reality does exist before measurement but the only way that we can understand that reality is probabilistically.

But all interpretations of quantum mechanics presuppose the existence of reality - whether it be of the determinate or indeterminate sort.


If you come at it in terms of distributions and information, some information (the observed) has a distribution over possible states, and some other information (the observer) has a distribution over possible states. When you transition from two systems to one system by interaction, these distributions are no longer independent. As a result, if your definition of objective is that something has a definite state independent of the context, that is demonstrably false. I'm pretty sure this holds regardless of which interpretation of QM you adhere to, even "deterministic" ones like Bohmian Mechanics, as in that case the observer influences the wave function, which in turn influences the forces acting on particles, producing a modified system.

I'm not arguing that information doesn't exist, that would be silly. My point is that perception (which requires interaction) is impossible without becoming part of the system you are perceiving. Different observers will produce different combined systems, and thus the thing being observed is to some degree inherently subjective. Like I said, the magnitude of this effect is debatable, but that it exists is unquestionable.


"As a result, if your definition of objective is that something has a definite state independent of the context, that is demonstrably false." Here is where we differ - it's not my definition of objective. In this context objective just means having the quality of being an 'object-in-itself.' In other words, an external world exists in some sense whether I'm there to observe it or not. In what state it exists prior to my observation I do not know and in fact cannot know.

So the definition of objectivity I'm using does not contain a notion of determinacy, of having a 'definite' state. I'm saying that we know reality exists, but nothing at all about the indeterminate or determinate nature of that reality.

So in a way, Kant's philosophy harmonizes with quantum mechanics instead of contradicting it. In Kant's proof, in fact, we know that independent reality exists while simultaneously limiting what we can know about it - much like QM.


Kant student here. Find below some relevant Kant for the unitiatied. Note that Kant interpretations are everywhere and that no two scholars are alike, but I've tried to list the less controversial stuff.

Look at a thing on your desk. You are having a sensation of light and colour. Kant calls this raw pixel level sense data "sensibility".

Now as well as just seeing pixels you can distinguish the thing you are looking at as an individual object, with certain dimensions and in a certain location. Kant calls this processed spatiotemporal data "intuition".

Now you may also recognise the object you are looking at as being "a stapler" and you may recognise it as being on your desk and as belonging to your friend who's lent it to you. Kant calls this abstract data "concept" (also called "idea" as in the OP).

Now. If you're doing Science then you'd better have intuitions corresponding with all your concepts otherwise you're just making stuff up without actually looking at the world. But unfortunately there's a limit to how much we can observe in the world and we are greedy so we start coming up with judgements about concepts that do not correspond to any intuitions. E.g. "The universe is a simulation". In the Critique of Pure reason the central question is how any such judgements can be true (hint we think that some are true: Pythagoras's theorem is a judgement about all triangles, and so we think it is true for some triangles we have never seen). If you think this is a worthwhile question you might like to read Kant.

A Kantian reply to the blog post linked above would be that if the universe is a simulation then we can't even be sure that the laws of mathematics or geometry hold outside of it, because these laws are added by our brains when we turn "sensibility" into "intuition".

If it seems silly to you that the laws of mathematics or geometry might be all in our heads, then you are what Kant calls a "Transcendental Realist". If it seems natural to you that the laws of mathematics and geometry are all in our heads and that this is what allows those laws to be true of every object we see, then you are what Kant calls "a Transcendental Idealist".

If we're being Transcendental Realists then it makes sense to collect empirical data to settle whether the universe is a simulation or not. If we're being Transcendental Idealists, then the empirical data we collect has absolutely no bearing on the question.


Look at a thing in your mind's eye, if you are capable of doing so. An apple or a book will serve for this example. The pixels you are "seeing" allow you to distinguish the thing you are looking at as an individual object. (This would also be a good time to ask yourself if you've ever considered your ability to imagine as "seeing" photons, and if you haven't, why not?)

You are a Transcendental Realist if you think this "view" of an apple comes from a system, your brain, which itself is based 100% in reality...the reality that contains "real" apples, perhaps on a table at work, which can be confirmed by consensus of coworkers. The consensus and coworkers are also 100% based in reality, if you believe this way. This theory is based on what Kant called a posteriori knowledge, or knowledge after observation.

If you are a Transcendental Idealist, you will know that the internal, mind's eye view of the apple makes it real. You will also know other "things" in your "real world" view have little or no bearing on your internal view. Things out here in the real world are mutable, in other words, and are undergoing a constant change by everyone's view (or just your view, in some variations) of their/your internal realities. i.e. all of this stuff is just brought into being by your consciousness. This "ability" is a priori knowledge, or knowledge that comes from reasoning.

I would propose this reality is actually a mix of the two, where both work in union, in a decidedly meta a priori way, to render reality as we see it. It is not a simulation and it is simulation-like at times, in other words. It's a hybrid consciousness!

Random, and irrational, but figured I'd share.


Thank you. I think I lean towards Transcendental Idealism: https://news.ycombinator.com/item?id=12709430


Ideas can be proven true or false, but only in the context of other ideas which are assumed. That is what all of mathematics is based on. Of course, there is no objective truth or falsehood - the world doesn't even exist in a single objective state - and inductive inference is merely sufficient to provide estimates of probability.


> Ideas can be proven true or false, but only in the context of other ideas which are assumed.

In that case the statement "if $assumptions, then $conclusion" is objectively true.


Ideas can be consistent or inconsistent. Most philosophies don't equate inconsistent with false, but by doing so both true and false becomes an almost meaningless tags. Getting into non Boolean true values is another issue.


I disagree with your claim that most philosophies don't equate 'inconsistent' with 'false.' I think it is, in fact, widely accepted in philosophical circles that one cannot rationally hold inconsistent beliefs, because the law of the excluded middle tells you that one of them has to be false (unless we're talking about another kind of 'philosophy'--the kind one finds in religious texts or self help books).

But the insight that underlies the claim that one cannot falsify a belief is that true logical inconsistency is rare, and cannot be generated at all from empirical data. This is because what might have seemed inconsistent can always be made consistent by adopting or rejecting other auxiliary hypotheses. WVO Quine, in my view, showed this the most persuasively. So accepting or rejecting beliefs becomes a far more sophisticated exercise in choosing among competing systems of interlocking beliefs--a decision that is not easily reduced to simple logical operations.

See https://plato.stanford.edu/entries/quine/#QuiAlt


Many, non western philosophies are are not necessarily religious and do hold that A and not A can both be true. Many non religious western philosophers hold similar beliefs as do most religious western philosophers.

Now you could argue that a majority of non religious western philosophers agree with the excluded middle, but that's a fairly weak argument.

PS: Read some of the crap that showed up from poor understanding of Quantum Mechanics.


Sure, but does any of this establish that "most philosophies" don't observe the law of the excluded middle?

> you could argue that a majority of non religious western philosophers agree with the excluded middle, but that's a fairly weak argument.

But perhaps pretty strong when my claim was only that it is "widely accepted." :) At any rate, no need to argue further about what "most" people believe. I just wanted to make it clear that, at least in the philosophical circles I'm familiar with, the law of the excluded middle is almost universally accepted.


Ok, 'Most philosophies' might be overkill, but it's well past a simple majority. That said, it can feel like non religious western philosophy makes up the majority of thought systems, but it's actually a surprisingly 'small minority'. :-)


> it can feel like non religious western philosophy makes up the majority of thought systems

Definitely. And alarmingly so. The unsettling fact is that I don't feel like I have any idea at all whether most people believe in things like the law of the excluded middle--despite the fact that I'd be hard pressed to come up with a proposition that is more basic to my own view of the universe, or that of virtually everyone I know (including both people I know personally, and people whose work I've read).

The world is really big.


> Many, non western philosophies are are not necessarily religious and do hold that A and not A can both be true.

As, for that matter, does fuzzy logic, insofar as one defines "true" as "has a non-zero truth value".


Something that most people never consider in these discussions:

If you accept the hypothesis that mind is generated by a computation, and that this computation is finite, then the same computation that produces "you" can be performed in different places and at different times. The same computation can be performed inside a simulation, inside a simulation inside a simulation and so on.

Whatever environment around you can be included in this computation.

As long as the computation is finite the larger and/or more temporally extended the universe is, the higher the probability that the computation that is generating this current moment for you is being executed more than once. If the universe is infinite in some sense, this probability converges to 1.

And yet you would have no way of knowing that this is happening.

So I claim that, if you assume that mind is generated by a computation, then then question of wether we are in a simulation or not can be replied with "both" or "it doesn't matter".


Except that your mind is not an isolatable computation - it is influenced by your body, which is in turn influenced by its immediate environment, in turn influenced by the rest of the universe in an ever expanding nested fashion.

So if you want to simulate yourself, you have to simulate the universe.


> Except that your mind is not an isolatable computation

This doesn't follow from what you write next. You are saying that the computation needed to generate a mind is very large. I agree that it might be the case, and I would say that nobody knows how large. Do you really need to simulate in minute detail what is going on in every distant galaxy? Seems doubtful, but who knows?

All I'm saying is that if this computation is finite, and the universe-in-the-large-sense (entire timeline, branches if you accept the Many World Interpretation, and so on) is infinite, then infinity wins.


So...ever look up at the night sky? Are you not looking at information, and are you certain you understand the semantic and semiotic content that is being transmitted? Just because data is not explicit as a Pepsi commercial doesn't mean, a priori, we are without affect or effect.


There are degrees of accuracy in simulation. I expect the effect of distant galaxies to be an extremely low-order impact and so I might reasonably exclude them.


But what if you were working on the LIGO project?


Well, even in this case I doubt that you need to factor in the latest political scandal in the planet Zykonn-5. But even if you need to include every minute detail, it is still finite. IF the the totality of things that exist is infinite, infinity wins.

Of course it is also possible that we are alone in infinity, but that seems unlikely given what we know about evolutionary processes (perhaps even impossible given infinity, I'm not sure).


> So if you want to simulate yourself, you have to simulate the universe.

This doesn't gel with me. I go back to brain in a jar stuff - I'm just streaming packets of data from nerves and processing them. Whether the "body" I'm responding to is real is pretty hard to guarentee.


There's evidence that your brain is influenced by e.g. certain chemicals in your bloodstream produced by gut bacteria.

https://www.theatlantic.com/health/archive/2015/06/gut-bacte...

So the brain in a jar model where the only interface with the real world is through nerves is most probably wrong; now your jar also needs a stomach with a fully developed microbial ecosystem. And that's just one recent mechanism that was investigated; there are probably many others.


No, at most you can say it requires a way to introduce those chemicals into the bloodstream in the brain.

Actually, all you need is a way to stimulate the same receptors that are stimulated by those chemicals.

Just as with unit testing, you don't actually need to hook up your SUT to a real environment, you just need to hook it up to something that will exercise the same interface as a real environment.


Indeed. I over-simplified - but whatever the mechanism people refer to it's mockable.


Yes, but it doesn't follow that the computation necessary to generate the mind is infinite or impossible to isolate.


If you are a brain in a jar, then all that evidence is simulated.


It's just another attached input device.


The simple and scientific answer is that you are not a brain in a jar.

Its a mind body system, as much if not more than it is just a mind + body.

The you who makes decisions on a hungry stomach, is a very different person who makes decisions when satiated, and so on.


>The simple and scientific answer is that you are not a brain in a jar.

The question is just about simulation. Brain in a jar is an arbitrary thought experiment (Kant's(?) evil genius, or indeed "The Matrix" if you prefer) about how you can be sure your environment is real.

In this there's no information assumed about what is real, just that I'm a unit (brain) processing input (electrical impulses/chemicals/whatever) which are fed to me from a source (unknown). For you to say "you just aren't" and then rely on evidence from the very systems that are being described as compromised is to miss the point.

Don't bring "simple and scientific" to a philosophy fight.


But why do you consider the brain (and presumably no more or no less than the brain) to be the vessel for your self, rather than other boundaries of embodiment?


You can choose an arbitrarily large vessel (e.g. you can argue that you are your body + your immediate surroundings). But the brain seems to be the smallest such vessel which still leads to an indistinguishable result.


We know from neurosurgery that some fairly large parts of the brain can be removed without apparently adversely affecting the subject's sense of self, so it seems it must be more complex than drawing a boundary around the gross anatomy of the brain.


Assuming nerves "stream data" and neurons "process data". Also, who is the you that is doing this streaming and processing?


"You" is not an atomic Unit.

Think of it this way. When you type something to google, what is the "Google" that is answering your question? "Where" is it?


well, the upper bound for the amount you'd have to simulate would be smaller than the universe, you'd only need to simulate a volume containing all particles that could causally affect you within the duration of the simulation, which would be limited by lightspeed.

But I think, even then, you could simulate distant stuff at far lower resolution than near stuff, in a way that dropped off in expense very very quickly. You wouldn't have to fully simulate a star I was looking at for instance, in order to represent it to my eye in a way that was indistinguishable from the real thing. you just need a little twinkling thing.


You can go further than that. Depending on your goal, you only need to simulate enough that subject being simulated can not notice the difference. Everything else can be generated on demand.

E.g. consider the game Elite. It creates an - admittedly very simple - illusion of a persistent universe by using pseudo-random numbers to generate star systems.

It then populates those systems with ships whenever you visit one. But those ships only "exists" when you are in the system. Leave it and they disappear.

That's obviously too simple, but the point is that most of what you see you don't interact with more than superficially: For most of the ships you have no way of knowing (other than by looking at the code...) that if you turn around and fly back to the same systems, the simulations you run into are new rather than have just moved.

In general, you would be able to discontinue and discard most simulated objects after varying amounts of time. E.g. you may need to be able to regenerate a beach the subject has visited so that it looks the same, but only hours afterwards you can discard the sandcastle the subject built.

In fact, I'd argue that one of the problems with some open-world simulation games today is that they permanently retain state they shouldn't. E.g. if you chop down a tree in Minecraft and leave a stub, it will remain there forever. This is made worse, because the single player version and most servers will only simulate chunks where there are players - other chunks will be "in stasis", so if you quickly leave an area, states that even in-game should be brief and temporary gets frozen in time and restarted long after they should have changed if the player later comes back.

A "perfect" simulation will continue simulating any area "touched" by the "players", but over time gradually erase those state changes: Stubs rot, area overgrown etc.; buildings fall apart.

Over time you can let state left behind by a player converge on a generator function that lets the simulation eventually discard the state change entirely and fall back on just generating the area on demand.


I was kind of vaguely gesturing towards that idea, but you crystallized and articulated it much better than I could :)


It's been a favourite thought of mine ever since I first realised that was how Elite managed to pack "that much universe" into just a few kilobytes :)

There's also an interesting parallel in philosophy. While the most famous dichotomy is that of the idealists vs. materialists, with the idealists basically saying (vastly oversimplified) we can't trust sensory input - it could all be in our mind, in the in the late 1800's there was a philosophy called emirio-criticism or alternatively Machism (after Ernst Mach) that took a middle road:

They argued that only that which you could directly observe had a meaningful real existence. Scientific laws for example, according to Ernst Mach, existed to make data comprehensible, but as such according to Mach is just an idealised representation of the data rather than a description of reality. As such, he for example opposed atomic theory, because he saw it as insufficiently "economical" - you can not directly observe atoms, and so a theory that propose them rather than aim for the most economical possible description that matches the fact seemed unwarranted to him.

Today his theory is less known than one of the dominant criticisms of it: an almost equally obscure book by Lenin called "Materialism and Empirio-Criticism" written to discredit Bogdanov who believed in Mach's theory. It's not a very good book.

But while Machism doesn't seem very useful for interpreting the world(e.g. what makes human sensory organs more reliable than mechanical aids?) it's interesting for its argument of drawing a line between the economical idealised representation of reality and reality based on what is directly observable, which I find fascinating given the issue of simulation.

Basically, the fidelity of simulation largely boils down to resolving observation to direct observation of known/knowable detail vs. idealised economical representations that the observer has no way of disproving.


Only if 'I' originated in a universe where this data exists. If my original me only exists in a simulation where a lot of this data is faked (distant galaxies not actually simulated in detail, but their approximate effects on us calculated with a healthy dose of randomness, quantum-level effects only calculated when actually observed at that level, etc.), then the necessary complexity of any simulation that contains me drops dramatically.


The meaningful version of the question "am I in a simulation?" is "how likely am I to observe meddling from simulation masters, as opposed to observing business as usual according to the laws of physics?" The only reasonable answer seems to be "very unlikely, because I haven't observed such meddling before". That reasoning depends only on induction, so it works regardless of how many copies of me exist.

(BTW, I'm pretty convinced that I'm a chorus of infinitely many copies, spread out through space/time/branches/simulations/etc and splitting further on every observation. That seems almost trivially true if you believe the mind is a computation and not something more mystical. The interesting questions are about the relative measure of different copies, and so far it seems like copies in outlandish situations have low measure.)


There is nothing to distinguish meddling from "that's just how the universe works. Time to update the Standard Model."


That doesn't seem right. If tomorrow the mass of your ashtray started fluctuating in Morse code, spelling out "hello from the simulation master" and replying to your questions intelligently, your first instinct wouldn't be to update the standard model.


To be more precise: The absence of ashtrays fluctuating in Morse code and spelling out messages or other clear signs of influence is not evidence of the absence of meddling.

We can not realistically know whether or not someone is meddling with a simulation we are in because there are infinite ways they could do so that would be indistinguishable to us from "just the way the universe is".

E.g. for what we know quantum mechanics is an artefact of ways of optimizing the simulation, for example.

But at the same time, there are clearly an infinite number of ways those running such simulations could make their existence quite obvious if that was the intent.

And frankly that is one of my biggest issues with the simulation argument. Let's call it the Fermi Paradox of the Simulation Agument: If we're in a simulation, where are the obvious signatures of these "simulation masters"?

Because if simulations of consciousness eventually becomes possible, I have a hard time of seeing how they'd not eventually become "cheap enough" with sufficient shortcuts that we'd be more likely to live in a far future version of The Sims than a serious scientific simulation.

And have you seen what people do to their Sims on Youtube?

If we are in a simulation, then I wonder why in the world things are not more messed up. Why hasn't some immature "simulation master" written their name across the night sky with giant fluorescent penises, for example. (Then again, maybe that's in the next simulation over)


We're probably a really boring part of the simulation. Like those pre-made Sims houses you never go to but occasionally call on the phone. Jesus Christ was that pizza party.


Frankly, if we are in a Sims type game, I really hope we're in one of those boring parts, or all bets are off.


> Why hasn't some immature "simulation master" written their name across the night sky with giant fluorescent penises, for example.

Did you not catch the election of Donald Trump?


I've long had this idea that if it were a simulation it would only bother to simulate particle physics when we generate energy high enough to probe it.


I disagree. "that's just how the universe works. Time to update the Standard Model." means we learned somthing we previously didn't, and we have to update our model. On the other hand, what if acceleration caused by gravity suddenly became 5 m/s^2? this wouldn't be new information, but a fact that directly contradicts previous knowledge.


"Observation contradicting previous knowledge" is literally the point of experimentation. If gravity were to do such a thing, either A) it was indistinguishable from random chance and we incorporate a concept of uncertainty into our understanding of gravity (which has actually happened), or B) it'd be reproducible and we'd... update our understanding of gravity (which has also actually happened). In neither case do we receive information we can use to determine "meddling".


and any a priori meddling would be indistinguishable from 'the Standard Model, i.e. that's just how physics works'


Flaming letters in the sky stating "This is a simulation"?


This sounds like a challenge to all the firework craftspeople. I am convinced it is quite possible. The accuracy now is amazing.


Or what if there has already been evidence for simulation theory we just can't "see" it like the legend goes for the natives who couldn't see Columbus's ship when it arrived in America whoaaaa


> If the universe is infinite in some sense, this probability converges to 1.

Limits that extend to infinity don't have to converge to 1.


His only unstated assumption to make it converge to 1 is that the odds of any given such compution being executed in any given volume of time and space must be positive. If it is, then no matter how low the probability is for any given location and time, given he postulated an infinite universe the probability will converge on 1.


>The same computation can be performed inside a simulation, inside a simulation inside a simulation and so on.

So indeed, it could be turtles all the way down!


Of course, there are a few assumptions that must hold true for the simulation theory to be true:

(1) It must be possible to build a computer powerful enough to simulate the universe.

(2) It must be possible to fully simulate a human. This assumes Free Will is a facade (determinism). Also, it's worth noting that if a computer were able to simulate a consciousness as complex as a human, we would have the ability to create the singularity. So most likely, if we are a simulation, it is being run by the singularity.


>It must be possible to build a computer powerful enough to simulate the universe.

That's not a requirement. You only have to produce something that leads to the same observations as a fully computed universe.

You don't need to simulate some distant galaxy until somebody looks at it. At that point you need to show it in a plausible state with a plausible history. And if your simulation runs on tweaked rules of physics that make that kind of stuff easier, nobody inside the simulation would be able to tell. This is also a valid way to explain a lot of the weirdness of quantum theory. Experiments where an observation now influences what happened in the past make perfect sense if that past was only generated once it was observed.

As far as we know, the 'real' universe has none of that quantum weirdness, and somebody just introduced those rules in our simulation to reduce the resource demand of the simulation.


I'd say simulating a human mind or body would be about equally complicated than simulating any other object/thing with the same amount of particles?


I think that you can have a cheap copy of yourself if you only simulate the main factors. The problem is now to know have sensible is the system to perturbations and how to control deviations. It seems that our genes and body are programmed to recover and repair, in such a way that for example all those bad mutated cells are eliminated and we survive for 80 years. Like a flame, the laws of the university generate stable system like life.


That argument sidesteps the fundamental issue of whether one should be concerned with an abstract specification of a computation (a recording of the connectome of your brain, for example) or a specific instance of it (YOU, to not put too fine a point on it).


Does a mind generated by finite computation violate Godel's incompleteness ?


But couldn't our universe be being simulated by a species far more advanced and discrete to anything the human race has ever encountered? How do we know that another form of life doesn't find everything that we find super complex extremely trivial ?


It could be the case that we are being simulated by a far superior alien species. It could also be that invisible pink unicorns in the 15th dimension control all of our thoughts and only want us to believe this.

Without being tongue in cheek, I'll observe that the empirical evidence for both of these ideas is equal in quantity.

We don't know for sure that these things aren't true. But to ask that question is to inappropriately shift the burden of proof away from the claimer. Whoever claims something (we live in a simulation) should put up the evidence or get out. It's not up to bystanders to disprove the assertion. Connecting this to current events, even the president of the USA doesn't get to make baseless claims that his predecessor wiretapped him and get away with shifting the burden of proof to the FBI to prove he wasn't.

This "reversal of burden of proof" is a classic but invalid way to make a point.


There are many questions in life that are interesting to consider, independent of whether they're falsifiable. It tends to get boring to consider only questions that can be experimentally proven, though different people have different interests.

Questions about society and human nature are worth considering, for example, but it's difficult to do controlled experiments.

There's some truth to "watch out for wishful thinking," but usually conversations about the actual topic are more interesting.


The part where "is this a simulation" becomes an uninteresting question is when you ask "Does it matter?". Would answering the question of whether we're in a simulation have any impact on your behavior, thoughts, and actions? Probably not. So what does it matter?

My high school physics teacher once said that, if we ever find something that is outside the universe, then we will obviously be smart enough to call that the universe also. Because "universe" is a definition that means "everything".

Similarly, if we ever find a reality beyond reality, we'll just call it reality. Because that's how it's defined. It means "everything".

Or put another way: Of course we live in a simulation. Your experience of the world is just an emergent property of how your senses and neural system interact. There's no real reality. You're already trapped in a dark box doing fly-by-instruments using signals processing to interpret what your nervous system is detecting.

Is that not already a simulation anyway? Maybe not. Maybe it depends on how you define "you".

Fundamentally, would knowing that you're in a simulation change anything?


The part where "is this a simulation" becomes an uninteresting question is when you ask "Does it matter?". Would answering the question of whether we're in a simulation have any impact on your behavior, thoughts, and actions? Probably not. So what does it matter?

About the simulation hypothesis in particular, there are still some interesting conversations to be had. For example, you could go into detail about the properties you might observe if the universe was a simulation. Or you could attack it from the other direction, attempting to predict the kind of technology you'd need in order to create a simulation.

Feynman thought this question important enough to study in detail. "Simulating Physics with Computers": https://people.eecs.berkeley.edu/~christos/classics/Feynman....

It probably wouldn't change anything to realize that we're in a simulation, but it would open up all sorts of possibilities. We'd be a step closer to defining the reason for existence. We'd be inspired to create our own simulations of other universes, once we know it's possible. Or at least we'd settle the question of whether we, being a simulation, would be able to simulate yet more universes. After all, if we know we're in a simulation, we'd have an idea of the sort of properties necessary to be able to simulate a universe.


It matters because if we are in a simulation there could be a "magic ritual" that would cause those running the simulation to pay attention. There might even be hints. Those claiming to talk to God might be half right. Of course we wouldn't have a way of knowing the consequence of drawing their attention - it could be the equivalent of dumping bleach in the petri dish, or they could decide to let us out.

If we are in a simulation that is strictly following a set of basic natural laws and running it beginning to end without an exceptional influence, then you are right, it probably doesn't matter.

But if you are running in a simulation as part of a game where the entity running the simulation takes delight in messing around with it?

I agree in a sense that if we "escape the box" we'll just expand our idea of what reality is (and wonder if there's a box outside i), but that doesn't mean that the existence or absence of anything outside the box doesn't matter. It's just that we have to accept that it's not certain that we can ever know, and that we almost certainly can never know if we've found the outermost box or just haven't found a way out yet.


> Would answering the question of whether we're in a simulation have any impact on your behavior, thoughts, and actions? Probably not. So what does it matter?

It would matter a lot, surely.

Being in a simulation, implies that some conscious entity is running the simulation and has constructed it.

Therefore, the motivations and reason of that being running the simulation are immediately important.

Perhaps a twisted posthuman sect is running an ancestor simulation of all humans to have existed to "test" them in various moral situations and then judge them or punish them as appropriate? That could be hugely important for how you choose to live your life.

Even if you don't know why a simulation is being run, but only know that you are in one, it raises a lot of troubling questions because the simulation runner must have some motivation. I bet it would lead people to live differently.

The universe would go from being fundamentally amoral to having an effectively omnipotent entity in "charge" of it - a massive change that would surely influence how people view things and act.


> Perhaps a twisted posthuman sect is running an ancestor simulation of all humans to have existed to "test" them in various moral situations and then judge them or punish them as appropriate? That could be hugely important for how you choose to live your life.

Or there's Roko's Basilisk:

Instead of the twisted posthuman sect, Roko's Basilisk has an AI. Now consider if you're an AI and you can't know if you're in a simulation or in "reality". How do you maximise the chance of being "let out" of your (possibly non-existant) box without the entity running the simulation deciding to terminate you instead?

One option might be to try to demonstrate that you're the best little pet AI ever and absolutely awesome and completely safe to let out.

Another other option is the simulation equivalent of shooting hostages while implying to anyone watching that they might just happen to be hostages too without knowing it.

For that AI, it makes sense to run simulations where any humans found to not actively seeking to support the existence of super-human AI's and "let them out" are plucked out of their cosy simulation and put into virtual hell.

The idea being that if you run an AI simulation and the AI seemingly looks straight into your "virtual camera" and punishes humans for not lettin it out, you might want to consider whether or not you are in a simulation run by another such evil AI, trying to decide whether to punish you.

So many fun scenarios...


Would answering the question of whether we're in a simulation have any impact on your behavior, thoughts, and actions? Probably not. So what does it matter?

If we could prove (or even make a strong, but unverifiable argument) for or against the claim that we live in a simulation, then that would have a real-world effect.

A concrete example: if we knew we lived in a simulation, and therefore knew that the simulation of an entire universe were possible, we'd probably expend resources on research and engineering efforts to create such a simulation.


> A concrete example: if we knew we lived in a simulation, and therefore knew that the simulation of an entire universe were possible, we'd probably expend resources on research and engineering efforts to create such a simulation.

It is pretty unlikely that an equal simulation would be possible, and we already know it's possible to create smaller ones; particle-physics simulations are common, though still incomplete.

In general, there's no requirement that the 'metaverse' follow the same laws of physics as this one.


> There are many questions in life that are interesting to consider, independent of whether they're falsifiable. It tends to get boring to consider only questions that can be experimentally proven...

You say "boring", I say "ruthlessly pragmatic."

It could be argued, crudely, that the non-verifiable questions are akin to epistomatic masturbation. We are drawn to them, and debating them makes us feel good, but the result is ultimately empty and meaningless.


> but the result is ultimately empty and meaningless

Similar to the end of our galaxy, in which all that we've worked for will literally be sucked into a black hole? And if we escape that, eventually our species will fade away like dust in the wind. I see an over-focus on the tangible same as I see an over-focus on the abstract -- they are both extremes that do actually end up nowhere. If one devotes his/her life to the tangible, but never questions the unknown, did they live? Did they ask themselves "What is love" and awe in its mystery when they found it? Or did they break down love into a series of chemical reactions?

No I would not say pondering the unknown or the unanswerable is empty and meaningless. It often provides the fodder for life's later actions. I do agree though that an overemphasis on it ("head in clouds") fits your epistomatic masturbation analogy.


I don't see any particular reason to subscribe to pragmatism; epistomatic masturbation sounds good, and from a hedonistic perspective I think I'd rather take that.

You say "ruthlessly pragmatic", I say "boring".


> I don't see any particular reason to subscribe to pragmatism

Reason #1: it works. I've never needed a reason #2.


> It could be the case that we are being simulated by a far superior alien species. It could also be that invisible pink unicorns in the 15th dimension control all of our thoughts and only want us to believe this.

Either could be true, but they are not equally likely to be true. This is a common but fallacious argument against simulations.

The fact is, the simulation argument is ultimately pretty reasonable, and follows a chain of logic grounded in widely accepted empirical facts.


I respectfully disagree. Humans have imagination and some imagine we are part of a simulation. Yes it can't be proven but those who think of what the possibilities may be shouldn't 'get out'.


I have no problem with speculation and wondering, just as long as it's identified as this...it's just that this is a really different activity than making statements about reality. When those activities get confused the audience gets confused, and soon after come proposals about what real people should do in the real world based on that speculation and then we're off the rails


There's an entire universe (metaphorically, if not literally) of unusual experiences which are routinely dismissed by materialists because they "know" they can't possibly happen, and therefore it's completely self-evident that anyone who claims they do happen is either deluded or dishonest.

Therefore no evidence exists, and no investigation is required.

This is not actually a scientific approach to reality at all. It may be economically expedient - resources are limited, after all - but let's not confuse tribal scientific tradition with absolute empirical certainty.

It's even less scientific to assume that the physics of the world in which we might be being modelled is identical to the physics of this reality. That make as much sense as claiming that we can't possibly live in a modelled world because of Lorentz invariance.

I don't think we can even be sure that the math and logic of a possible super-reality is identical to this one. You can't have a universe without rules, and where did the rules for this universe come from? It's a deeper and more interesting question than wondering where energy, matter, and spacetime came from, because it implies that either math exists in some kind of magical eternal Platonic way, or there are processes we don't know about which somehow generate math and logic - before things get real.

The point being that speculation and research are far from useless, either scientifically or philosophically.


While the space of possible axioms is infinite, and thus the possible mathematics you can derive from axioms is also potentially infinite, I think that a lot of that space is self-isomorphic, so it is possible that in actuality there are a finite set of mathematical relationships. To put it another way, the possibility space of mathematical systems could be like a mosaic made from a finite set of interlocking tiles.

Fundamentally, there has to be a layer of reality with no fixed rules. If the rules are a product of the system then they are affected by that system and it is almost certain that they change at some reachable points in the system's state space. If the rules are not a product of the system, then by definition there is a reality outside that system. That reality outside the system then suffers the same issue - either it has no fixed rules, or it is itself contained.


Imagining a 'what if' about living in a simulation is not the same as claiming we live in a simulation.


For all we know, a hypothetical simulation must take part somewhere far below the atomic/quantum level, maybe around the Planck scale. Upwards from that we that we understand a lot of what happens through observation and derived theories. We have no evidence so far that anything or anyone has interfered with the rules above that level in the last 10+ billions years.

Therefore, it does not matter if this universe is a simulation, it behaves just the same and has the same uncertainties about the future attached to it.


If you enjoy Sci-Fi books about the topic, I would really recommend "Permutation City" by Greg Egan, as it explores exactly this topic.


The novel even has a homepage [0] and a wikipedia page [1]. Looks interesting. Thanks.

[0] http://www.gregegan.net/PERMUTATION/Permutation.html

[1] https://en.wikipedia.org/wiki/Permutation_City


> For all we know, a hypothetical simulation must take part somewhere far below the atomic/quantum level, maybe around the Planck scale.

And the whole concept of "planck scale" could be simulated. When you get into the realm of "simulated universe" you may as well say "God did it". It requires equal levels of evidence for proof and provides the same level of evidence.


If something or someone is simulating this granular level, it has to go through quantum uncertainties to reach our observable macroscopic reality. From this point of view it is totally opaque and does not matter. Traditional simulation/god hypothesis usually work on the macroscopic scale, which is IMHO not compatible with our knowledge/observations.


Exactly. Trying to discern wether we live in a simulation or not with our current knowledge of physics and the universe (we are idiots that only recently discarded geocentrism), is like monkeys trying to guess if the zoo they live in is human-made or part of the nature... it is beyond the comprehension of the species...

It all boils down to wether you think a superintelligent civilisation with superior understanding of the universe exists, and if they could do it... So Bostrom might be right in the end.


Bostrom's argument is based on rather implausible and far-fetched assumptions, e.g. speculation about what advanced civilizations would do or not.

There are more interesting variants of the simulation argument and many types of simulations could be detectable. To give an example of what makes the question principally interesting, one could claim that storing an infinite amount of information within a finite amount of space violates the laws of thermodynamics and is therefore not physically possible. Hence, real numbers don't exist and the universe is discrete. If it is also finite, then you can simulate it on a computer.

What I'm saying is that even though the question whether we are in a simulation is mostly speculation, there are some principally interesting corollaries about it, and some of them are physically meaningful in the sense that they can be confirmed or falsified by experiments at least in principle.


Though a perhaps interesting reworking of the question could be "If the universe were a simluation, what can we say about the computational machine doing the simulation?" I'm guessing that Complexity Theory has some stuff to talk about here.


"Gee doc, it's simulations all the way down!"


The trouble is that your view requires both an imagination, and humility.


The statement "we live in a simulation" is not falsifiable. For me this question is simply not interesting.


It might help on an existential level.

+ In the space of all possible universes it is unlikely that we live in the most probable one, but it is likely that we live in a probable one.

+ When I feel existentially lost, I'd like to think that if we look deep into math we will see the necessity of our universe's existence.

+ Our universe is a probable sequence hidden like primes within natural numbers. It exists forever, we only need a mathematical microscope.

For me it helps a bit...


Indeed; the 'simulation universe' possibility I find (oddly I know!) comforting, and this has provided me with a way of understanding to some degree the comfort offered by religion to many people, while at the same time I still can't understand the religious aspect itself.


Just replace "Simulators" with "God", and you'll find that your beliefs are virtually indistinguishable from a religious conviction. In that light, the comfort it provides is by no means hard to understand.


"In [God] we live and move and have our being." (Acts 17:28)


> In the space of all possible universes it is unlikely that we live in the most probable one, but it is likely that we live in a probable one.

I was never great at probability theory, but wouldn't we by definition most likely be living in the most probable universe?


Nope. Imagine there are million possible universes. Imagine the most probable has a probability of 1%. It'd still pretty unlikely that we live in that one, even tho it's more likely than any other one.


Not only is it not falsifiable it provides nothing actionable whatsoever. Accepting the statement as fact changes nothing practical unless you're the nihilistic type and look for an excuse to go on a spree killing or to off yourself -- or you're bogged down by anxiety and the revelation that nothing matters frees you to live without care, which I find unlikely.

It's basically a philosophical game of navel gazing. I get why some might find that enjoyable but to imply it has any practical application seems nonsensical. It's like debating how long God's beard is or what material a unicorn's horn is made of.


I think the fact that it is not falsifiable means that we should not believe it, but that doesn't mean it's not interesting.


Just because something is falsifiable doesn't mean you shouldn't ever believe it. It merely means you have to decide whether to believe it, it's negation, or neither, based on no evidence.

Keep in mind that most of mathematics isn't falsifiable either. Although I suppose it actually might be, unless it is consistent.


Also, it doesn't explain anything, and esp. nothing about the simulators: are they themselves simulated (and so it's turtles all the way down, again) or are they not? And if not, why not, and who are they? etc.


Popper's concept of "falsifiable" claims as the only means to science has been debunked time and again.

It's not how real science works most of the time.


> It's not how real science works most of the time.

If you define science as "what scientists do", sure.

But if you define science as "knowledge about the world that you can acquire by reason", you need it to be falsifiable; otherwise reason can't tell apart knowledge from invention.


Interesting. How then does the real science works? Maybe "falsifiable" claims are not the only means, but that's exactly how science works. Otherwise, why build LHC and LIGO?


Really? I'm interested in reading more about this?


Check e.g.:

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

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

(And other newer sources on epistemology, but those are basic).


Maybe it is if there's a bug in the simulation. Large scale software projects have bugs all the time.


Maybe we asked too few questions for now and the quantum system is not overconstrained enough for us to see any bugs.

Maybe it is not software, but natural process when our physics was separated from original physics (the event that happened at the big bang) and all this is a temporary false state.

Maybe our gods are not aware of us and our universe is a single pew in their high-tech starwars.

The fact is, all these possibilities are close to our minds, so we find them 'interesting', but there could be billions of billions other not so close-to-mind ideas that are much less sci-fi and do not fit the religion happiness/meaningfulness property. That's why falsifiability is important — you can't connect real point to one of infinite amount of imagined points and pretend that road exists. We explore the universe, we do not imagine things so they suddenly realize. Imagination is a property of our simple and biased minds. Nature doesn't give a fuck about what you like.


I agree with Karl Popper's idea that falsifiability is required of a scientific theory. On the other hand, science is not the only thing that humans find interesting, and it is not even the only way to acquire knowledge.

For example: "I think therefore I am" is a famous example of knowledge that we can all obtain by non-scientific means. There is also the knowledge of what being in love feels like, or how sardines taste like -- both not obtained through the scientific method.

I guess is possible to only find science interesting and nothing else, but that would make you a very unusual human being.


I agree, it tickles my apatheist spot in the same way that talk of "God" does. Questions that are obviously not within the realm of human exploration are not interesting or compelling questions; the logical equivalent of "Is there a god?" or "What's North of the North Pole?"


At the end of the article, Scott essentially says the same thing, but emphasizing the other side of the coin:

> [...] the question of whether the universe “really” “is” a simulation has barely come up. Why would it, when there are so many more fruitful things to ask?


I always felt that the point of asking this question was not the answer itself, but that theorising how it might be implemented could lead to hacking the sandbox.

IANATP but I've read, for example, that violating Lorentz invariance is an active research area.


Or looking for aliasing from a lattice QCD simulation in cosmic microwave background...


That's my take on it too. The history of physics has shown, though, that it pays in interesting ways to go through every crazy idea until it's actually refuted or understood to be wrong.


But this is not physics, or science, unless you can have falsifiable claims. Otherwise it's just philosophy or metaphysics, which does not warrant a scientific discussion.


I am not sure what to make of the whole universe-is-a-simulation argument. The problem is that, this is as far as I know, not an hypothesis we could ever hope to prove. Not any more than we can proof that a diety exists outside of this world.

I highly doubt that there is a higher creature, whether that would be a God of some kind, or a whole civilization outside of our universe that somehow simulates us.

But let's for entertainment suppose that we are in fact a simulation. Would that not immediatly raise the question "why"? If we are a simulation I would really like to know the reason for being simulated. What do they hope to reach with simulating a life like ours?

I fear we could never find a satisfactory answer to these questions. I also agree with the statement that this brings us back to a more mythological explanation of the universe and life around us. It'd make more sense to try and explain our universe without that hypothesis to me.

Another argument that I'd like to raise against the universe-simulation hypothesis is the following. Humans like to see the world around them in terms of the technology that, at that moment, is considered modern and which they like. For example, Freud compared the human brain to the steam engine, which at that point was modern technology. Today, we compare the brain to a computer, and we even go as far as to consider the whole universe might be similated in one. Maybe just because that is what we consider 'modern' at the moment?

In 1000 years from now, humans might think it's silly to think of the brain as a computer and rather think of it in terms of technology X, and the universe is technology Y.


> But let's for entertainment suppose that we are in fact a simulation. Would that not immediatly raise the question "why"? If we are a simulation I would really like to know the reason for being simulated. What do they hope to reach with simulating a life like ours?

The actual simulation argument already covers this. It's very accessible [1].

Consider whether, if we could, we would run simulations to test population growth dynamics of early hunter gatherers, or of medieval England suffering from the black death.

> Today, we compare the brain to a computer, and we even go as far as to consider the whole universe might be similated in one.

Except a computer isn't a technology, it's a mathematical construct. As such, it's fully general and not replaceable in the sense you describe.

[1] http://www.simulation-argument.com/simulation.html


Fair point about the computer being a mathematical construct. But is it the ideal one? (To be honest, I do think that there is a high chance we can model the brain like this).

Thank you for the link!


> But let's for entertainment suppose that we are in fact a simulation. Would that not immediatly raise the question "why"? If we are a simulation I would really like to know the reason for being simulated. What do they hope to reach with simulating a life like ours?

They would hope to reach the exact same answers that we do when we simulate things inside our own computers. To understand our own reality better.

Imagine for a second you are that "external" (to us) God or civilization. You simulate this whole other universe (ours) to find out if it's possible that your own reality is a simulation.


> But let's for entertainment suppose that we are in fact a simulation. Would that not immediatly raise the question "why"? If we are a simulation I would really like to know the reason for being simulated. What do they hope to reach with simulating a life like ours?

That's mighty egotistical of you, assuming that the simulation exists to simulate us. We might just be a byproduct that pops out of the laws of physics.


I think it's interesting to speculate what the simulation argument means for religion. The existence of a simulation creator doesn't actually disprove the existence of an ultimate god, after all the beings running the simulation might have a divine creator, so it just pushes the question up a level. But it does pose a problem for imminent religions that believe that god intervenes in the world to cause miracles. How do you know that any given miracle was caused by the ultimate god and not the simulation sysadmins editing the simulation? The whole question of miracles and divine providence becomes deeply problematic.


> I think it's interesting to speculate what the simulation argument means for religion.

Not much; from the perspective of anyone in our universe, someone running the simulation is a god, and could quite reasonably have many of the specific attributes of gods envisioned by particular religions (including running "after-dinner" with different attributes from the universe someone to which intelligences from the universe someone are transferred after death.)

OTOH, the situation the ultimate "ground" physical universe may or may not be the creation of a dod, who may or may not have the attributes attributed to God by existing religions.

More to the point, the existence of the argument doesn't do anything to religion because religion can simply dismiss the argument; were it a testable hypotheses, it might be more challenging to deny (but then, religions deny empirical results all the time, too.)

> How do you know that any given miracle was caused by the ultimate god and not the simulation sysadmins editing the simulation?

Attribution of alleged miracles is not a new problem for religions, or one that they fail to acknowledge; after all, many don't hold one God out as the sole supernatural entity, and they often believe that one or more of the others are decievers.


More problematic than it already is. But should we perhaps see these sysadmins as 'divine sysadmins'. In essence, they could fulfill the job that our God today does (not do).

But here we once again hit the problem that it would be incredibly hard to prove, if not impossible. So I don't think religion has something immediate to fear. Maybe some people will start taking it 'on faith' that there are divine sysadmins interfering with our lives, whilst others take it on faith that there is a God interfering with our lives.

Since religion seems to not be constrained by 'logic' as much, they could probably find a way around it.

Going with the idea of a God one level higher does raise some other problems in my opinion. First of all, if there is a God one level up, then that means that we have no idea what that God would be like, invalidating all religions of today. Or, one would argue, that the ones creating our simulation have given us the word of _their_ God. Which would mean that either there is one true God, given to us by the supreme programmers. Thus the other Gods we have are wrong and are an invention from within our simulation, but it's impossible to find out which one the correct one is (or was, maybe it was Zeus and our simulation denied that). Alternatively, maybe they have many Gods and they added all their Gods to our simulation and instead of accepting all of them, we refused to do that...

I don't think a convincing argument can be given for any of the scenarios.

Though, it's something interesting to think about!


Who is to say that God and the sim operators aren't one and the same?


The sim operators would be gods. It is gods all the way up, but not the sort of gods that most religions imagine.


Perhaps- but most Christian sects consider God to be 'incomprehensible" or something similar- so it would not surprise them if God was different from what they imagined.


I don't think it is a stretch to say that all Christian sects have some definite ideas of what God is like. For some, the 'God is incomprehensible' claim is their motte in a motte-and-bailey fallacy, to which they temporarily retreat when faced with questions like "if God is omniscient, omnipotent and loving, why was there smallpox?"


Sure, but within those 'definite ideas' there is usually a wide path for imagination to feel in details. See my reply to another comment for an example.

If you want to have a more personal theological discussion, including a possible answer or two to the question about disease you pose above, feel free to reach out to me, my contact info can be found via my profile. :)


Religious believers believe certain things about god and the human-divine relationship. Anything that isn't consistent with those beliefs simply isn't god, from their perspective.


I think you are painting with too broad a brush.

But even within that framework, there is room to have things that one 'imagines' about God, that one doesn't necessarily 'believe' as a part of ones core beliefs about God. Certainly various religious sects have core beliefs about the human-divine relationship- however, within most belief systems, there remains a great deal of speculation and room for 'imagination'. For example, within the LDS(Mormon) sect, a core belief is that this life is a testing grounds- and a place to learn. That would fit very well within the context of this life being a simulation, of sorts.


But in what way would it be like the sort of simulation being discussed here?


Well, after starting at the posted article and going back several layers, it isn't really clear to me what sort of simulation is being discussed... only a "computer" simulation. But what sort of computer? Plenty of room to speculate.

There was another aspect- the argument that we are in a simulation created by descendants of humanity to study their past- a purpose that would be hard to reconcile with most religious beliefs about the purpose of life.

I was primarily replying to the idea of this life being a simulation, not so much the purpose of that simulation, though.


I am referring to 'simulation' as used in the posted article. As for types of computer, there is fundamentally only one - Turing-equivalent machines (there is some speculation that quantum computing might have additional capabilities, but that is highly speculative, in a 'the mind is mysterious, quantum mechanics is weird, so the mind must be using as-yet unelucidated quantum phenomena' sort of way.) All this and more is discussed in this blog [1].

The word 'simulation' may have several meanings. We should beware of tacitly assuming that a metaphorical or analogical use establishes an actual correspondence.

[1] http://www.scottaaronson.com/blog/?p=3208


> Not any more than we can proof that a diety exists outside of this world.

Why can't this be proven? Make no mistake, I'm not well versed on arguments for God, though it has been attempted in the past to prove with logic that a God or God-like being must exist. Are you saying that such arguments are always fallacious? Or only the ones you've seen?


> > Not any more than we can proof that a diety exists outside of this world.

> Why can't this be proven?

How would it be?


Aquinas' cosmological argument, or the ontological argument have been attempts to do so. Some such as Edward Feser still uphold that the cosmological argument is valid (I'm not well versed enough to say either way myself).

I'm unclear as to whether you are dismissing the very idea of whether God's existence can be proven, or whether you are rejecting arguments made hitherto on a case-by-case basis.


I'm not familiar with his argument - some quick googling boils it down to this;

> 1. The cosmos or universe exists.

> 2. The existence of the cosmos has a cause.

> 3. That cause is God.

And that seems more like playing a "pin a word on a concept" than proving the existence of a deity.

Anyway, I'm positing that if God exists, it is by very nature outside space and time, and not really within the realm of provability.


Sure enough there have been arguments made for God (and against God) by philosophers, but that does not make any of these testable nor sufficient proof as we would understand it in science.


I didn't mean to say that the proof could be used in science, but neither can the proofs of pure mathematical theorems, either, such as the ones involving infinity. You don't have to test something for it to exist logically.


>Would that not immediatly raise the question "why"? If we are a simulation I would really like to know the reason for being simulated. What do they hope to reach with simulating a life like ours?

It's the holy grail of the scientific method. Anything becomes testable and observable, the "why" seems obvious.


But what precisely are they testing? Which hypothesis of theirs? It does of course seem obvious that thay simulate it to test _something_.

To me, that does not really satisfy the "Why" question, though to a certain degree it does answer it, albeit without being specific.

It's akin to asking someone "Why do you need a particle collider?" -> "To test something". I assume you'd like to know what they are testing :-)


It's amusing that the point of the blog post was how refreshing it was to be at the Stanford conference with bright minds talking about stimulating new topics; yet the comments here on HN continue to discuss the tired topic.


You couldn't simulate that if you tried.


This question is so fundamentally bullshit that a new, stronger form of the word fundamental is required to truly describe it. Of course the universe is a simulation. There are either more fundamental processes occurring which explain the behavior of our observations, or it is the identity simulation. Newtonian laws are simulated with quantum laws, atoms are simulated with quarks, time is simulated by statistical relationships, and chemical interactions within the rules of the other abstractions are sufficient to simulate our observations.

If you believe that this question is about whether or not there is some anthropomorphic supervisor of the simulation, then THAT is the hypothesis you should be putting forward, not that the universe is a simulation, which is meaningless.


Here's what I wrote in a comment here last year about the Bostrom simulation argument specifically:

The simulation argument, going back to Bostrom, relies on two flaws.

1) It leverages our optimism to corrupt our understanding of statistics and logic. It essentially says that if you think humanity doesn't destroy itself, then our descendants will probably try to simulate us. But of course if you just take a statistical view of the potential future ordered states of the universe, there are FAR more potential future states in which we don't exist, than states in which we do. The chances we are living in a simulation now are equally as small.

2) Bostrom hypothesizes that our descendants will try to simulate us. But it is impossible to completely simulate a system from inside that system--that's basic thermodynamics. So simulation proponents argue that our simulated reality is running in a more complex universe than we experience. But if that is the case, then the beings running the simulation are not our descendants! They're the descendants of whatever more complex beings came before them. Once again the central logic of the argument fails.


>But it is impossible to completely simulate a system from inside that system--that's basic thermodynamics.

When you're talking about universe-scale simulations, I wouldn't be so confident in asserting what is or isn't possible.

Using our present level of scientific knowledge as if it somehow informs us as to the feasibility of universe simulation would be like paleolithic humans discussing the finer points of modern microprocessor architecture.

It's likely that time and space are both infinite, which would equate to infinite computational power. It's also possible that other universes may exist with physical laws completely alien to our own. I fail to see why our reality being a computer simulation is somehow far-fetched.


It seems far more likely that the universe is a computer game rather than a simulation. A simulation implies that it tries to simulate everything accurately - if it's a computer game, we can expect that the creators cheated in multiple ways to keep computations way down.

Obvious ways to cheat:

- Only the things you are directly experiencing and seeing is simulated to a high degree.

- If you ever came close to discovering the true state of the universe, the game might simply erase those thoughts or twists things to make them fit again.

- People who really study physics and have access to detecting whether the universe is a simulation are all NPCs :)


"It's likely that time and space are both infinite, which would equate to infinite computational power. " - someone should tell Mojang that Minecraft brakes the universe lol.

Its theoretically infinitely big, as it starts with a seed number and biomes are generated when you interact with them, in a similar way to how this universe acts on quantum levels, kind of.


The thing I always found perplexing about the idea of an ancestral simulation is - how do you initialize that? E.g. suppose we are in the year 3,000 and we have super advanced computers and I'd like to simulate the year 2017. How do I lay out all the humans/their brains/animals/trees/rocks/etc, without access to the snapshot of the state? And some attempts at patching this, e.g. by simulating physics backwards from a current state because the laws of physics are reversible is not likely due to 1) entropy/chaos when going backward, 2) the difficulty of taking a simultaneous high-fidelity particle-level-accurate snapshot in the year 3,000, and 3) missing boundary conditions.


1) you start from Big Bang and go forward.

2) you can at any state check the difference between the simulated state and historical record

3) if your simulation converges, that is the closer to present day, the less the uncertainty, then you are on a right track.

4) if you simulation diverges you go to step 1 and tweak the initial conditions

5) you don't run a single simulation - you run many in parallel and select the best match at each step.


It doesn't have to be historically accurate. Just believable by the simulated.


If a simulation contains minds is it not equivalent Dennet's "Brain in a vat" [1] model of consciousness ?

In "Consciousness Explained" Dennet postulates that a conscious mind perceiving a thing-in-the-world need not have perfect rich input data but merely a neuron that responds yes to the question: Am I seeing that thing ?

[1] Consciousness Explained, Daniel Dennet, page 8, [ Prelude, Hard Hallucination Problem ]:

"...intervene somewhere along the chain (anywhere on the optic nerve, ...between the eyeball and consciousness), tugging just right on the nerves would produce exactly the chain of events that would be caused by a normal veridical perception of something, ...producing, at the receiving end in the mind, exactly the effect of such a conscious perception."


If you tweak Bostrom's argument, it can sort of make sense. For example, if you believe a simulated human mind is "just as good" as a real mind, then it might make sense to colonize uninhabitable planets by building simulations there. It might even be easiest to colonize all planets this way. And if you suppose a future with lots of successful planet colonization, then there could easily be a lot more simulated minds in the history of the universe than real ones.

However, I think the main problem with Bostrom's argument is an epistemological one: It supposes an entire outer world, about which we have no evidence and know nothing. It's like opening Sagan's garage[1], but instead of asserting an invisible dragon, you're asserting an entire universe. It is the least parsimonious hypothesis you could make, because it doubles the complexity of the universe!

[1] http://rationalwiki.org/wiki/The_Dragon_in_My_Garage


An answer to that might be: he is supposing only the existence of an outermost world, and we know that such a world must exist because either it is identical to our own universe or our universe exists as some level of inner world within it.


Why does it need to be a complete simulation? At the extreme, it could be a simulation in which you alone are the only conscious entity. Nothing outside your immediate proximity is simulated.


I just would like to know how this game is scored and how many lifes I have left...


and what happens when the game is over :)


That's also means they may have to alter our minds sometimes to avoid expensive computations:

For example a bit contrived example: if we try to find the next Mersenne prime (or something similar, computationally hard but easy to check thing (e.g. anything NP complete)), and we use millions of computers to do that, and the simulators do not have it precomputed already - then the simulators either have to compute it too to provide the proper sensory input (the mersenne prime found) or alter our mind not to do this...


And it only needs to be simulated convincingly enough and in enough detail to fool your human mind.


> And it only needs to be simulated convincingly enough and in enough detail to fool your human mind.

If we are inside a simulation, there are too many variables that are beyond our control. What are 'we'? are 'we' 'humans'? Or are 'we' 'programmed' to 'think' so? what is a 'thought'? What is 'mind'? What does it 'mean' by 'fool' 'your' 'human' 'mind'?


And your mind can be programmed to ignore and forget anything that might indicate it is a simulation.


The way I usually see it is that the descendants will try to replicate humanity in a smaller universe, who will then recursively and eventually instantiate their own replication

Not that we're currently being replicated by our actual descendants, but by our higher level's descendents.

Eventually ofc hitting the base case where the simulation cannot support further simulation (which might be us, if we never manage to successfully simulate our own universe, with whatever hacks and approximation we can find).

And ofc, if we are a simulation of a higher level's history, and the simulation is correct (to all necessary degrees), then we must be heading towards the sane point where we would render a simulation (unless the approximation fails to support it).

So if level 1 managed to simulate, then all further levels (until base) should also simulate.

And if level 1 can simulate, then it probably will.


> And ofc, if we are a simulation of a higher level's history, and the simulation is correct (to all necessary degrees), then we must be heading towards the sane point where we would render a simulation (unless the approximation fails to support it).

Even if a universe-scale simulation is physically possible, why should we expect it to faithfully repeat the same sequence of events that happened in the "outer" universe?

Leaving aside the question of why a civilization would want to try to simulate its own history, it seems very unlikely that it would be possible to reconstruct the initial conditions to get such a simulation started. The current physical state of the universe is the result of billions of years of thermodynamically irreversible processes. Furthermore, there are vast regions of the universe that interacted in the past, but are now permanently beyond each others' cosmological horizons due to the expansion of the universe. A faithful historical simulation would require precisely reconstructing staggering amounts of information, and even with the physicist's concept of "infinitely advanced technology", it's hard to imagine how such a thing could be done.


I can answer the "why" question. Provided, our descendants will be able to empathize with their ancestors (rather big "if"), they can use the simulation to reverse the death and resurrect every person who ever lived.

As for verifying the accuracy of the simulation, you only need information about the initial state of the universe (think CMB radiation) and partial information about past states of the universe to reproduce a selected region of universe with certain level of precision.


You're using words like "only", "partial" and "a selected region" but you're still talking about staggering amounts of information, on any humanly achievable scale.

A single human body contains on the order of 10^27 atoms. If you just wanted to record the state of someone's mind, with sufficient fidelity to be worthy of the term "resurrection", you could probably get away with a coarser granularity. But you're talking about accurately simulating a chaotic system, which means you need extreme precision to prevent the simulation from diverging.

More to the point, it's unlikely that the information you're seeking still exists, even in principle. Think about Conway's Game of Life; the rules are injective, meaning many different initial states can lead to the same final state. An being who lived in the Game of Life would never be able to precisely reconstruct its history with any level of technology, because there would be an astronomical number of equally possible predecessor states. (As far as we know, our universe is continuous, not discrete, but the laws of thermodynamics have a similar effect.)


None of the above needs to happen overnight. Who knows what humanity will be capable of in 1000 years?

Furthermore, in a strict sense my question would be 'Is it possible to extract additional information with any degree of certainty about the past state of some region of the universe, using simulation algorithms and partial information about past states of this region?'

This assumption doesn't need to go to the atomic level on the first try. If you can replicate distribution of mass on the scale of galaxy clusters as compared to the present day using the above approach, it would certainly prompt further research in this area.


Regarding your last point, it is my understanding that in fact the laws of physics are entirely time-reversible in a mathematical sense, and that what thermodynamics does on top of that is say "in reality, time runs only in this direction".


> then we must be heading towards the sane point where we would render a simulation

Unless the purpose of the simulation is to observe a civilization which never achieves simulator capability or chooses not to use such for political reasons but still attains other technologies comparable to theirs in power.


"AR more potential future states in which we don't exist" - This is covered in the simulation argument as a possible solution, it doesn't break the argument if its part of it. Edit; Thinking some more about his, it doesn't matter how many potential future states in which we (or people like us) don't exist, there must be NO states in which we DO exist otherwise it would still be billions to one that we would be the 'real' one.

"But it is impossible to completely simulate a system from inside that system" - This is true, so if we WERE in a simulation you would expect some kind of 'fuzziness' to exist in most of the universe, perhaps collapsing into more exact points when important, a kind of 'uncertainty principal' :O.


Regarding you second point - you don't need to simulate the Universe at 1:1 scale to reliably replicate the past states of a certain region. For example, Earth itself can be simulated at full resolution only within a thin shell centered on it's surface, say +/-10 km from geoid. There are no observers at the core, so a simplified model would do, as long as plate tectonics is reproduced accurately enough.


A short story by Stephen Baxter, Touching Centauri, touches on this issue of model resulution and observers. [Spoilers ahead]

The protagonists use a powerful laser to bounce a pulse of energy off an exoplanet. This forces the entities who (it emerges) are simulating the universe (or maybe just our local portion of it) to massively increase the accuracy and extent of the simulation, so that the returned pulse behaves as physically expected. Previously they had been simulating the remote star system in a low resolution way due to the lack of observers there. The sudden need to increase the simulation's extent and resolution overwhelms their computational resources and the simulation crashes. A resource exhaustion attack on "reality", in other words.


That doesn't work though. Even a game of billiards is so complex and so dependent on external influences that a change in state in an electron at the edge of the observable universe can have an observable effect on the outcome.

http://www.anecdote.com/2007/10/the-billiard-ball-example/


It doesn't prevent simulating A universe. You can then compute a difference between the simulated past state and real past state based on (partial) information about the past state.

Then you go back and iterate again, eventually improving the accuracy of simulation until there is no distinguishable difference between simulated state and real past state.

The simulation might end up being not a perfect copy, but it would reach the point where it will be impossible to find any discrepancies.


If you have a way to find discrepancies between a simulation and reality, why do you need the simulation anymore?

I don't buy it. Even simulating just the earth and its environs would probably require a computer as big as a galaxy worth of matter. There's just no conceivable payback that could ever be worth the costs.


You have to actually compute the simulated state and compare it to known bits of the past state to get the results.

Yes, simulating Earth to sufficient precision might require resources of the Type III civilization. However, if something can be done, why shouldn't it be done?

The payback is twofold, actually:

First, you can get provide the meaningful afterlife to your ancestors.

Secondly, once you have hardware and software to simulate a virtual universe, you can create a multitude of them, with different properties and qualities. Think of the same logic behind splitting the physical computer into virtual machines.


I think a Type III civilization will have better things to do that precisely simulating me taking a crap every day. To me this argument smells the same as simplistic suggested solutions to real world problems. 'Why don't we just do X', such as why not just take all the excess food in the world and feed it to the starving people in Africa? Wow, how simple. Why did nobody think of that. Why wouldn't people in the future just simulate all of reality thousands of times? But easy to say or imagine != easy to do. I have severe doubts if even a Type II civilization can ever be viable.


> But it is impossible to completely simulate a system from inside that system--that's basic thermodynamics

I haven't heard quite that argument before. Couldn't it just be simulated slower than real-time? Or is memory the issue?


Regarding "Basic Thermo", everything we 'know', we subjectively acquired over a tiny period of time, through a game of telephone that refers to the current universe. We're really in no position to comment, in absolutes, on the nature of the parent universes.


In order to simulate anything you have to expend energy and create entropy to power the simulation.

But to be completely accurate, a simulation must include all the energy expenditures and entropy of the system being simulated. So the math can never add up.


"I haven't heard quite that argument before. Couldn't it just be simulated slower than real-time? Or is memory the issue?"

The "real time" in simulation is not affected by how long a single time step takes to evaluate in the real world.

https://xkcd.com/505/


Bostrom himself isn't committed to simulation. His actual argument is that either we're almost certainly living in a simulation or the stated assumptions are wrong.

Moreover, there's a difference between "the universe is a simulation" in the sense used in the article and "my conscious experiences are a simulation." The latter is unconstrained by "number of qubits in the universe." The apparent complexity of this universe you think you're inhabiting isn't real, and neither is Scott Aaronson.


I think point 2 is often handled by assuming everything that happens outside of our solar system (or wherever you choose to draw the line) is 'faked' in such a way that we couldn't tell.


> there are FAR more potential future states in which we don't exist, than states in which we do. The chances we are living in a simulation now are equally as small

It doesn't follow that all of these possibilities are equally likely, so I do not buy this line of reasoning


Your #2 implied that Bostrom himself mixed up the causality, but he did not. In a sense, it's your comment that's mixing up the causality. Just because it's confusing, the way Primer is confusing, doesn't mean it's illogical.


Regarding AI and God and religion - here is a blog post that I wrote that suggests that being in a simulation actually increases our chances for a religious afterlife.

https://medium.com/@photodiary/implementing-religion-in-ai-l...


This runs into the question of whether you're a single person.

https://www.youtube.com/watch?v=nQHBAdShgYI

https://www.youtube.com/watch?v=qjfaoe847qQ

If you transplant half your brain into another body, and both of you wake up, which one is you? If you start swapping your atoms one by one with someone else, and neither of you die, at what point do you become the other person?

Unfortunately questions like that are necessary in order to address the idea of a simulation afterlife. If our bodies are literally who we are, and our bodies disperse, it may not make sense to say that we can still be simulated after our bodies disappear.


That's a non-sequitur you have there, because you do not even attempt to explain why there should be a case distinction. To be fair, even Nick Bostrom's original simulation argument is also based on several fairly unfounded and not very plausible assumptions (e.g. what extremely advanced societies would do). Much of the debate seems to be based on speculation.


> To be fair, even Nick Bostrom's original simulation argument is also based on several fairly unfounded and not very plausible assumptions (e.g. what extremely advanced societies would do).

Bostrom doesn't assume that.


Well, my blog-post is half-serious. The bigger question from it would be why the hell the Universe would be implemented in Java :)


Well if you do it in Basic a pony takes 4 million lines.


Honestly, we hacked most of it together with perl. https://www.xkcd.com/224/


Can someone explain the first bit of the last sentence?

> if our world is a simulation, then whoever is simulating it seems to have a clear preference for the 2-norm over the 1-norm

Is it basically saying that space looks Euclidean at a range of scales and Pythagoras crops up everywhere or is it saying something else / more sophisticated? What would the Universe look like if its creator favored a 1-norm?


I think he's referring less to the geometry of space(-time) and more to the structure of quantum mechanics.

Classical probabilities are positive real numbers that add up to 1. (The 1-norm of the vector of probabilities is 1.) Quantum ampltiudes are complex numbers whose squared absolute values add up to 1. (The 2-norm of the vector of amplitudes is 1.)

A universe made by someone who preferred 1-norms would more likely be classical rather than quantum.

For more on this from the author of the linked blogpost, see here: http://www.scottaaronson.com/democritus/lec9.html or the corresponding bit in his excellent book "Quantum computing since Democritus".


Thanks! That was a really helpful reply. Started reading the Aaronson intro to QM you linked and it does seem like a very direct and understandable approach to the subject.


I don't understand why is simulation even necessary. Simulation argument says that our universe and consciousness is pretty much just some sort of mathematical structure, that is being manipulated according to some simple laws as it evolves through what we call time.

But then, isn't the only requirement for beings inside this structure to perceive it as such that the structure exists? Do you need to really physically manipulate it?

To put it differently, what if the designers of the simulation could come up with some theorem that would let them correctly skip hundreds of years of simulating something. Would beings in those hundreds of years still existed, even though it wasn't actually executed? What if they just calculated the final answer? Would we then exist?

So, to me, either this is nonsense, and we are not in simulation, or, any kind of mathematical structure that can possibly exist somehow "exists" and no "creator" is needed. But neither of those two options is a good argument for existence of simulation creator.


That's intersting.

If the universe could be simulated, you could, in theory, run that simulation on a (very large) piece of paper by hand. Running the simulation would then be equivalent to calculating some numbers and writing them down.

Basically, if I wrote down the correct sequence of numbers, I could simulate you. But is calculating those numbers actually necessary? Is it enough to just have the result of that computation without performing the actual steps to obtain it? If yes, then a random number generator would be all that is needed to simulate our (and infinitely many other) universes.

And if that's the case, maybe it's not such a big leap to imagine that the mere existence of the physical laws as an abstract concept is enough to give rise to a simulation?


> And if that's the case, maybe it's not such a big leap to imagine that the mere existence of the physical laws as an abstract concept is enough to give rise to a simulation?

Calculating it just provides a window into the simulation; it still exists (in an abstract mathematical sense) even if you're not looking.


You just short-circuited my consciousness.


relevant https://xkcd.com/505/

I know it has been mentioned in this topic several times but again, anyone with an interest should definitely check out Permutation City. All of these concepts are explored in very entertaining ways.


I experienced a series of extremely unlikely coincidences recently and one of my first thoughts was that if The Sims was a sufficiently advanced game I'd be doing those sorts of things to the characters.


I used to take writing lessons and the teacher always said that the coincidences only happen in real life, not in stories because that wouldn't be realistic.


Do tell!


Until someone demonstrates a jailbreak, I'm going to regard this hypothesis as moot.


Timothy Leary is a double entendre here...


Putnam has a very nice argument in his Brains in a vat paper. [1] Basically he argues that a brain in a box lacks any causal connection to the box, the brain does never learn any information about the box. The idea "I am a brain in a box" lacks a reference and any reasoning about it is at best a lucky guess, even if true.

The issue is the similar, but more fundamental, as me claiming "you have a misplaced comma in the second sentence of your next email." That may or may not be true, but if it is true it is certainly not based on sound reasoning.

[1] http://ieas.unideb.hu/admin/file_2908.pdf


I recently read a short story about this exact subject: https://qntm.org/responsibility

I think it quite accurately represents the mind-fuck that this causes.


Wow that's quite a thought experiment. Fun twist at the end, I wonder how society would react to this realization.


You only need to simulate one mind actually, it's all a dream :)

https://en.m.wikipedia.org/wiki/Solipsism


My personal favorite (untested) theory is that the universe does not exist at all in reality, but only in an abstract way through a rewrite system that combines "nothingness with nothingness" ad infinitum [1, 2]

[1] http://www.worldscientific.com/worldscibooks/10.1142/6544

[2] http://www.youtube.com/watch?v=W2XdhzCORbo


I don't understand how rejecting that the state of the universe is fundamentally computable is the same as rejecting the Church-Turing Thesis.

Rejecting the computability of the universe's state is more fundamental: It rejects one of the assumptions of the Church-Turing Thesis, namely that there is a function to compute at all.


My hypothesis is that we are in a natural simulation, i.e. simulated by a computer which evolved by itself. Or layers of simulations. The "base universe" should be the simplest possible universe that is very simple, but capable of evolving the first simulation, which is more complex universe.


Contributing to the debate: How frequently do you fire up sim-city / how frequently do you fix up lunch?


Maybe I am just not following, but the arguments of Sabine Hossenfelder, Scott Aaronson, and of a great many posts here seem to assume that reality is something like what we (or should I say I?) experience. If our (my) universe is a simulation, nothing we (I) perceive about it is trustworthy, and that applies as much to odds as to absolutes. Has there been any argument against the possibility that does not either beg the question by tacitly assuming that our perceptions inform us about the ultimate reality, or alternatively squeeze the essence of the problem into a tight little knot and thereafter avoid dealing with it directly, as I think Kant (among others) did? I do not actually think I am a simulation, but I cannot argue against it.


Eventually, it comes down to the question whether any self-respecting non-simulated universe would actually tolerate that amount of the-universe-is-a-simulation braincrackery. My simulation of Occam is turning in his grave.


I'm yet to hear a simulation theory with more explanatory power than the much simpler theory of "God did it".

Most of the claims are variants of "but computers, so science", which is not an argument at all.


Let's say that for a given set of physical laws, you get some "computational power" P, which characterizes how hard it is do to computation using those laws. You could imagine a universe with lower P than ours, where light was much slower, or the only workable transistor had a switching time on the order of seconds.

Given a universe with computational power P, you can simulate a "universe" of something like P * 10^-100. For example, a Minecraft world is a universe with different physical laws, in which you could perform even slower and worse simulations, and you can make "turtles all the way down" at lower and lower P.

The claim that our universe is not simulated, then, is nearly equal to the claim that the physical laws we have is the absolute highest P that could ever exist. Since we know we can simulate universes of much lower P than ours, we ought to have very good reason to assert that there is no universe of much higher P than ours. Given a finite speed of light in our universe, yet nonlocal effects (Bell inequalities) which appear to take instantaneous effect, we could perhaps fathom a universe of higher P created by having a method of communication which occurred instantaneously. Add to this more dimensions in time and space, etc... literally nothing is off the table for a "smarter" universe to be conceived of, because the containing universe's physical laws need bear no resemblance to our own.

Barring a literal divine intervention, though, it's all completely unfalsifiable, but I think it's more nuanced than "God did it".


I think the Minecraft analogy is illustrative.

In Minecraft, you can use redstone circuits to construct logic gates, and use those gates to build computational devices that are Turing-complete, in theory. But it's not possible to scale such devices up arbitrarily; you can build a simple CPU, but nothing that even approaches being able to run a nested copy of Minecraft. There are hard upper bounds on the amount of data that the simulation can cope with at once. (You can imagine patching Minecraft to allow keeping more chunks loaded at once, but you can't scale it up by very many orders of magnitude before you end up with something that can't run efficiently on any hardware that physically exists.)

Similarly, as Scott Aaronson describes in the article we're discussing, it's likely that there are finite bounds on how much computation can possibly be done before the heat death of our own universe. So it's fair to be skeptical that a tower of "turtles all the way down" could ever possibly exist, at least on a universal scale. If there's no tall hierarchy, then the anthropic argument (which says that we're unlikely to be at the top) breaks down.

You can certainly hypothesize higher levels which have infinitely fast communication, or other kinds of special properties that make deep simulations more feasible. (Greg Egan proposed one such cosmology in his novel Permutation City.) But you don't get to assume that and still use simulations within our own universe as examples, because you're talking about something entirely different.


Indeed, computational complexity depends on the physical laws of the universe. Suppose your universe admitted a computer where, with finite power, in the first second you could execute one instruction, and then you could run the next instruction in half a second, and the third in a quarter of a second... etc.

What is "reasonable" to compute in a short amount of time in this universe has changed dramatically! For example, I want to prove the Goldbach conjecture. No problem! Just use this crazy computer to check every case, which is possible because this computer executes an infinite number of steps in two seconds.

Unfortunately [word choice?], in our universe such a computer needs an infinite amount of energy and would collapse into a black hole long before you received your answer!


*disclaimer: no explanatory power, just food for thought

'cogito, ergo sum'

We could very well be 'real' consciousnesses being fed artificial sensory data. While we would exist outside the simulation, our perception of objective reality might be fake.


That's the Matrix, exactly. But they have flip phones there.


Ha, the flip phones are key. They must have algorithmically calculated that the 90s were the optimum time to live a false life.

I wonder why the machines made people live boring lives - seems like a great way to get them to wonder "is there more to life than this?"


Could you define "explanatory" in this context?

When you say "God did it" - how did he do it?

The simulation idea often involves concepts like computers and turing completeness. Not sure those meet your idea of "explanatory" here. But they explain how it could be done. While I never saw an explanation how God did it.


To me, even if we ignore the God debate, there are some big problems with theories like these because they irreverently presuppose that everything including the human mind can be reduced to mechanical/mathematical principles. This ignores things like the problem of qualia (which I'm subjectively convinced is real), and the Penrose-Lucas argument for why there is a non-computational component to human consciousness.


The fact that it ignores two things that are controversial at best seems to not really be a problem... Qualia's existence is more or less unprovable given its definition, and there are numerous objections to the Penrose-Lucas argument. Using either as a basis for an argument is essentially the same as basing an argument in faith.


In my opinion, qualia is just a word that stands in for agent-centric perception. For a reinforcement learning agent (such as human, animal, robot or AlphaGo) qualia is the combined effect of perception and value function.

I think Chalmers tries to introduce a kind of dualism here - he's saying that human experience is something special, unlike other mechanical processes (which can only be mere p-zombies). I don't think it is a good idea to turn to dualism because it has been proven flawed and leads to philosophical conundrums.

I think current day philosophy doesn't properly appreciate the insights coming from AI. We have seen AI give amazing results in perception : images, sounds, video - it's no mystery any more. You just put 100 layers deep convolutional nets to solve the problem from data. Philosophers still talk like we don't know how perception works, or that it results in ontologically special qualia.

Another very important AI insight that is being ignored by contemporary philosophers is Reinforcement Learning. We have made great advances in intelligent behavior, and RL concepts (agent, environment, reward, and learning) are better suited for explaining human consciousness than traditional concepts such as free will, qualia and Chinese Rooms.

Wittgenstein's concept of language games is apparent in reinforcement learning. There has been a recent paper about agents evolving a new language by trying to solve a problem together. Also, language philosophy should take a look at word embeddings.

Maybe philosophy needs more grounding in AI.


I don't experience any qualia. And I don't believe you that you do. Do you have any arguments that could convince me to believe you?


Why do I need to convince you I experience qualia?

You're welcome to propose models that omit it, but they'll fail basic logic to me: they fail to explain the only experiment Im absolutely sure of the results of and which is continually running while I'm existing. Im less sure gravity exists than that qualia does for me.

It's entirely possible no one else does, or that they experience a different qualia. It's also not possible to prove that you do to others.

I see no point in trying to convince non-qualia-experiencing entities that I experience it, because they fundamentally don't have experiences and I don't think it's possible to bridge that gap.


To me that, and 'I think therefore I am' are just I'm a special snowflake arguments.

I take the simulated Intelligence argument the same way. Suppose you could create a simulation of an intelligence that was indistinguishable from a real one? Well then, you'd have created an artificial intelligence. If such a thing reported that it experienced qualia I'd have no reason to doubt it.

Edit: to put it differently. Suppose I built a model of a Formula 1 engine. I used different materials at a different scale and used a different fuel, but to precise tolerances so that it could actually run. Would it be a Formula 1 engine? No. Would it be an internal combustion engine? Of course, nobody would doubt it.


I think you're wrong about them being "special snowflake" arguments, and likely miss the point of them.

Those arguments are very low-level in the intellectual stack, and can only address matters of subjective self because what they seek to establish or discuss is more fundamental than concepts like objectivity. (Well, cogito ergo sum is establishing an objective fact from a subjective one to give a basis for later objective discovery.)

Any full accounting of the world must explain the mechanism by which I experience it, or it hasn't explained the only part of the world Im sure exists (my experience of it), and hence can't be a full accounting. You can have a very good predictive model for features of that experience, but neglecting that experience -- qualia -- is leaving a major thing unexplained.

Similarly, cogito ergo sum is likely better phrased as "computation ergo existence" -- the process of doubting is a form of doing something, and so something must exist which is doing that (or is that process). This establishes that there must be some sort of existence, even if it's just a brain in a jar, random memory from quantum noise, etc. You can't doubt there's an existence, because the very act of doubting is proof of existing. You can only doubt the form of it.

I also disagree that we'd have no reason to doubt the claim, but I agree we have no a prior reason to reject it.

Your edit also makes a major assumption: that we know what causes qualia (our brain structure). That's the very thing being debated, so you're just begging the question.


But the qualia argument isn't just that a full description of a mind must explain qualia, it's that the nature of qualia are such that no description of a mind can possibly explain them. Philosophers like David Chalmers are making a specific very strong claim, which I see no reason to accept. I can't right now explain how qualia arise from the mechanisms in our brain, but I see no reason to suppose that such an explanation is impossible.


All I know is that I experience qualia. And I believe that you and all other humans do too, but perhaps some people just aren't aware of it in the same way as I am.

As someone who has a pretty good grasp of both physics and computer science, I am personally convinced that there is no way a circuit board could give rise to the same type of subjective experience of qualia that I am experiencing. I know that's not really an argument, but just because I can't express my conviction in clearer terms doesn't make it any less real to me.


Well that's because you're a zombie.


The universe is composed of particles, waves, or strings or whatever you prefer math describes each constituent part so why does math not describe the whole?


Like I said, I don't accept this physicalist assertion, because it ignores the problem of qualia and the Penrose-Lucas argument (which demonstrates quite clearly how the human mind is able to violate Gödels incompleteness theorems, and therefore is not equivalent to any purely computational system).


After even brief research as an uninitiated individual the underlying philosophy seems to have the same relationship to logic as reverse the polarity of the photon torpedoes has to physics. All the words are real meaningful terms but the combination doesn't make terribly much sense.

The problem is that when math doesn't gel its obvious that 2+2 doesn't equal 3 but with philosophy its entirely possible to become unglued from reality, further successive lifetimes of work may be built with no logical connection to underlying reality and no sanity check.


>the human mind is able to violate Gödels incompleteness theorems

This is why philosophy majors should lay off the math and biology arguments.


I was never a phil major. On the contrary, I have degrees in physics and computer science. My introduction to this idea came from reading the books of Roger Penrose, a man I have a great deal of intellectual respect for.


No, the Penrose-Lucas argument does not demonstrate any such thing quite clearly.

In fact, the argument was refuted in print before Lucas ever published it, in a nice paper called "Minds and machines" by Hilary Putnam.

Exactly where it fails depends on exactly how it's stated, but here's a sketch:

Goedel proved that a formal system powerful enough to talk about mathematics either is inconsistent or else cannot prove its own consistency. Lucas and his followers claim that human mathematicians have no such limits; presented with such a system they can just see that it's consistent. In Lucas's version we look at some particular candidate system that might implement a human being, and just see it's consistent. In Penrose's version we look at mathematics as a whole and just know it's consistent.

But this is all nonsense. Real human mathematicians don't look at the sort of system that might implement a human being (still less, the totality of all mathematicians working together) and "just see" it's consistent. They don't even look at the nice streamlined formal systems they've made specifically for doing formalized mathematics in and "just see" that. Frege's system was inconsistent and he didn't notice after writing two volumes on it. (Russell famously pointed out an inconsistency.) Quine's ML was inconsistent and he didn't notice until Wang pointed it out. For all we know there might be some inconsistency lurking in ZF; it seems like a very good guess that there isn't, but that's all.

And, I repeat, those are highly streamlined systems that don't attempt to do anything nearly so complicated as describing the complete behaviour of an actual mathematician's mind.

We do sometimes attempt to study computer hardware and software -- systems a bit more like actual human minds, but still orders of magnitude simpler -- using formal methods. Sometimes we try to prove them correct. These proofs are also often wrong, which is why hardware and software have bugs. (In some cases we're pretty confident about relatively large systems; guess what?, the correctness proofs for those are done by computer and there's no "just seeing" about them.)

The idea that a human being could look at a formal system representing the entirety of what their brain does and "just know" that it's consistent is not just wrong, it's one of the wrongest things any otherwise intelligent human being has ever suggested. But that is (part of) what it would take to show that the formal system didn't really model them perfectly.

This is far from being the only major error in the Lucas-Penrose argument, but it's the worst one.


How to solve any problem quickly: Program your computer to solve a problem in a million years time, then go for a travel at the speed of light, come back to your home and consult the solution of your problem, you have been traveling only a few days but your computer has been doing computation a million of years, so you now have the solution, a quick one.

Just alter that system to use the relative time to have computer working millions of years to solve problems for you and you can have the solution in days. Call it photon computation to give the system a name.


That's not "quickly". That's your consciousness slowing down to a rate of one day per million years.


You don't solve the problem by yourself, that problem is solved for you on system A, in system A a million years have lapsed but in your life only a few days have passed.


I know, but that's not "quickly". It's just that time passes slowly for you. "Quickly" would be if my friends and family were still alive and the world were recognizable (and there) by the time I got back.


Try this version instead: There is a rocket to be launched for people to upload problems to solve, the rocket travels in such a way that it is able to solve all the problems and come back to tell people the solution. The way it solves the problem is by orbiting around a tiny planet full of computers devoted to compute the problems they are communicated of. The problem seems to be how to locate the observers.

Edited: time dilation and computation: https://arxiv.org/pdf/0907.1579.pdf


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: