Agnosticism and Intelligent Design

[excreationist]

.....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)

I thought bone cancer would be due to a blind mutation. On the other hand things like thorns would have been designed if God exists. The Bible explains why God designed thorns.

 

[excreationist]

. 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.

If there was a grading for fitness that wasn't all or nothing, then things like the ball in the cup skill could evolve far more easily. (like how the robot was making progress even though it was failing)

....There is no intent, or goal.

If it can tell the relative fitness then there is a direction in which it can evolve.

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.

If its all-or-nothing and none are quite successful then it makes it much more unlikely for the thing to ever evolve. It's a bit like abiogenesis. If there is no way of selecting for partial life then it makes the whole process a lot less likely. (maybe I have no idea about what I'm saying)

 

[DrZoidberg]

.....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)

I thought bone cancer would be due to a blind mutation. On the other hand things like thorns would have been designed if God exists. The Bible explains why God designed thorns.

Where cancer arises is pretty predictable. So cancer is not really a mutation in that sense. It's just the result of bad design. It's an incredibly complicated system. And as any engineer can tell you, the more moving parts, the more likely it will break. It's either a fault in the systems regulating when to grow. So the growth goes out of control. Or it's a fault in the system that tells genes that have been growing to stop growing. Leading to the same problem.

The body has six (I think) independent systems all designed to track down and kill cancer. Your body getting cancer is actually very common. You probably get it every day. Many times. But your body sorts it out on it's own. It's only when all six of these fail that we say that we have the disease.

Why nature has "chosen" to do it this way is because nature does everything using recursive algorithms. It's the kind of system that would arise spontaneously if left alone. This is yet another incredibly strong piece of evidence for evolution.

 

[excreationist]

....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.

I think it is very reasonable to assume that there will be simulations that look real. Have a skim through these characters that are being rendered in real-time:
https://www.youtube.com/watch?v=Vh9msqaoJZw
and these environments:
https://www.youtube.com/watch?v=DDsRfbfnC_A
40 years ago there was space invaders and battlezone. Graphics in 5 years time will be even better - often it is in 4K, and what about in 40 years time?

We don't yet have true AI,

That's kind of a separate issue. Well things like Siri are quite impressive.

or simulations so good that they're indistinguishable from our experiential reality.

The jump between 1978 and now is huge... it is reasonable that the jump from now to 40 years in the future would also be huge.

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.

I'm sure Elon Musk is acting like we are in base reality (though I haven't in the past - see my other thread about my "cheating in the game of life" story)

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.

Well it does explain things like why space and time are pixelated and how information can travel across any distance instantly, and why it matters whether things are observed. (in games it ignores things that aren't observed/relevant)

 

[excreationist]

While neural networks are interesting for machine learning it's unlikely it is how our brains work.

The exact algorithms used would be different from our brain but I think it is clear that our brains do involve neural networks. They explain many things. Though our brains have about 100 billion neurons with each connected to about 10,000 neurons. It seems some artificial neural networks can also have billions of neurons:
https://www.quora.com/What-is-the-highest-number-of-neurons-weve-modeled-in-neural-network-code

They require a lot of processing power, and waste a lot of energy on calculations that don't go anywhere.

It seems the neural network system that can quickly beat humans at Atari 2600 games can run on home computers: (even though our brain is more powerful than a home computer in many ways)
https://www.reddit.com/r/MachineLearning/comments/2xhe36/running_deepminds_atari_ai_on_a_homepc/
As far as calculations that go nowhere - well at least it is checking - just in case.

The puffer fish's brain is tiny. It doesn't have the real estate nor energy to be a true neural network.

https://en.wikipedia.org/wiki/List_of_animals_by_number_of_neurons
An adult zebra fish has about 10 million neurons so I think a puffer fish would too. Neural networks can even involve a few hundred neurons so a puffer fish would be considered a true neural network.

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

Yes, instincts - these would be neurons that have pre-defined weights.

The brain isn't a tool for us to understand the world. It's function is to keep us alive.

What about people who understand that trains kill people and decide to jump in front of them? Or they sacrifice their life for a cause? (edit: well the brain does put up quite a fight in those cases)

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.

I think the fear of snakes and spiders is learnt depending on the attitude of the person to the unknown - or maybe, like in the case of cats and cucumbers, it is instinctual? It doesn't matter how likely we are to see snakes and spiders. What would be weird is if we had a fear of them if they posed absolutely no danger to us. But people can die from spider and snake bites.

 

[DrZoidberg]

The exact algorithms used would be different from our brain but I think it is clear that our brains do involve neural networks. They explain many things. Though our brains have about 100 billion neurons with each connected to about 10,000 neurons. It seems some artificial neural networks can also have billions of neurons:
https://www.quora.com/What-is-the-highest-number-of-neurons-weve-modeled-in-neural-network-code

Dude... it's highly speculative. We don't know how neurons work. We haven't the merest clue as to how our brain stores memories. That's a huge part of the puzzle missing. It's way too early to hitch your wagons to any cart. When it comes to neurology we're where cartographers were in the 14-hundreds. There's a lot of white on the map.

Also.... why do you feel the need to do it? You seem to think it very important to belong to a certain camp. Why? Why not just accept when you don't know something?

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.

I think the fear or snakes and spiders is learnt depending on the attitude of the person to the unknown - or maybe, like in the case of cats and cucumbers, it is instinctual? It doesn't matter how likely we are to see snakes and spiders. What would be weird is if we had a fear of them if they posed absolutely no danger to us. But people can die from spider and snake bites.

Not in Sweden. Our most poisonous snake can't even kill a baby. None of our spiders are able to pierce human skin. None of them poisonous. I think there only exists dangerous snakes and spiders in the tropics. All the temperate zones are safe.

While bears and wolves are seriously dangerous. There's zero innate fear of those. Kids just want to go and pet them. They think they're cute.

Not to mention traffic. Cars are seriously dangerous. If it's a learned behaviour to the degree you propose nobody would be riding a motorbike.

 

[excreationist]

Dude... it's highly speculative. We don't know how neurons work.

This video involves precisely simulating the neurons in a worm:
https://www.youtube.com/watch?v=eYS7UIUM_SQ
Artificial neural networks (ANNs) are based on what we know about biological neural networks. Why is it that a ANN can allow a robot to learn how to catch a ball in a cup or read hand-writing or do speech recognition?
https://medium.com/@ageitgey/machin...h-recognition-with-deep-learning-28293c162f7a
Is it just a coincidence that ANNs are a very efficient way to replicate human-type skills? ANNs can explain "hunches" or "intuition". You can get an ANN to predict something based on an untrained input. When I studied ANNs in university I made a very simple ANN which learnt a form of grammar - you'd train it for some examples and it could guess unseen examples and it would initally over-apply the rule (despite exceptions in the grammar) like how kids might say "runned".

We haven't the merest clue as to how our brain stores memories.

The weights that a ANN tweaks as it learns are a form of memory. Humans can also have memories linked to language. BTW in the case of ANN the memory of what catness or dogness is in a picture is smeared across many neuron weights. So memories in a ANN are similarly mysterious.

That's a huge part of the puzzle missing. It's way too early to hitch your wagons to any cart. When it comes to neurology we're where cartographers were in the 14-hundreds. There's a lot of white on the map.

You seem to think ANNs have next to nothing to do with biological brains. I think it is good to try and understand as much as we can rather than learning nothing about this.

Also.... why do you feel the need to do it? You seem to think it very important to belong to a certain camp. Why? Why not just accept when you don't know something?

ANNs have a huge number of applications including describing images, speech recognition, natural language processing, etc. This means better and better AI including better AI in games - which would approach levels that appear to be self-conscious. I'm not saying we know everything but it is interesting to see that we do know a lot.

 

[excreationist]

....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 think the following is a relevant analogy about the origin of life....

say there were particles that were randomly arranged:

and say "life" involved obvious geometric patterns - e.g.

There is an incredibly large number of patterns that could meet this requirement. But it still would be incredibly unlikely to happen by chance.

 

[Keith&Co.]

.....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)

I thought bone cancer would be due to a blind mutation. On the other hand things like thorns would have been designed if God exists. The Bible explains why God designed thorns.

If evolution is unguided, that's one thing. If reality is the product of choices made, i have questions.

The Bible also explains bone cancer, as God is the source of evil.
(Isaiah, somewhere)

 

[Keith&Co.]

If it can tell the relative fitness then there is a direction in which it can evolve.

if WHAT can tell?
How?

If its all-or-nothing and none are quite successful

If none are successful, we have extinction.
What evolution needs is successful ENOUGH.

then it makes it much more unlikely for the thing to ever evolve.

unlikely is not as important as you think.

It's a bit like abiogenesis. If there is no way of selecting for partial life then it makes the whole process a lot less likely. (maybe I have no idea about what I'm saying)

You don't.

 

[DrZoidberg]

This video involves precisely simulating the neurons in a worm:
https://www.youtube.com/watch?v=eYS7UIUM_SQ
Artificial neural networks (ANNs) are based on what we know about biological neural networks. Why is it that a ANN can allow a robot to learn how to catch a ball in a cup or read hand-writing or do speech recognition?
https://medium.com/@ageitgey/machin...h-recognition-with-deep-learning-28293c162f7a
Is it just a coincidence that ANNs are a very efficient way to replicate human-type skills? ANNs can explain "hunches" or "intuition". You can get an ANN to predict something based on an untrained input. When I studied ANNs in university I made a very simple ANN which learnt a form of grammar - you'd train it for some examples and it could guess unseen examples and it would initally over-apply the rule (despite exceptions in the grammar) like how kids might say "runned".


The weights that a ANN tweaks as it learns are a form of memory. Humans can also have memories linked to language. BTW in the case of ANN the memory of what catness or dogness is in a picture is smeared across many neuron weights. So memories in a ANN are similarly mysterious.

That's a huge part of the puzzle missing. It's way too early to hitch your wagons to any cart. When it comes to neurology we're where cartographers were in the 14-hundreds. There's a lot of white on the map.

You seem to think ANNs have next to nothing to do with biological brains. I think it is good to try and understand as much as we can rather than learning nothing about this.

Also.... why do you feel the need to do it? You seem to think it very important to belong to a certain camp. Why? Why not just accept when you don't know something?

ANNs have a huge number of applications including describing images, speech recognition, natural language processing, etc. This means better and better AI including better AI in games - which would approach levels that appear to be self-conscious. I'm not saying we know everything but it is interesting to see that we do know a lot.

Yeah, well... I've heard (and read about) the same speeches since the 50'ies. It didn't pan out then. We'll see.

 

[bilby]

....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 think the following is a relevant analogy about the origin of life....

say there were particles that were randomly arranged:

and say "life" involved obvious geometric patterns - e.g.

There is an incredibly large number of patterns that could meet this requirement. But it still would be incredibly unlikely to happen by chance.

Nope. It's almost certain to happen.

A random scattering of sand or granulated sugar on a metal plate will spontaneously form patterns like that if you tap the edge of the plate to make it 'ring' like a bell. It starts off looking like your first picture, and in less than a second it looks like the second.

No intelligence needed. Just random bouncing of the grains, but with some parts of the plate where the grains that land are less likely to bounce as far as they would from other parts of the plate.

Natural selection in action in an incredibly simple physical system. You can reproduce it in your kitchen.

 

[bilby]

Here you go - I found a video of it.

They use a signal generator to vibrate the plate at different frequencies, to get a variety of patterns; But all that's happening is a metal plate vibrating with sand on it.

[YOUTUBE]http://www.youtube.com/watch?v=wvJAgrUBF4w[/YOUTUBE]

 

[excreationist]

Here you go - I found a video of it.

They use a signal generator to vibrate the plate at different frequencies, to get a variety of patterns; But all that's happening is a metal plate vibrating with sand on it.

[YOUTUBE]http://www.youtube.com/watch?v=wvJAgrUBF4w[/YOUTUBE]

I wrote "say there were particles that were randomly arranged" but that's really interesting

 

[excreationist]

...Yeah, well... I've heard (and read about) the same speeches since the 50'ies. It didn't pan out then. We'll see.

Well now we have ANNs with billions of neurons. They can beat humans in a lot of things (including Atari 2600 games and chess). I agree that researchers in the 50's were over-confident.

 

[bilby]

Here you go - I found a video of it.

They use a signal generator to vibrate the plate at different frequencies, to get a variety of patterns; But all that's happening is a metal plate vibrating with sand on it.

[YOUTUBE]http://www.youtube.com/watch?v=wvJAgrUBF4w[/YOUTUBE]

I wrote "say there were particles that were randomly arranged" but that's really interesting

Did you watch the video? The particles WERE randomly arranged at the beginning. The patterns arose simply from vibrating the plate.

It's natural selection. A sand grain that lands on a part of the plate that is moving fast will be thrown off a long way in a random direction. If it happens to land on a part of the plate which is moving more slowly, it gets thrown a much shorter distance (still in a random direction). Almost immediately, grains on the fast moving parts of the plate become 'extinct'.

 

[excreationist]

Here you go - I found a video of it.

They use a signal generator to vibrate the plate at different frequencies, to get a variety of patterns; But all that's happening is a metal plate vibrating with sand on it.

[YOUTUBE]http://www.youtube.com/watch?v=wvJAgrUBF4w[/YOUTUBE]

I wrote "say there were particles that were randomly arranged" but that's really interesting

Did you watch the video? The particles WERE randomly arranged at the beginning. The patterns arose simply from vibrating the plate.

It's natural selection. A sand grain that lands on a part of the plate that is moving fast will be thrown off a long way in a random direction. If it happens to land on a part of the plate which is moving more slowly, it gets thrown a much shorter distance (still in a random direction). Almost immediately, grains on the fast moving parts of the plate become 'extinct'.

I'm talking about patterns in a random arrangement with no mechanism to consistently bring about patterns.

 

[bilby]

Did you watch the video? The particles WERE randomly arranged at the beginning. The patterns arose simply from vibrating the plate.

It's natural selection. A sand grain that lands on a part of the plate that is moving fast will be thrown off a long way in a random direction. If it happens to land on a part of the plate which is moving more slowly, it gets thrown a much shorter distance (still in a random direction). Almost immediately, grains on the fast moving parts of the plate become 'extinct'.

I'm talking about patterns in a random arrangement with no mechanism to consistently bring about patterns.

Why? That's very boring. With no mechanism to bring about patterns, no patterns will arise. Yawn.

You do realize, I hope, that the very simple mechanism by which those patterns on the plate arise are exactly analogous to the very simple mechanism by which diverse species of life arise?

Evolution is not 'no mechanism'; It's a very simple and straightforward mechanism.

1) Offspring are similar to, but not identical to, their parent(s).
2) In any population, some individuals are more likely than others to become parents, in part due to those small differences between them.
3) Over time, this leads to life forms that differ enormously from their distant ancestors. (For the same reason that if you keep taking small steps, eventually you can end up a long way from where you started).

That is it. That's the entire theory of evolution in a nutshell. What part of that simple mechanism do you think is in any way wrong, mistaken, implausible, or hard to understand?

 

[Keith&Co.]

I'm talking about patterns in a random arrangement with no mechanism to consistently bring about patterns.

Mechanism?
I thought we were trying to figure out if evolution required intelligence?

This is an example of patterns being produced by a non-intelligent process.

 

[excreationist]

bilby:
I'm trying to show that there is a large number of valid patterns but it is still highly unlikely if there was no mechanism to consistently create patterns.