Agnosticism and Intelligent Design

[Jobar]

More. This presumption that reality is not a simulation can be called the Prime Axiom; the presumption that our senses are reporting a reasonably accurate representation of external reality, and we aren't just brains in jars, or patterns of bits in some googol-flop computer.

True that our senses don't infallibly report perfect models of reality; do enough drugs, suffer a head injury, or be mentally ill. Your mental model can vary considerably from the consensus model other people agree on. Why, just do a search for optical illusions on the net. But those misfunctions of our sensory apparatus can be explained by other observers; even those who suffer some sensory or brain dysfunction may be shown what their problem is, and with careful analysis come to understand why their mental models are flawed.

 

[excreationist]

But if it is easy to describe by language, why can't you make it work the first try?
Why do we have to learn the way we do?

Well that robot arm in the video partly learnt by people demonstrating how to catch (by moving the arm with their hands). BTW if there are major problems using language to describe catching a ball than surely it would be far harder to describe the puffer fish design.

And i think you oversimplify the fact that we do learn, and what is involved with it.

And you don't seem to have much knowledge about neural networks.
Neural networks like DeepMind are VERY good at learning
https://www.youtube.com/watch?v=V1eYniJ0Rnk
here it learnt to play a game better than most humans without being programmed in the rules of the game.

https://futurism.com/google-deepmind-ai-finds-way-doom-like-3d-maze-sight/
"the AI navigates through the maze simply by “sight” or by looking at the screen and deciding on what to do next, just as humans would play it."

The puffer fish doing something by instinct means it is a process that has taken more than one lifetime to develop.

In essence, it HAS been learning, by trial, error, and selection, for thousands of generations. A little bit with each successive one, progress stored in the DNA.

Us learning something even half as complicated, within one lifetime, that's far more impressive than the puffer fish.

An instinct in DNA only has one try per generation. It gets modified every generation. But when a human learns, they can learn once every few seconds and have tens of thousands of tries. In a neural network it is VERY efficient at learning. If it gets something wrong it adjusts the weights. Or if it gets something right it also adjusts the weights. But instincts in DNA just make the whole DNA more or less likely to survive rather than specific neurons being adjusted in an optimal way. You really need to understand how neural networks work since they are fundamental to how behaviours work.

 

[excreationist]

....who- what intelligence- is doing the simulating?

In a lot of cases it would be for entertainment.

How was this Matrix created?

Using computers - or I think in the case of eXistenZ it involved a biologicial device. In the Matrix the characters are usually controlled by actual brains.

Obviously by an intelligence- but how can that intelligence be certain that it is not also being simulated?

An episodes of Rick and Morty and the movies eXistenZ and the Thirteenth Floor involve simulations within simulations.

Is it simulations all the way down? Down to what?

Base reality of course.

Here is Elon Musk's reasoning:
https://motherboard.vice.com/en_us/article/8q854v/elon-musk-simulated-universe-hypothesis

 

[excreationist]

I'm aware of the major problems in the Bible. But sometimes I wonder if somehow there is some truth to the Bible - partly so I can have open discussions with Christians at my church.

Define "some truth". The bible, for instance, says killing people is bad, which is objectively right.....but then the murder rate in the bible by gawd or true xtians is ferocious. But the vast majority of the bible is just plain wrong or awful or illogical or downright self-contradictory. And you can't really have open discussions with xtians, as amply demonstrated by any thread with LIRC in it.

I've been reading Christian apologetics books and talking to Christians. They are often aware of all those problems but they have counter-arguments. It is conceivable that an injust monster of a God exists who likes Christians saying he's perfectly holy, etc.

 

[excreationist]

....Anticipating the parabola of a hurled object directed towards us is advanced trigonometry in real time. It's an absolutely amazing piece of evolutionary design.

...If you use evolutionary principles to create a programme to make a robot catch a ball thrown at it... good luck with that. Our top minds have been trying for decades. Its not going so well.

About this again:

The robot was taught and practiced like a human:
https://www.youtube.com/watch?v=qtqubguikMk
https://www.youtube.com/watch?v=W_gxLKSsSIE

That's thanks to neural network learning. But that process of optimised learning wouldn't have been at play with the puffer fish since the DNA doesn't know exactly which neurons need to be adjusted (which is done during neural network learning)

 

[excreationist]

Keith&Co:
Just a tangent:
As far as piles of cards vs houses of cards go, there is also the concept of "entropy" which seems to be measured in joules per kelvin...? (though that measurement doesn't have a clear relationship with the complexity of the behaviour required to create the arrangement)

 

[steve_bank]

....Rough estimate implies that you could spend the time to make a fine estimate, but you really don't have a method of doing so, do you.

Because it would actually depend on the exact results desired, and tge existing knowledge base of the brain involved.....

Say I wanted to compare the intelligence involved to make a pile of cards vs a house of cards.... a crow could maybe be trained to make a pile of cards. The house of cards is a geometric pattern while the pile of cards is not. Maybe some people would say that the algorithm for a house of cards is simpler... but the actual fine motor skills required is more complex.

That is the creationist Clock Maker argument. You find a mechanism you have not seen before and you assume somebody or something must have made it. Therefore it follows that the universe which looks like a mechanism must have been designed.

By the way there is a video somebody got by chance of a wild crow bending a piece of wire to get at a piece of food in a bottle. There are videos of wild squirrels figuring out how to defeat squirrel proof bird feeders, sometimes including cooperative group behavior.

Who taught seagulls to drop clams on rocks to open them? What you are calling human centric intelligence is an evolution of the brain.

As to entropy the technical definition you posted refers to the efficiency of a heat transfer process doing work and the theoretical efficiency limit. Philosophically entropy is described as order vs disorder. Left alone all systems run down. Turn off a refrigerator and it will warm to the background temperature. A refrigerator is an 'entropy reverser'.

Without my active working against entropy my office at work always descended into chaos and clutter. Literally.

The question of learning and trial and error including a robot is the question of a priori knowledge. We are not born knowing anything like driving a car or math. We have a gentic capacity to develop knowledge through trial and error couples with imagination.

At the time of the first self replicating organisms there was no 'knowledge'.

 

[Keith&Co.]

No, I'm just ignoring the discussion about neural networks learning. I can't see how it affects the puffer, and it doesn't quite match your other claims.

You are wrong about instinct getting only one try per generation. Evolution takes place in gene pools, not individuals. So the gene pool gets a try with every new mating. A hundred males will make a hundred patterns, or tries. Fifty females will pick the ones that got it 'right.'
Slight variations in every born male will alter the tries, slight variations in the females alter the definition of 'right.'
The pattern represents ten thousand generations of winners, but it also represents hundreds of thousands of tries.

 

[DrZoidberg]

I think for the idea of how unlikely evolution is to be intellectual, it would require me to be very informed.
e.g.
https://en.wikiquote.org/wiki/Fred_Hoyle
Sir Fred Hoyle - "...the chance of obtaining them all in a random trial is only one part in 10^40,000..."
So I'm not sure how informed I am about this. And like I said maybe there is a near infinite number of parallel universes so evolution would be inevitable in many of them.

While true creationists often miss why it's not a problem for evolution. If the goal of evolution is what we have now then that would be incredible. But evolution doesn't have a goal. If it hadn't been like this, it would have been some other way. I could have been typing with tentacles or claws. Or something else entirely. Life might have stayed as unicellular goo. A world like that would have exactly the same odds of existing

I am still agnostic but I'm curious about things that are unknown to some degree.

I'm also curious about the unknown. But I see no reason to title myself a believer, non-believer or agnostic. Why not just call yourself a "curious person with an open mind"?

I have also never understood how the supernatural was ever a contender in the marketplace of ideas. The supernatural can never be more than "yeah... wicked concept, man... let's smoke some more".

I know it's important for a lot if people to have a faith. I just don't understand why. Why not just have an open mind about all of them? Who gives a shit if heaven is real. If it is you'll find out eventually anyway. Why waste energy thinking about it?

In the case of Christianity the traditional church view is that most people go to hell forever. If that is true then it would be good to avoid it. So I'm spending a bit of effort into seeing how accurate the Bible is.

Meh... If getting in hinges on faith in God then God is a vile creature, and the best option is to stay away from him as much as possible. Why would you want to go to a heaven ruled by somebody like that?

Also... You only have God's word for it. Why trust God? What if Hell is the nice place to go?

Me personally I put a lot in the bucket of "I don't know and can never know". And just move on. Whether God exists or not is in this bucket.

 

[excreationist]

....You are wrong about instinct getting only one try per generation. Evolution takes place in gene pools, not individuals. So the gene pool gets a try with every new mating. A hundred males will make a hundred patterns, or tries. Fifty females will pick the ones that got it 'right.'
Slight variations in every born male will alter the tries, slight variations in the females alter the definition of 'right.'
The pattern represents ten thousand generations of winners, but it also represents hundreds of thousands of tries.

Sorry I didn't know what I was talking about.

I think an analogy of a child trying to catch a ball is relevant. They know if their skill is improving and get better and better. But if they were blind folded and have no help, it would be much harder to learn to catch a ball. In evolution how fit or successful you are can improve gradually - like the kid learning to catch a ball - or it is a case of pass/fail so it doesn't know how to gradually improve how successful you are.

I'll reply more later.

 

[DrZoidberg]

While neural networks are interesting for machine learning it's unlikely it is how our brains work. They require a lot of processing power, and waste a lot of energy on calculations that don't go anywhere. The puffer fish's brain is tiny. It doesn't have the real estate nor energy to be a true neural network.

Most likely all brains, human and other animals, have a lot of pre-programmed bits.

The brain isn't a tool for us to understand the world. It's function is to keep us alive. That's why people living in cities still have an instinct to fear snakes and spiders, even though we're unlikely to ever come across them.

 

[Jobar]

In a lot of cases it would be for entertainment.


Using computers - or I think in the case of eXistenZ it involved a biologicial device. In the Matrix the characters are usually controlled by actual brains.

Obviously by an intelligence- but how can that intelligence be certain that it is not also being simulated?

An episodes of Rick and Morty and the movies eXistenZ and the Thirteenth Floor involve simulations within simulations.

Is it simulations all the way down? Down to what?

Base reality of course.

Here is Elon Musk's reasoning:
https://motherboard.vice.com/en_us/article/8q854v/elon-musk-simulated-universe-hypothesis

Far as I can tell, Musk's ideas are all just hand-waving.

Not *impossible* hand-waving, no; but he gives us no evidence, nothing but speculation. We don't yet have true AI, or simulations so good that they're indistinguishable from our experiential reality. Until such evidence is presented, I think our most sensible course is to presume, and act like, we are in the 'base reality' which *may* eventually generate such perfect simulations that we can't tell which is reality, which is simulation.

There are similarities between Musk's notions, and the idea that we live in a universe created by a deistic god, which doesn't interact with its creation. Both speculations are pretty much bootless; we learn nothing about the ultimate nature of reality from them, nor benefit in any way.

 

[steve_bank]

There can be no supernatural. Anything that exists is by definition natural.

If a you see a ghost there are two possibilities. It is created by your mind as an hallucination or the ghost exists and there is a causal link to your brain.

Anything that interacts with reality by definition has to be natural. If I recite an incantation and a demon appears, there has to be a causal link.

Goedel wrote that he thought an anlog to the human mind could be developed and raised as would a child.

Back in the 80s I was frinds with a family with a rug rat.

I had her out in the yard. She started crawling towards a big rock. Every few feet she reached out trying to touch it. When she got to the rock she touched around it.

One night music was playing on floor speakers. She crawled across the room feeling the air with her hands until she got to the speaker.

We are genetically programed to explore and build an internal picture of reality.

 

[Keith&Co.]

. In evolution how fit or successful you are can improve gradually - like the kid learning to catch a ball - or it is a case of pass/fail so it doesn't know how to gradually improve how successful you are.

It does not KNOW anything. It's blind variation, with all-or-nothing grading.
There is no intent, or goal. Or perfect.

In any generation, some will be the fit enough, some will not.
The fit-enoughs will by definition live to reinforce the fitness qualities going forward.
And that's it.

 

[Jobar]

Here's an interesting, and entertaining, take on reality-as-simulation. It's from the webcomic Questionable Content, which I read regularly. It takes place in a near future where AIs have been perfected, and are commonly installed in many sorts of robotic bodies, many of them humanoid. Start at https://www.questionablecontent.net/view.php?comic=3238, and read up to number 3242. (I've linked to this comic before, in the course of discussions on both artificial intelligence, and the possibility that we live in some virtual reality.)

 

[Keith&Co.]

Here's an interesting, and entertaining, take on reality-as-simulation. It's from the webcomic Questionable Content, which I read regularly. It takes place in a near future where AIs have been perfected, and are commonly installed in many sorts of robotic bodies, many of them humanoid. Start at https://www.questionablecontent.net/view.php?comic=3238, and read up to number 3242. (I've linked to this comic before, in the course of discussions on both artificial intelligence, and the possibility that we live in some virtual reality.)

Meh.
Whether our reality is virtual or the result of intelligent manipulation, the question remains: someone took the time to design bone cancer for children. What the fuck is up with that?

(With all due credit to Stehan Fry)

 

[steve_bank]

Steve pawn in game of universe simulation.

 

[skepticalbip]

I think that Elon Musk must be a gamer that got entirely too consumed into "Forge of Empires".

 

[steve_bank]

Steve pawn in game of universe simulation.

https://www.youtube.com/watch?v=SKRma7PDW10

 

[bilby]

Keith&Co:
Just a tangent:
As far as piles of cards vs houses of cards go, there is also the concept of "entropy" which seems to be measured in joules per kelvin...? (though that measurement doesn't have a clear relationship with the complexity of the behaviour required to create the arrangement)

The entropy of a system in Physical Chemistry is proportional to the (natural) log of the number of 'microstates' in the system, and is a measure of how 'disordered' it is - microstates are defined by the fact that new information is needed to describe them (rather than saying 'just like another one we already looked at'). The number of microstates is obviously a dimensionless number, but the constant of proportionality (Boltzmann's Constant*) in this case has units J.K^-1, and this allows us to see that statistical mechanical entropy (ie as defined by the number of microstates) is equivalent to thermodynamic entropy - the change in entropy of a system is the change in the internal energy of the system divided by the absolute temperature - hence J.K^-1.

Entropy has nothing to do with 'behaviour'; It's a measure of disorder, and can be thought of as the amount of information needed to completely describe the system. In your picture, once you describe the position and orientation of the first card, you can describe group A without doing that separately for all four cards - card 2 is 'the same orientation as card 1', and the position of cards 3 & 4 is a function of the positional difference between cards 1 & 2. Group B is more disordered, and has a higher entropy, but still cards 1 and 4 share a vertical orientation (unlike in group C), so group B has lower entropy than group C does.

In most cases, one cannot determine the absolute entropy of a system, but only the change in entropy - to know the absolute entropy, one would need to know all of the possible variables, including any (as yet undiscovered) sub-quark structures that may or may not even exist. So you can say that a particular change to a system has increased the entropy by ΔS J.K^-1, but not what the absolute starting (or ending) entropy was. By convention, entropy is denoted by 'S', so change in entropy is given as 'ΔS', where Δ ('delta') is the conventional symbol for the change in a quantity.

So we can precisely determine how 'disordered' a system is relative to another system.

Disorder is all about unpredictability. The less easily you can predict (or describe) the next part of the system, based on the parts you have already examined, the higher the entropy of the system is.








*k_B = 1.38064852(79) × 10^-23 J.K^-1, is equal to the Ideal Gas Constant divided by Avogadro's constant, and relates temperature to energy per particle in an ideal gas (the Ideal Gas Constant relates temperature to energy per mole).