The Simulation Argument and the Simulation Barrier
michaelfeathers.silvrback.com
michaelfeathers.silvrback.com
And the very end of the day you have to take something on faith and proceed. Solipsism is internally inconsistent because it presupposes a mind that can evaluate a truth whilst simultaneously taking the argument that existence isn't knowable. Empiricism presupposes that our senses and mind can be relied upon to some degree. Theocratic foundations for selfhood are rooted in faith in a higher power.
So in a limited sense I agree with this article's argument. Once you allow that the simulation argument may be true you invalidate that your mind is capable of determining if this is likely.
That said, unless you believe your mind can evaluate what is true it has to be (as far as I can tell) a faith-based stance. That isn't much different from a religious world-view, except there's not a higher power. There's only you.
What bugs me in philosophy is that you can read theories and ideas (good ones), but no one asks if base terms are too much human to be important. Is there a school that doesn’t ob/subjectivize “items” and thinks as a whole?
This is a very different idea than believing that only the self is real. I don’t think it’s possible to strongly argue that. However, solipsism is a real case to make. Prove that your mind and body are connected and you will have debunked solipsism.
Mystics claim that it is possible to do something similar for this reality. It's all "I," where "I" ultimately points at something some might call "God." Hence the common instructions to look very carefully within.
>any posthuman civilization is extremely unlikely to run a significant number of simulations of their evolutionary history (or variations thereof);
[emphasis added]
>This lack of necessary overlap in metaphysics between the world that creates a simulation and the simulation itself could be called the simulation barrier.
[emphasis added]
Why would there be a lack of necessary overlap in metaphysics between the world an evolved organism inhabits and a simulation of its own evolutionary history? Bostrom's reasoning for why these simulations would exist in the first place clearly depends on extrapolating from various aspects of human behavior, so his argument necessarily posits that the beings running the simulations are at least in this way (and probably in others) similar to humans. The similarity of the Universe, largely, follows.
Furthermore, the simulation argument fails to be really different from the dream argument when we omit Bostrom's motivation argument. If we can be in a simulation within matryoshkaed simulations thousands of layers deep, then surely we might be in a dream within a dream thousands of layers deep. We can provide all of the necessary metaphysical computational capacity to the dreamer that we would a simulator, can't we? The whole bit where creatures are motivated to simulate themselves seems, to me, to be crucial.
It probably does for those who created the simulation. Not sure they have infinite patience to see the outcome of their experiment.
So our relative spacetime scale of a billion years here, might just be a few hours inside some of those crazy dense SUPER MASSIVE blackholes
And with that in mind and some loose sense our whole simulation might just took 1 second of the creator's relative time because...time is relative , just a loose assumption
This answer on Quora says the rate at which time passes on Jupiter is only a few nanoseconds off compared to Earth.
https://www.quora.com/Will-time-pass-slower-if-we-go-to-Jupi...
[!] Disclaimer: I don't know if that interpretation is still considered valid.
Edit: Formatting
Or try it the other way: simulator’s universe has 13 time axis and they can fast-forward on one by slowing down on one of the others, then reporting final results to all dimensions (and putting sinful souls in a special eternal suffering simulation supervised by worst criminals ever¹³).
There could be absolutely no feedback on the success or failure of an attempted escape until the post processing is complete (after the heat death of the universe). Escape might depend on some entity or automatic routine finding something interesting enough to warrant a second look.
To escape, you'd litteraly have to cause a universe wide anomaly, then social engineer your way past the ones running the simulation.
For example if I think there's a very low chance a valuable enough use for such a simulation could be found, and there's a low chance such a simulation would meet the criteria to be useful, and that there's a low chance such a simulation would need to be so accurate as to have actually conscious simulated inhabitants to meet those goals, then the chance that we are in a simulation comes to be vanishingly small.
Where else could it exist?
> wouldn’t a procedurally-generated universe alleviate the need for high-fidelity?
Could you elaborate on this, I'm not exactly sure what you mean.
We've all had the experience of dreams that seemed very "realistic", only to awake and realize that they didn't really make that much sense.
I'm not proposing this as the situation, but if we're talking simulations, it's fun to think about.
For example, in another part of biology, DNA seems to be a very inefficient as a data structure and tool for genetic replication. But you’d expect and forgive that from a system that self organized through chance, iteration, and a vast amount of time. Evolution has short horizons and lacks the long term planning required to be economical.
Human engineered systems are by contrast far more efficient with their resources. You can now sift through photos of cats and not_cats with a GPU and a few decades of collective technological progress. To get to the stage where a mind, artificial or natural, has self awareness, I can completely believe that it would require the level of processing only made available by a naturally inefficient evolution process.
We have minds that are focused on hunting, gathering, procreating, toolmaking, and surviving. Humans have succeeded at doing these things by growing vast brains that only accidentally became good at spotting incoming lions and finding mates. We are lucky that our complex brains evolved in such a way[1] as to allow conscious thought to emerge as a side effect.
Ok, you can assume an interstellar hegemonic civilization that can consume the resources of a galaxy, but if such a thing were possible how come we haven't already been absorbed by one yet? Ergo such a thing seems to be unlikely to be possible.
We could look beyond the limitations of thermodynamics, and consider that the contents of spacetime may just exist in the same way as the contents of pi or the Mandelbrot set exists.
Why must the universe be "created" at all by computation, when computation doesn't even seem to be a fundamental, but only an emergent property of it?
The computational complexity hierarchy could be different in a parent universe.
I do wonder though what things might have to be the same. I would be very surprised if any reality that allowed intelligent life forms able of creating and running simulations didn't have causality. What about the concept of number, or certain parts of logic? I think there may well be reasonable assumptions that can be made about a parent universe, even if they don't reach the level of complete confidence.
The choice of axiomatic system limits what is provable in mathematics, as demonstrated by Godel's theorem in the case of natural number arithmetic. The world may limit that choice of axiomatic system or the computability of theorems. There may be mathematical theories which are true but we can not prove within the physical bounds of our universe.
There's an xkcd comic titled "A Bunch of Rocks" at https://www.xkcd.com/505/. Could dropping rocks in the sand to simulate a turing machine produce a conscious being? If not, why can a computer?
Given that meaty goo can be conscious, and assuming that this arises from physical properties rather than some form of actual magic, it is reasonable to guess that the consciousness arises from what the goo does rather than from other properties such as pinkness or gooiness, so if we can cause another physical system to do the same things then it’s likely to have the same subjective experiences while doing it.
If a simulation cannot produce consciousness, what is special about life that makes it unrepresentable with physics? Where is the boundary between a conscious system and a non-conscious one?
Conscious minds certainly seem bound by their physicality, you can figure out how feelings and experiences are represented molecularly and electrochemically and affect those feelings and experiences experimentally. So far we've not found anything that cannot be broken down into physics (not to say we won't), but if we assume there's nothing else _but_ physics – then I think we're left with:
- The physics consciousness relies upon is fundamentally uncomputable
- Consciousness _is_ computable and all the weird consequences that follow (like sand computers) are true
I lean towards the second option, but I struggle to understand it in any intuitive sense.
What are your thoughts on why a simulation cannot be conscious?
So, consciousness being computable may not necessarily imply that any given computer could do so. One potential way for this to be the case is if consciousness is tied to some form of hypercomputation (https://en.wikipedia.org/wiki/Hypercomputation), which can provide outputs that are not turing-computable. This would mean that a traditional computer couldn't simulate consciousness, but that another hypothetical machine might be able to.
Bostrom is assuming there's a thing called "computation" which is somehow a fully-specified solved problem. (It isn't).
Worse, he's also assuming that this "computation" can somehow simulate something called "consciousness" - which unfortunately for him, is nowhere near to having the formal rigorous definition that "computation" as we know it today would require.
And finally after all that, you get something which is subjectively indistinguishable from life itself.
Well - obviously. If you don't define any of your terms and assume your argument is correct because you don't show your working or get into specifics, you can argue whatever you like.
For example, if I'm living in a simulation and I move my hand an inch to the left, it occupies an uncountably infinite number of intermediate positions during that motion. A Turing machine can't calculate or even represent all of those positions, since the set of Turing-computable numbers is only countably infinite.
Of course, what I've written above assumes that spacetime is continuous. Is it?
That can be tested by experiment:
https://physics.stackexchange.com/questions/9720/does-the-pl...
According to that link, spacetime shows no sign of discretization down to fourteen orders of magnitude below the Planck length. (The Planck length itself is so small that if an object that size were expanded so as to be a millimeter across, then a proton, at that scale, would be larger than the distance between the sun and Alpha Centauri.)
That suggests to me that while Turing machines can provide useful approximations to the behavior of physical systems, they can't represent or simulate them precisely.
Experimental evidence (the Fermi satellite data discussed in the link in the grandparent comment) seems to support standard physics.
Authors can write stories with believable worlds and compelling characters. Given infinite time, an author could detail how a character would react to every possible situation. But no matter how much they wrote down, their words would just be words. There isn't some breaking point where the author writes enough to create a pocket universe.
Why are turing machines different?
Just because the universe is governed by physics does not mean that physics are sufficient to construct a universe. I expect that "physicality" is a real property of some kind, which doesn't appear just because you described it accurately.
So much of this is guesswork, I may well be wrong—but it strikes me as the most logical conclusion.
absolutely nothing
> The physics consciousness relies upon is fundamentally uncomputable
This would be "huge if true", but I don't think it is. At least I haven't seen any indication of why it would be.
> Consciousness _is_ computable and all the weird consequences that follow
I think this is true, and I think the weirdness is a result of misunderstandings on our part (we are notoriously bad when reasoning about systems that contain us in some way).
> What are your thoughts on why a simulation cannot be conscious?
I think this whole consciousness confusion is due to mixing levels of abstraction on our part. Replace consciousness with other higher-level cognitive or social functions, like "love" or "patriotism", and see how silly it sounds. Is "love" computable? Can "loyalty" be reduced to physics? Yes, but the concept makes sense on a whole different conceptual level, which would be silly to reduce to, say, quantum physics, and is much better to reason about at the right abstraction level.
https://www.reddit.com/r/MachineLearning/comments/2xcyrl/i_a...
Your link is quite technical.
> "...the agent could be called self-aware."
"Could be called" being the key phrase. Yes, you can call such a model self-aware, because indeed it incorporates a model of itself so it is self-aware in a very literal sense.
However, there's no reason to believe that a model with the ability to represent itself would have conscious experience of the kind humans do. Why should we expect it to be any different from any other ML model or computer program, just mechanically executing its program without any conscious experience? What stops it from being, basically, a machine version of a philosophical zombie?
This is the hard problem of consciousness in a nutshell, and the Schmidhuber comment doesn't address it in any meaningful way.
edit: Out of curiosity, why would one consider him infamous?
IIRC claims were made that LeCun etc. were unfairly credited for work pioneered by him. As in, he pretty much invented modern ML but wasn't credited at all.
Perhaps look up the case with Ian Goodfellow's work on GANs. Schmidhuber said the GAN paper doesn't give sufficient credit to his earlier work on predictability minimization and was very vocal about this also in conferences (interjecting etc.).
In fact, it pretty much boils down to the same thing, as far as I can tell. You have a simulation? Someone set it up. He/she/it is the creator of this universe.
[Edit: I should say that a simulation winds up being the same as a religion, but I'm not sure that a religion necessarily winds up being the same as a simulation.]
Stating the cause/creator first and looking for evidence of it is not very reasonable.
You think this is some special matter? It's just carbon atoms all the way, which even started out as simple chemicals in a pond.
And just like you can't see "consciousness" in the brain (it's just a piece of meat with some electrical signals moving inside), you can't see it in the rocks either.
But the consciousness is in the encoding of the rock structure not each individual rock. Just like molecules taken out of the brain are not conscious, or the brain itself, when dead with no one to "move the rocks" (send neurons electric signals) is not conscious. It's the structure that is encoding the consciousness.
[Edited after reading the XKCD strip, in which by "throwing rocks on sand" means immitating a turing machine with rocks and sand].
Except if you think it's in an external, not meat/neurons/info soul.
Of course you can't test it empirically, because there's no empirical test for subjective experience. You can only test for the appearance of subjective experience - specifically for external behaviours, loosely defined, which might be considered to imply subjective experience to you as an observer.
This isn't a solvable problem, and Bostrom is begging the question by trying to hand wave it away with a sneaky "If we just accept this premise..."
Care to mention any of those "good practical reasons"?
Penrose gave it a try (though even there, there was nothing about how it can't be modelled with code for example, in which case, rocks are fine too for a simulation a la XKCD, just not as a direct substrate for an Android).
>Of course you can't test it empirically, because there's no empirical test for subjective experience. You can only test for the appearance of subjective experience - specifically for external behaviours, loosely defined, which might be considered to imply subjective experience to you as an observer.
It's not really that we can't test it, it's more that we can't measure it. But we can, and do, test it, e.g. with psychological tests, the same kind we give to people to understand different aspects of their personality (which we also can't just measure with some equipment or simple measurement). If it quacks like a duck, and looks like a duck, and cries like a duck, and has other sentimental duck qualities, it's a duck. We only accept the subjective experience of other people from its appearance (their actions etc) too, anyway.
The only reasoning I have seen on this subject so far has been "people are special" or "God made me". Sometimes the arguments are dressed up more. Do you have a good source that has a different type of reasoning?
Edit:
> You can only test for the appearance of subjective experience
Yes, and we solve this in people by assuming that if they sound, act and look like a person they are a person. I've never seen a compelling argument why it is somehow wrong to apply this principle to everything else, except of course the huge trouble philosophers go to to then argue what exactly needs to be looked at before we are satisfied. For some reason the Turing test is not adequate, but to my mind it's only not accurate if you believe that you are somehow special and that there is more to you than what can be observed.
Let's say I build a chatbot with a predetermined database of perfect responses to every possible string of words. This is quite clearly possible in theory, given a database with infinite room. This chatbot would pass the touring test, even though all it's doing is retrieving prewritten sentences. Is it conscious?
Ergo, it's quite clearly possible to build an automaton which acts like a human, but actually has no clue what it's saying. Which means the touring test isn't adequate.
---
Here's another take:
You have no proof that anyone in the universe is conscious other than yourself. However, given that you were born in the same way as everyone else, and you're conscious, everyone else probably is too. Why would you be so special?
Well, a computer program was not born in the same way that you were, so that assumption goes out the window.
For all intents and purposes, yes. Though without actual consciousness it would be impossible to build it.
>Well, a computer program was not born in the same way that you were, so that assumption goes out the window.
Well, a computer airflow program was also not created in the same way as actual air, but it can simulate it to the degree that we use it to build actual planes.
A computer program might not have been 'born' in the same way that we are, but we can make it simulate the "way that we are" (giving enough computer power, down to the atom or quantum element). So?
On one hand it is a wonderful outlook where computation and imagination creates universes as real as our universe feels to us (or rather as real as each creatures individually experienced universe seems to them). On the other hand, it does concern me that our own and fellow creatures experiences are perceived by this school as formally attributable to controllable processes, distinguished in "reality" only by beholders individuated interpretations, be they someone, a convincing replication of someone, or anything really which any individual decides to treat as conscious or not.
Either: - It can, and so can rocks. - It can, but there's something else unknown that's necessary to generate consciousness. - It can't, consciousness doesn't exist.
> it seems reasonable to allow that simulations, as experienced by their inhabitants, could have entirely different models of math, physics, and even consistency.
Which is a (shallow, informal) restatement of Hilary Putnam's work on externalism and its many responses. Just read some of that instead.
Software developers (or worse, tech entrepreneurs) often assume they have something to add to the discussion of simulationism just because computation is the simulation substrate du jour in Bostrom's work. But the philosophical discussion goes back directly to Descartes and indirectly even further. It's broad, complex, and not something a dilettante can just jump into after a teenage diet of scifi.
(This doesn't apply to Bostrom himself - the article's objection is a subset of his #2 option. If the "necessary overlap in metaphysics" is impossible or prohibitively expensive future civilizations are not likely to run them.)
Einstein was a patent clerk. Big, important ideas can come from anyone. They usually come from outsiders connecting the dots between different areas.
This isn’t me defending the writer of this article’s argument.
When Einstein was a patent clerk he organized a reading group to study Poincare and Mach. His annus mirabilis was six years after his first published paper. It's not impossible to for someone to swoop in and revolutionize a field with no background knowledge, but it's vanishingly rare in one as old as philosophy of mind, and Einstein is not an example of such a thing in physics.
This isn't an outsider connecting the dots; like most responses to Bostrom from non-philosophers it's barely on the right side of crankery.
It's 13 years after "dabblers and blowhards" and this site still hasn't developed the requisite self-awareness.
It's also not something we expected from TFA (to "revolutionize the field")
We just wanted to read a thought regarding the Simulation Argument that occurred to a fellow developer, and we got that.
Well, I didn't want that. And you apparently didn't want my thought regarding comments on the simulation hypothesis either.
My complaint is not that the article doesn't "revolutionize" the argument, but that it shows little to no engagement with the existing body of philosophy before, after, or even around it. Statements like:
> In the Western tradition, we have a deep assumption of the universality of mathematical truth.
reveals a view that seems disengaged from any developments philosophy after the early 20th century other than Bostrom's.
A stretched analogy: If someone posted an article about how no one's ever pointed out how bad it is that PHP variable names start with $ and this is difficult to type on many keyboards, and this is a key reason to avoid PHP, they would not be taken seriously - even though their conclusion, that other languages should generally be preferred to PHP, is probably valid.
I'd say the "western tradition" is more about what 99% of westerners would think and have traditionally though, including the majority of scientists and mathematicians (which would be closer to the "universality of mathematical truth") than about what some philosophers of epistemology postulated in the 20th century...
A blog post is a blog post, not a claim to academic supremacy. Blog posts are for thinking out loud and meandering, and that's welcome here. No one needs to be an expert on HN, and the job of genuine experts here—like yourself, I presume, given how you've been posting—is to share some of what you know in a friendly way, so the rest of us can learn. If you can't or don't want to do that, please don't post.
Edit: I took a quick look through your account's history. Everything I saw was completely fine and sometimes superb. I hope you'll continue contributing like that, and not so much like this.
Sites do have culture, both organically grown and cultivated. You can't say HN does not have collective viewpoints and attitudes while at the same time chastising someone for transgressing them.
So what? People should stop thinking for themselves and relegate it to the pros, lest they repeat informally something already investigated and rejected?
One could also point out how openness and encouragement of layman participation and collaboration with field outsiders have also been discussed extensively, and point people to read the relevant papers/books instead of your comment...
I didn't say anything about thinking. But yes, maybe many programmers (and celebrities and entrepreneurs - also called out by the article itself) should think more and publicly speak less.
As an argument it may or may not be convincing, but it’s not obviously actionable. It’s hard to see how it could be tested.
You seem to be saying that amateur philosophers shouldn’t jump in with their own naive takes on it, as they don’t have the required educational background. I can certainly believe that, but it takes away another avenue by which we non-philosophers can potentially engage with the work.
If there has been any notable followup or rebuttal, it hasn’t made nearly as much of an impact in the public consciousness (and of course the simulation argument is a pretty niche thing in the first place).
How does, or should, this simulation argument affect the average person? How might it affect us in future? What does it change?
This is a good question. This paper is popular well in disproportion to its impact, because it's a sexy scifi topic. I don't know Bostrom but I suspect it's popular far in disproportion to his desires also - it's just one small piece of his work on superintelligence, which is in turn only one part of his work on measuring existential risk.
The paper, it should be noted, is not an argument we are living in a simulation (it's often taken as one, but the simulation hypothesis predates Bostrom, the 20th century, etc.). Rather, it's an argument that there is a trilemma of which two options are material, and potentially even testable - so in a sense, it does allow the simulation hypothesis to be tested, if you buy his trilemma and reject the other options. (Based again on his other work, I suspect Bostrom may personally believe #1 - I personally am suspicious of both the validity of the trilemma and independently suspect #2.)
So this plays the kind of role in philosophy that a discovery of some unique way to synthesize a chemical we already have plenty of ways to synthesize in chemistry. It's a workaday paper - interesting if you are interested in the field, but of no great impact to most people's lives.
The result is that there's also not a lot of great ways to engage with the paper, in the same way it's difficult to engage with a refinement of the equivalence between inertial and gravitation mass.
If you want to engage with the simulation hypothesis and metaphysical realism more broadly, I would begin with Putnam's Reason, Truth and History. The philosophical work around the simulation hypothesis is maybe best thought of not as an attempt to actually answer it conclusively, but as particular lens on the relationship between reality, self-perception, and sense-perception. Clarifying those relationships is of immense use to the average person.
This thread would have people stop writing things on the Internet altogether, unless they have approached it the way one approaches writing an academic paper for peer review.
There's a value in writing down that flash of insight, even if it "have been considered, named, and much more eloquently stated hundreds of years ago".
For one, nobody can be familiar with all that "has been considered, named, and much more eloquently stated hundreds of years ago". Especially if it's not their main domain or main hobby.
Second, an amateur stumbling with his insights might also come at an interesting insight once in a while.
Third, it doesn't preclude anyone from going to the "real" experts. If anything, by discussing such things, it makes it more possible.
The only problem would be if said amateur thought he had the answers and that all the experts or previous discussion were wrong (like people who claim they have invented the perpetual motion machine, etc).
But I don't think we have that here.
In this case, I commented more harshly than I usually do because it's important for authors to be able to come here to contribute and not be met with garden-variety, albeit highbrow, toxicity. That makes HN strictly worse: it disincentivizes thoughtful people to participate and encourages hostility and snark. Feedback loops like that are all too easy to get going and they're an existential threat to this place.
michaelfeathers has been a good HN contributor in the past, as well as a distinguished author on software topics. Who cares if his blog post was "shallow, informal"? That's what blog posts are for. If someone else has deeper and more formal knowledge of a topic, that's wonderful: it puts them in a position to teach something to the rest of us here, which is obviously in the spirit of this site. If people want to play the intellectual bully, though, they should find another place to do that.
If you say you had no intention of being a jerk or putting anyone down, I believe you. But you did break the community standards (https://news.ycombinator.com/newsguidelines.html). I wish I could find a better way to persuade you not to post like that, because you obviously have valuable contributions to make.
Edit: maybe this would help. Sometimes people post harshly critical comments that are modeled after the communication style of more specialized environments, like literary reviews or academic colloquia. In those smaller environments, critique is often done in a concentrated and devastating style; that's how the game is played and how mastery is demonstrated, and it spurs the other players to get better.
The thing to understand is that this game doesn't work in large, flat internet spaces that are open to everybody. Here it just starts fires, drives away thoughtful participants, and encourages others—who are neither knowledgeable nor witty but far more numerous—to join in on what they think is the game, and it just gets degraded and degrading. The cost isn't worth it.
It took me years to understand this, because I don't like blandness in communication. Who wants verbal blancmange? What I didn't realize is how context-specific this is. The highly-critical game doesn't translate into the broader internet playing field. It just makes things nasty and dumb. I've written about this in the past with this analogy: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so....
Also https://news.ycombinator.com/item?id=21237524 and more linked from there.
Basically, it's a case of the-medium-is-the-message. We need to adjust our commenting style to the reality of the internet forum, or this place will get worse than it already is.
But why must it be post-human from future? why not a "pre-human-ancestor" from "relatively ages ago" that may have created a simulation to peek into simulated future?(so they can learn and won't make the same mistakes we did)
You may notice a contradiction in my example is IF the creator are instead from the past? How can they even be a human at all if they had to exist before human does?
The thing is chicken and egg "only" if we assume our creator themselves must be human, on the other hand
Today, we are designing/training artificial intelligence life form barely resembling us in simulations and most of their purposes are generally "forecast/solution/learning/gaming"
With the penultimate goal to create ASI to solve problems in nature, perhaps the reality is our universe may just be one of the branches as a result of such humble "exploration"
OBTW are you aware "I AM AI" a slogan favored by Nvidia is a cute palindrome itself that can also be rearranged to "AI AM I", "AM I AI", "AI I AM", "I A I AM" :D