Do our computers live in a simulation?

René Descartes is famous for countless things in mathematics—Cartesian products, Cartesian coordinates, Descartes’ rule of signs, the folium of Descartes. He is also famous for his work in philosophy and the notion of an evil genius. The evil genius presents a full illusion of a reality, and “fools” Descartes into believing there is a reality, while actually there is none.

Today Ken and I want to talk about the evil genius, and its relationship to the simulation hypothesis.

The hypothesis (SH) can be traced back far into history; even Descartes was not the first to wonder about reality and the possibility that it is an illusion. Recently the SH has surfaced as a “serious” thought. So Ken and I decided that we would comment on SH: did we really decide this on our own, or are we just part of the simulation? Oh well, who knows.

## The Simulation Hypothesis

The SH is that we live not in what we take to be reality, but rather inside some type of simulation. Indeed what is reality is a longstanding issue that has puzzled humans since the beginning of sapience. Descartes over four hundred years ago asked basic questions about mind vs body, which can be viewed as early questions on the SH.

Jump forward to 2003. A paper by Nick Bostrom gives a probabilistic argument for the SH. Bostrom argues that a future civilization would have enormous computing power—plenty left over even after the demands of Windows Edition 4532—OK, I added that. (Or did I?) Then he uses a Carl Sagan type argument: there are so many civilizations, that it is likely that some advanced civilization would run such a simulation. Robert Scherrer, Professor and Chair of Physics at Vanderbilt University and a Princeton 1981 classmate of Ken, wrote a science fiction story hinting that the argument can further be viewed as an instance of the Copernican Principle:

Earth enjoys no special cosmological privileges.

If there are so many simulations out there that the weight of numbers favors being in one, then it even takes special pleading to assert that we are not in one.

This argument apparently convinces many, but we are not convinced. It goes beyond the “brain-in-a-vat” concept:

—to assert that the mere processing of information feels like consciousness. You make your own decision. Or perhaps the simulation itself makes the argument seem correct.

## Detecting Simulation in Our World?

Now jump to 2012, or rather to last week. The paper Constraints on the Universe as a Numerical Simulation by Silas Beane, Zohreh Davoudi, and Martin Savage is an attempt to suggest possible empirical tests for SH. Roughly they argue that a simulation of reality would have to use a discrete grid type simulation of the kind physicists already use. But these simulations tend to make certain types of detectable errors, like the lamp labeled “Sirius” falling from the sky in the movie The Truman Show.

Their test turns on the fact that we human physicists do our own gridlike modeling of subatomic events as our way of being able to collate and interpret the high-energy data. Their argument seems to go that if Nature is “real” then we enjoy full oversight of the limitations of our models. But if Nature is being generated by someone else’s model, then the models could well “clash” in a detectable manner. In particular, they argue that cosmic rays would be registered as traveling across diagonals of our grids, despite our models being designed to register them along grid edges, and the rays would favor interactions in certain directions instead of a uniform pattern.

This strikes us as unconvincing on two counts. It seems first to claim that the super-advanced civilization would employ the same type of algorithms we use, just on faster hardware, and hence would be liable to same kind of modeling errors—which would tend to clash with our own like differently angled sheets of polaroid. But what if they had better algorithms? Second, what if their simulation already subsumes the kind of models we—the ones being simulated—are allowed to make?

Still, we buy the argument that a fortuitous proof of something tantamount to simulation is possible. Consider strings ${R}$ of (say) 50,000 “truly random” bits generated via quantum processes at John Walker’s HotBits service, or at Random.org. Suppose someone happened to find that ${R}$ can be generated exactly from a seed of length 1000 bits by a previously-known program. Some “universe-is-a-computer” proponents predict this kind of pseudo-randomness, while in a “real” Nature as standardly understood in quantum mechanics and relativistic causality, the processes must be truly random. Ken wrote a post on this blog on how these issues might be probed in a more subtle manner, where ${R}$ initializes a tabulation-hashing scheme for exponential-time programs that happen to be playing chess.

However, we can investigate similar issues being played on a battlefield right in front of your eyes, on almost every computer in the world. Let’s do that next.

## Simulation As A Defense

Right this second, all over the world, there are countless on-going attempts to discover whether “they” live in a simulation. Yes, we are serious. The issue is that the “they” are not people, but programs. These programs are battling to discover whether they are being simulated.

The context is the battle between dark hats attacking computer systems with malware, and white hats defending them. One method for trying to stop malware is to use simulation. Actually hardware people prefer to call it emulation, rather than simulation. We will stay with simulation to synch up with what the others say.

Simulation is used as follows to help detect malware:

1. A piece of code ${C}$ arrives at a computer—it could be your phone, your tablet, your laptop, or your server. Rather than just run ${C}$ and hope that it is okay, almost all systems check out ${C}$. One method is to look to see if ${C}$ is some known malware. But assuming that it is not, a good strategy is to simulate ${C}$ in a controlled environment—sometimes called a sandbox. Essentaily, simulate ${C}$ instruction by instruction.
2. If during this simulation the program ${C}$ does nothing strange, then it is declared okay and is run it in the normal mode on the real hardware. If it does anything suspicious during the simulation, it is declared malware. And the user is warned, or other remedies are taken.

This sounds like a pretty powerful method, which should be able to detect lots of dangerous malware. It does work quite well in practice. But. The “but” is that there are numerous ways that well-written malware can detect that it is being simulated.

Hmmm. This seems very close to the problem of detecting whether or not we live in a simulated world. So how does malware detect it is being simulated? Let’s look at some of the popular ways that it can do this. Security terminology is that there is a “failure of transparency” if there is an aspect of behavior that allows malware running on the simulator to distinguish it from real hardware.

## Defeating A Simulation Defense

Suppose again that ${C}$ is our malware. When it runs it is being simulated an instruction at a time. Of course the simulator is smart enough so we get the same results as if we were run in a normal mode. If this were absolutely correct, then there would be no way for the malware to detect the simulation. But getting absolutely the same results is not that easy. Here are some tricks to see a difference, a violation of transparency.

${\bullet }$ Use Time: One idea is the malware can try and see if it is running slower. It could get the time, do some computing, then get the time again. If the time elapsed is larger than it should be, then ${C}$ has detected the simulation. Clearly, this is possible to block: have the simulator “lie” about the time. But, of course, this is quite delicate. The simulator must return the exact correct time, even a small error from what the malware expects could be revealing. It is possible, it can be done, but it is not trivial.

${\bullet }$ Use External Values: Another idea is the malware can ask to examine external values. They can be from other web sites or even just local information from some files. Again the simulator can “lie” and fake the values. Again this is delicate and must be done in ways that fool ${C}$.

${\bullet }$ Use Bugs: One more idea is that the malware can exploit the fact that the real hardware has “bugs.” If the simulator does not know about a bug, then it will give the wrong answer to ${C}$ and it will reveal that a simulation is in progress. Modern processors are very complex, so getting the instruction by instruction simulation correct is again very delicate.

In all of these ways, the malware is detecting clashes between its model of the true machine and the simulator’s imperfect model of the machine it is trying to defend.

## More on Bugs

Even fifty years ago processors were complex and hard to simulate correctly. When IBM introduced their first IBM/360 mainframe in order to make it easy for customers to switch to it, they included a “simulate 1401 mode.” An IBM/1401 was then a popular machine that many companies used for their computing needs, such as payroll. IBM engineers had a special feature built into the 360’s that directly simulated the 1401 instruction-by-instruction. This allowed customers to run existing payroll code, or other programs, on the new machine, thereby making the transition to the new machines easier and hopefully increasing sales.

A problem arose. Customers began to complain to IBM that their new 360’s machines did not work. In particular they did not run the old 1401 programs properly—the payroll checks were often incorrect. IBM of course quickly investigated this, looking for an error that they must have made. There was no error. The IBM engineers had exactly simulated the 1401 instructions. Well almost. The issue was that many customers used certain “illegal” instructions, ones that were not officially supported by the 1401, in their programs. They used these instructions to increase performance—some illegal instructions helped make programs go faster.

Even though the IBM engineers knew the the exact specification of a 1401, they did not imagine that they would need to handle illegal instructions exactly as the actual 1401’s did. This was quickly fixed. Clearly, even if the processor is yours, and you have full documentation, it is not always easy to simulate it perfectly.

## Open Problems

The main question is, can we use the above and other ideas from malware attacks against simulators to help us see if we live in a simulation? I wonder of some of the above, or other ideas, could be exploited in the physics world. What do you think? And can we prevent an evil genius from gaining control of our computers?

[changed subtitle]

1. February 22, 2014 11:12 am

These are the forms of time, which imitates eternity and revolves according to a law of number.

Plato • Timaeus

2. February 22, 2014 12:01 pm

That Earth enjoys no special privileges is a philosophical position that denies observation.

Even inside the Solar System, Earth is very special, with a magnetic field acting as a shield, a stabilized rotation axis, a large Moon, an active plate tectonic and CO2 spewing volcanoes, and a huge core nuclear reactor.

These facts are related:
http://patriceayme.wordpress.com/2014/01/14/life-giving-nuclear-earth-reactor/

However, although it has been proposed that there are 40 billion planets in habitable zones in the Milky Way alone, nobody has pondered how probable and necessary it is for a planet to have both water for billions of years, and a very active core nuclear reactor.

Moreover, one has to take into account Galactic Core explosions:
http://patriceayme.wordpress.com/2013/11/06/40-billion-earths-yes-no/

Hence it seems likely to me that the Earth is very privileged, and we are the only civilization. That makes a simulation highly unlikely.

• February 22, 2014 12:36 pm

Agreed in the larger sense, but to define the Copernican Principle I chose the qualifier “cosmological” to distinguish away from such terrestrial considerations and also “anthropic” ones.

3. February 22, 2014 12:42 pm

I am not sure if using time to detect inconsistencies in the computing context is directly translatable to a path to prove SH. We don’t have anything to compare our physical universe to (i.e. a hypothetic universe running ours for which we’d know the timings), whereas as a malware creator, I can figure out if my timing measures are accurate with respect to the real hardware (known).

February 22, 2014 3:34 pm

I wonder about this as well. Is it possible to estimate the likelihood that the simulation uses different physics compared to one that uses the same physics as in the real (not simulated) universe? From the article, it might appear that changing physics is easier than trying to recreate them in a non-detectable manner.

Assuming it were actually different, then it would be sufficient to simulate only any simplified and consistent version that does not expose any behavior it shouldn’t, i.e. it would be undetectable, if I understand this correctly.

• February 23, 2014 12:28 am

I agree. The primary trait shared by all three listed malware detection strategies is the ability (for the creators) to compare known “real” hardware behavior versus a possibly different “observed” context. That process is fundamentally deductive. If we are not delivered a universe tagged a priori as “real” against which we can compare, then it’s hard to see how any observations would come to be categorized as “not real”.

• February 25, 2014 3:30 pm

Agree. And still, not only time, but how can we identify a bug? How to find a reference of the normal behavior of this big simulation called life? We will always be limited to the outputs/results of the program… Can’t find a trustable source of comparison

February 22, 2014 1:27 pm

If we are the “malware” trying to detect the simulation, we must also be lucky enough to exploit a “bug” in the sandbox that hasn’t been patched yet. And good malware stays dormant until after the sandbox trial period, and gets to run on the “real” hardware.

February 22, 2014 4:14 pm

Assume, if possible, that “my” comment” is not simply a banal subroutine and/or explicit malware (a strong assumption). I read somewhere that Descartes thought he existed but I haven’t ever read a satisfactory definition of what “pure thought” really is or where “interesting” ideas really come from.

As far as I know, Prof. Gödel regarded the human mind as a perfect instrument of reason and attempted to prove that such a mind is bigger than ANY machine. After this became known (malware’s worst nightmare), he openly speculated that the mind might be a computing machine attached to a SPIRIT, presumably not intrinsically evil. So perhaps the key project is to conduct a scientific investigation of (non-alcoholic, sub-material) spirits.

Is there a “completely convincing argument” (E. Bishop’s informal definition of a “proof”) that a purely mechanical procedure can simulate one-half of an ongoing “stimulating” conversation? I live in “another” world–very far off (course)–and haven’t received news that a machine passed the Turing test (say, for example, during a conversation with Pierre Deligne where “they” discussed how to compare apples to oranges and whether such a comparison might yield valuable insight into an efficient method for radically simplifying Wiles’ proof of Fermat’s Last Theorem).

More to the point, if this world is so perfectly artificial, then how could a concept such as “The Genuine Article” come into existence? Or is the belief in such a concept a kind of malware… perhaps “belief” is simply a string of zeros and ones written on the Möbius strip that I’m treading upon… (while listening to Judy Collins sing “I’ve looked at life from both sides now””). Chomsky must be proud that he recognized how “they” manufacture consent (its one type of sociological malware).

6. February 22, 2014 6:44 pm

ah man no discussion of this topic is complete without some obligatory/gratuitous mention of the Matrix! :?: … so looks like you played it safe with just the blue pill.
yes there are some [sensitive inferometer] physics experiments being done to determine if we “live in a matrix”….
personally think the universe is built out of solitons aka the Hindu Purusha/Prakriti :smile: (physics simulation experiment links in that blog)

February 23, 2014 12:14 am

The malware idea is very interesting. However, it seems to me that the core question remains the same — can we deduce something about the “actual universe” based on the universe we live in? Or at least take a “reasonable” guess? It’s a problem with which malware doesn’t need to deal.
Maybe we’ll only gain insight on SH when we have means to simulate a universe of our own, and let it resolve its own SH… :-)

February 23, 2014 12:23 am

In regard to the hypothesis that human reality is a simulation, my two favorite works of fiction — or … are … they? — based upon this theme are Stanislaw Lem’s short story Non Serviam (which appeared in Lem’s collection A Perfect Vacuum, 29 reviews in GoodReads) and Ian Watson’s chess-centered novel Queen Magic, King Magic (three reviews in GoodReads).

Lem’s story is fun; Watson’s book is even *more* fun (as it seems to me), and perhaps Ken Regan might the latter almost as much as Gödel’s Lost Letter readers might enjoy Ken’s recent Chessbase column “Ken Regan’s Tandem Pawn Chess” (a Google search finds it) … perhaps some Gödel’s Lost Letter readers don’t know that Ken is an International Master and world-respected authority on computer chess.

To continie, it’s mighy surprising that no-one has drawn the (obvious!) economic conclusion regarding this hypothesis, namely invest in D-Wave.

The chain of logic leading to this conclusion is irrefutable prima facie

•  If the universe is a simulation, then it is a PTIME simulation. Otherwise no entropy difference could be large enough to run the computer!

•  If the universe is PTIME simulation, then it is a varietal simulation … as are the vast majority of existing quantum simulations (needless to say)

•  If the universe is a varietal simulation, then thermodynamics “just works,” and conservation laws “just work,”, and Emma Noether’s theorem “just works” — all subject to restrictions and caveats — but the superposition principle is held hostage by varietal singularities.

•  To cloak the universe’s varietal singularities — and thus preserve humanity’s 20th century delusion that the universe’s state-space is an infeasible-to-simulate Hilbert space — the masters of the universe specify a Hamiltonian flow such that varietal singularities are occult.

Are they evil for so doing? Or merciful? The world wonders!

• Conclusion  Varietal state-spaces of less-than-Hilbert yet more-than-classical dimension are large enough to fulfill all of D-Wave’s corporate ambitions — especially in universes whose dynamical flow is restricted to the singularity-cloaking gauge-interactions of the Standard Model — yet dimensionally small enough to respect the no-go theorems of BosonSamplers and ShorFactorizers.

To paraphrase Feynman, we already know enough math, science, and engineering to be entirely confident that There’s Plenty of Room in the Middle … dimensionally speaking.

Remark  Varietal algebraic dynamics naturally explains why the Standard Model looks contrived. The Standard Model *is* contrived … contrived to cloak varietal singularities that otherwise would elicit acausal informatic paradoxes.

In a nutshell  D-Wave can’t simulate precisely the quantum processes that the Standard Model can’t implement.

It all seems so incredibly obvious, Mandrake … after reading the imagination-stirring works of Stanislaw Lem … Ian Watson … and Joseph M. Landsberg too. All of which are GOOD!

February 23, 2014 1:18 am

Caveat  Occasionally I try to be funny, and invariably the grave, even worried visages in audience remind me of my limitations in this regard (as the 1907 cartoon Mark Twain in London illustrates).

The humorous story is told gravely; the teller does his best to conceal the fact that he even dimly suspects that there is anything funny about it.

A comment that grapples directly & drily with these tough-yet-topical dynamical issues — so that quite likely, some readers will find humor in it — can be found on Gil Kalai’s fine weblog, associated to Gil’s essay BosonSampling and (BKS) Noise Sensitivity.

• February 25, 2014 2:50 pm

————————————
Bulletin of the AMS, April 1, 2020
On the unexpected efficacy of SAT solvers

A new generation of SAT-solvers has revealed unexpected structure associated to Intel’s hardware random number generators (including both the 80802 Firmware Hub and its successor Bull Mountain): these bitstreams can be emulated by an GNU-standard MT19937 generator (“Mersenne Twister”) initialized with seeds found by SAT-solving. Early expectations that this phenomenon might be confined to Intel hardware have been confounded by the surprising discovery — confirmed in recent months by multiple independent laboratories around the world — that data records associated to photon-counting experiments invariably derive from GNU/MT19937 initialized with SAT-guessable seeds. Are we to conclude that the human experience of reality, in its entirety, is a computational simulation? What can we deduce about the algorithms and even the author(s) of the simulation software? And what are the implications (if any) of these experimental findings with respect to PvsNP? This Bulletin of the AMS special issue is devoted to a discussion of these questions, by a panel that includes Lance Fortnow, Bill GASARCH, Dick Lipton, Scott Aaronson, and the Quantum Pontiffs, with moderators David Deutsch, Doron Zeilberger and Richard Stallman.
————————————

9. February 23, 2014 12:30 am

It seems that every generation since the dawn of human history a few toddlers stumble out of Plato’s Cave and come back babbling this story.

Nothing New Under The Sun $(0NUTS)$

February 23, 2014 3:17 am

It seems to me that one of the best arguments for a simulation is Heisenbergs uncertainty principle, which has a lot of properties in common with the noise of a simulation run in finite resolution.

11. February 23, 2014 4:58 am

Iff P=NP, then we can live in a simulation? ;-)

12. February 23, 2014 5:00 am

Iff P=NP, then we can live in a simulation ;-)

13. February 23, 2014 5:04 am

Reblogged this on mt2mt2.

14. February 23, 2014 5:06 am

(Sorry I have wordpress problem.)

15. February 24, 2014 3:40 am

Quantum mechanics seems to be good sandbox. May be the dark matter is an indication we are simulated. Or may be this is level up puzzle for human zoo lab, where advanced civilization is experimenting what ingredients one need to add to this artificially grown colony of barbarian so they can solve more difficult problem. OR if you are more human centered, those are left overs of the preparation – they did not expect (did not care) human would be smart enough to detect this.

February 25, 2014 5:16 am

Reblogged this on abhineshwar’s Blog and commented:
Is Simulation Hypothesis (SH) real?

February 25, 2014 6:11 am

SH seems utter drivel to me. It’s intelligent design Jim, but now as we know it.

February 26, 2014 10:12 am

I thought it was the symptom of schizophrenia, to believe you’re in a simulation…

February 26, 2014 10:47 am

Serge

A good point. We thought should discuss since in the news. But of course you may be right