Philosophically speaking...

[Dawnstorm]

I think "indistinguishability" here means kind of what you do -- the consciousness within a given reality can't discern that its reality is simulated relative to some other reality.

But somehow a trajectory of "Pong --> photo-realistic 3D simulation --> ???" is supposed to the set-up of an argument here. And I don't get how.

 

[RX-79G]

But somehow a trajectory of "Pong --> photo-realistic 3D simulation --> ???" is supposed to the set-up of an argument here. And I don't get how.

I had thought the point was not how anything was "rendered", but that the advancement in complexity from Pong to today demonstrates that eventually a computer would have the power to "run" a reality as complex as ours.

There would be little point in having "Pong bats" if we only have Pong minds. The granularity of reality goes along with the mental capabilities possible in that reality. Our minds are only as complex as the physics that underlie them.

 

[Justin Swanton]

Per the Anthropic Principle, though, the physical conditions of the universe must physically allow for sapience. Otherwise, you're putting consciousness into the realm of the supernatural.

I gather, then, you're arguing from a theistic perspective.

The Anthropic Principle makes the point that however unlikely conditions are for life to exist in the universe, this universe (which assumes there are other parallel or prior universes) has them since it has sapient living beings able to think about it. But that doesn't mean that this universe has the necessary conditions to generate life, even less to generate intelligent life. That's an assumption.

Personally I believe in a God, but in this context that doesn't feature. My point is that intelligent consciousness cannot be created by the mathematical processes of a computer, no matter how complex that computer is. Intelligence is of a different order entirely.

Right. The article I posted addresses this.

At the purely functional level, our bodies are only slightly more complex machines than the garage door opener, really. We could, in theory, simulate an "apparently intelligent" human that is, in effect, nothing more advanced than a really fine-tuned garage door opener. It could look, function, and act like us. Given enough inputs and enough variables, there is every reason to expect that Siri will become even better than the "Computer Lady" in Star Trek, and if we give Siri a human body, there is certainly a line at which she will become indistinguishable from a human, which is what Dick gets at in DADOES.

No. The most sophisticated human robot will be able to mimic preprogrammed human responses, nothing more. There's an easy test: start a discussion on some concepts of which the verbal expression has not been entered into the robot's database. It will become very clear very fast that the robot has no idea what you are talking about.

This is also the whole idea at play in Asimov's Robot Novels. If we can create a human-analogue that appears sentient/conscious/sapient, then how do we really know that it's not?

With the above test.

If the robot asserts that it is thinking (I think, therefore I am), then who are we disbelieve it?

If it repeats the phrase when a certain circumstance triggers the response that means nothing. The robot has to prove it can understand intelligent concepts.

 

[RX-79G]

My point is that intelligent consciousness cannot be created by the mathematical processes of a computer, no matter how complex that computer is. Intelligence is of a different order entirely.

What order is it, and why?

 

[Fung Koo]

I'm not sure how to respond to this, Swanton.

Are you really suggesting that the life on our planet isn't from this universe?

(And if so, isn't that actually an argument in favour of this universe being a simulation?)

 

[Justin Swanton]

I'm not sure how to respond to this, Swanton.

Are you really suggesting that the life on our planet isn't from this universe?

(And if so, isn't that actually an argument in favour of this universe being a simulation?)

Well....if I understand Elon Musk correctly, he thinks it almost certain we are computer constructs created by a superadvanced alien civilisation. We are however conscious, thinking creatures. But given the nature of thought and consciousness, no technology no matter how sophisticated could replicate it. Therefore Musk is wrong.

But rewatching the video it's possible that Musk thinks we are superadvanced aliens who have created this simulated world for ourselves and are currently all living in it, unaware of our true nature. That's theoretically possible. The only question remaining is why would an advanced alien species create such imperfect surrogates for itself and put them in such an imperfect virtual world? (Answer: their true natures are much worse than ours and they want to forget the fact)

 

[Fung Koo]

Sorry, but how do you square your second paragraph with your first? It's impossible for aliens to do it, but it's possible we could do it to ourselves?

But given the nature of thought and consciousness, no technology no matter how sophisticated could replicate it.

What's your proof of this?

And what, pray tell, is "the nature of thought and consciousness" anyway?

 

[Justin Swanton]

What order is it, and why?

Intelligence is not a physiological or biological function. An intelligence can grasp immaterial concepts, i.e. realities that cannot in themselves be visualised (though we might attach a visual image to them for ease of reference). When a cat 'knows' that a dog is bad news, it does so purely from a combination of instinct and memory. Instinct gives it an impulse for self-preservation. Memory retains unpleasant sensations of dogs acting in a way that kicks in that impulse. You could programme a computer to retain similar memories and act in consequence.

Human have this kind of knowledge but they have more. The word 'beauty' for example has a clear and distinct meaning to the mind, but it doesn't correspond to any image or even similar set of images. A beautiful song, painting, dance, face - they are all beautiful but there is absolutely no material connection between the beauty of one and of the other. You can't quantify beauty nor reduce it to a mathematical equivalent. Animals and robots will never understand it. Beauty is an intelligible concept that corresponds to something immaterial that very diverse material beings have in common. A faculty that can grasp such an immaterial concept must itself be immaterial, not reducible to biological or physiological causes. So the intelligence and in consequence consciousness is immaterial - fused to a material body but not identified with it. Different order.

 

[Justin Swanton]

Sorry, but how do you square your second paragraph with your first? It's impossible for aliens to do it, but it's possible we could do it to ourselves?

OK.

Paragraph 1: it is impossible to create a virtual world in which the avatars have any real intelligence or consciousness.

Paragraph 2. It is possible to create a virtual world in which intelligent creatures - already existing and not created as part of that world - are immersed, thinking that the virtual construct is real.

 

[Matthew Hughes]

I like to think you all (the universe and everything) are just figments of my imagination.

At last, I know whom to blame!

But I have to congratulate you on the power and subtlety of your imagination. I mean, making each snowflake unique -- brilliant, if a little OCDish.

 

[N. E. White]

But I have to congratulate you on the power and subtlety of your imagination. I mean, making each snowflake unique -- brilliant, if a little OCDish.

You can thank the nub of hardened dead flesh on my left pinkie-toe. It's always shedding.

 

[RX-79G]

You can't quantify beauty nor reduce it to a mathematical equivalent. Animals and robots will never understand it.

I guess we'll just be disagreeing. Beauty is an emotional response to instinctual feelings about other beings or environments that have a positive association with comfort, survival, reproduction, etc. Animals have emotions that are connected to an appreciation of attractiveness in their environment or other animals.

It just appears that you are doing the same thing as humans have for a long time: Only people make tools, only people mourn the dead, only people have "real" emotions, only people have language. And then we find that all of those things are false.

We're animals. We're cognitive machines. The brains of tiny worms with only 302 neurons are still not understood how they function. Most larger brains are grossly more complex than any computer. We don't know anything about the most basic cognition, let alone higher order animals. There's no reason to suggest some weird line between humans and animals that isn't just a form of cultural imperialism. We just don't know enough at this point to make distinctions.

 

[Dawnstorm]

But rewatching the video it's possible that Musk thinks we are superadvanced aliens who have created this simulated world for ourselves and are currently all living in it, unaware of our true nature. That's theoretically possible. The only question remaining is why would an advanced alien species create such imperfect surrogates for itself and put them in such an imperfect virtual world? (Answer: their true natures are much worse than ours and they want to forget the fact)

That's sort of how I understood him, except that we're not currenly "all" living in them. I have yet to watch the video, though; whenever I remember to check this thread I'm not a computer with sound. The argument, as I understand it, is this:

Virtual and real worlds are indistinguishible. --> We're living in a world which is either real or virtual, but we can't tell which just from living in it. --> Since the real world contains billions of computers each of which runs one such world, probability tells us that the probability we're living in the real world is 1/billions.

I'm not sure if it's any more sophisticated than that. For example, I'm not sure if he believes that we could be sentient AIs, or if he even cares. Potentially, this set-up allows for stochastic solipsism: If I'm the playing a single-player game, then I'm the only sentient being in this simulated universe, but I can't know that I'm playing a single player game. If I'm playing a multi-player game, I have no way of telling the difference between a player or an NPC. Of course, the way human beings are, it's only a matter of time until someone who believes that Musk is right convinces himself that everyone who disagrees with any of his positions is just an NPC and thus its morally acceptable to get rid of them.

 

[Fung Koo]

OK.

Paragraph 1: it is impossible to create a virtual world in which the avatars have any real intelligence or consciousness.

Paragraph 2. It is possible to create a virtual world in which intelligent creatures - already existing and not created as part of that world - are immersed, thinking that the virtual construct is real.

The argument in the first establishes this topic as being inarguable. So, to proceed with the conversation, it would have to be stricken. We have to accept it as possible to create real consciousness without requiring a divine figure.

However, I'll play devil's advocate... if we take it as given that this is the case, and it really is impossible to create real consciousness, then let's step to the second point.

For the second argument, if we tackle this logically and test the hypothesis that this universe right now is a simulation and that we, sentient things within it, might not really be "from" here...

Presumably, the concept of a "soul" is what's at issue. There is some "otherworldly" thing imbuing the human body/avatar with consciousness. It is not "from" here.

In order to be perceptually ignorant of its otherworldly existence, per earlier point about senses, then while the "soul" is providing consciousness to this simulated body then the perceptual processes available to the "true" body of the soul's origin universe are... shut off? inaccessible? other something thing like that.

The first argument of the impossibility of creating consciousness means, therefore, that consciousness can only be created through whatever the natural means are to the true reality. The consciousness is not ever created to exist only inside this simulated reality -- of necessity, it pre-exists relative to its experience of the simulation. Consciousness is therefore alien to this experience and, effectively, maladapted to the simulation.

The sensory apparatus of the soul would thus have its origin in a universe with either similar or dissimilar features to the one it appears to be in, but per the first point, if we are incapable of simulating true consciousness because no technology could ever be powerful enough, then given the observations of the size of this universe -- including the galactic supercluster level to the Planck length -- and the functions of matters within the infinite universe -- including black holes, supercondensates, and quantum entanglement at the least -- then the simulation that is running would have to have infinite processing power yet still not be capable of simulating actual intelligence.

If infinite processing power were possible, we would of course have to allow the possibility of a created consciousness. "Infinite" in this case would mean "powerful enough to account for absolutely everything." But, given that you're saying no such technology could exist, then there is no way that a simulation of an infinite universe is possible. This simulation would be a finite universe.

Ergo, this reality cannot be a simulation anymore than any technology could create consciousness. Though the human mind is inordinately complex, the universe in its entirety, I would wager, is equally as or more complex than a human mind.

So your first argument invalidates the possibility of the second.

Unless the simulation, of course, really is not "complete" and does not render the universe in its entirety.

In that case, per the above requirement that the simulation receives already-existing consciousnesses, our consciousness' true existence would be perceptually attuned to its own physical perceptual processes from its true existence. The consciousness itself, per the earlier provided argument that sensory input proves reality is real, would then necessarily be "out of place" in the simulation. No matter how closely it appeared to match "real reality," it would never be a "correct" match to the perceptual inputs demanded by the consciousness.

In effect, the conscious mind would by necessity recognize the simulation as unreal, or a simulation.

I suppose, theologically, this is a core concept of several theistic frameworks -- but that they posit this reality is real.

So the only way you can square those two paragraphs, frankly, is if there is neither the possibility of creating consciousness nor of simulating a reality capable of deceiving a consciousness.

There would be only this one reality with its consciousnesses in it.

But that's not a fun way to continue this conversation

 

[Fung Koo]

Virtual and real worlds are indistinguishible. --> We're living in a world which is either real or virtual, but we can't tell which just from living in it. --> Since the real world contains billions of computers each of which runs one such world, probability tells us that the probability we're living in the real world is 1/billions.

He has mentioned this on more than one occasion, I believe.

The argument is that given the advances we've seen in computing technology in the last 40 years (from pong to advanced MMOs, VR, and AR), if we assume that computing technology continues to progress and get better and better, then over enough time we will eventually be creating simulations/programs that can produce worlds that are indistinguishable from reality. Given time and the assumption of continual and progressive development, this is a near certainty -- is the logical foundation of the argument.

Given this foundation, he argues that the likelihood that the reality we are currently perceiving is "base reality" is extremely low.

The idea that such a simulation could spawn internally-developed consciousnesses stems from the treatment of the idea by Benny Belvedere, author of the article I posted (and this is a common aspect of the argument -- if a simulation is indistinguishable from base reality, and base reality can produce consciousness itself because of its properties, then the simulation, too, should be able to produce "native" consciousness).

In one interpretation of the concept, there would be a base reality with potentially billions of simulations running within it. So again, likelihood is that this isn't base reality, but one of the billions of sims. But in another interpretation, if we then correspondingly assume that an indistinguishable simulation means that over enough time such a simulation could even be developed within a simulation, then the odds of this reality being "base reality" are even lower.

 

[Justin Swanton]

You can thank the nub of hardened dead flesh on my left pinkie-toe. It's always shedding.

Noooooo!

 

[N. E. White]

Oh, Justin...I haven't had breakfast yet!

 

[Stephen Palmer]

My point is that intelligent consciousness cannot be created by the mathematical processes of a computer, no matter how complex that computer is.

Oh dear, not this old chestnut again...

 

[Justin Swanton]

Oh dear, not this old chestnut again...

Old chestnuts are good for conkers, no?

 

[Fung Koo]

Adding a layer to the matter -- what if all the aliens have already upgraded into super intelligent AI?

http://motherboard.vice.com/read/th...y-superintelligent-robots?utm_source=vicefbca

If Musk's assumptions on development of simulations and the possibility of real created consciousness are probable, then life elsewhere could all be artificial, more or less.