If a computer was powerful enough would it generate consciousness?

[fractionMan]

yes is could, just not yet. And you'd need some software too.

 

[Crispy]

yes is could, just not yet. And you'd need some software too.

Ah, but as pika says, some of the brain's properties might be inherent in the precise chemical and physical structures therin - things that by their quantum nature, can't be simulated.

Personally, I think that we will eventualy create a machine that thinks, but it won't think like us, and it'll have a really hard time convincing us that it's 'alive' - the politics will be fascinating

 

[axon]

If the brain was that magically simple, we'd have done alot of it already, and we'd probably have a computer that was functionally identical to the brain. The fact that people have been investigating this issue pretty much since medical science and, later, computers came about gives some indication of just how complex the brain is and how little we know about how it works.

I disagree. We have not replicated the brain already as it is vastly more complex just in terms of the number of bits than anything else. We have been researching lots of bits of the body for a long time, it's still not known exactly how sperm and eggs fuse, and this is a hell of a lot simpler system than a brain.

The trouble with brains is they have a lot of neurones. And a lot is the worlds biggest understatement. Off the top of my head I think it is something like 100 million. And then they have a lot of connections, I think of the order of a hundred billion. Some individual neurones can have over 100 000 connections. So just to build the right number of parts would take more than an afternoon. And then the neurones are not all identical. And then each neurone couldn't be built as a single piece of wire due to their complex shapes and branching patternsm and the differing positions along the neurone of the synapses. So you'd need a bit of wire (plus a device that gives a probability value to 'fire' or not) for each small subsection of a neurone. I guess I'm saying in theory you can build a perfect replica brain but in practise it would take several hundred years to know how, and then several hundred more years to actually build it.

Heck, some recent research suggests that the brain may even work on quantum levels, meaning that even if we could replicate the brain in an atom-for-atom fashion, it still wouldn't work.

the brain uses QM no more than any other set of atoms in the universe. Damn you Penrose!

Fact of the matter is that understanding the brain is quite likely decades away, if not centuries.

I'll give it a few centuries.

 

[axon]

Assuming that it can be replicated using bits of wire and solder - which it can't.

why not?

 

[Blagsta]

yes is could, just not yet. And you'd need some software too.

No it couldn't. It's not made of wires and solder. How can you replicate something made of organic material that works electro-chemically with wires and solder?

 

[axon]

Ah, but as pika says, some of the brain's properties might be inherent in the precise chemical and physical structures therin - things that by their quantum nature, can't be simulated.

But the physical and chemical properties all go together to create electrical impulses, which are not QM in nature as the impulses are simply created by moving large ions about and not electrons (generally) as in a wire. You don't need to invoke QM to get away from determinism. Simply throw a dice, the result of an individual throw cannot be predicted exactly, but can be assigned a probability.

 

[sleaterkinney]

No it couldn't. It's not made of wires and solder. How can you replicate something made of organic material that works electro-chemically with wires and solder?

You program the wires and solder to act the same as a brain. The problem is understanding what the brain does, not being able to replicate it once we do(understand it).

 

[Crispy]

But the physical and chemical properties all go together to create electrical impulses, which are not QM in nature as the impulses are simply created by moving large ions about and not electrons (generally) as in a wire. You don't need to invoke QM to get away from determinism. Simply throw a dice, the result of an individual throw cannot be predicted exactly, but can be assigned a probability.

You assume that the electrical impulses are the only important part.

 

[butchersapron]

You program the wires and solder to act the same as a brain. The problem is understanding what the brain does, not being able to replicate it once we do(understand it).

Is it?

 

[Blagsta]

You program the wires and solder to act the same as a brain. The problem is understanding what the brain does, not being able to replicate it once we do(understand it).

Can wires do the same thing as neurotransmitters? That's rather a big assumption.

 

[sleaterkinney]

Can wires do the same thing as neurotransmitters? That's rather a big assumption.

You can write software that would mimic a neurotransmitter, yes. If you really wanted you could build one out of wire. What makes you think you can't?

 

[sleaterkinney]

Is it?

Yes, we currently have little understanding of how the brain works. If you do understand something it's a small step to actually build it...

 

[Crispy]

It'll be a long time coming. At the moment we can barely muster the computing power to accurately model the behaviour of individual atoms bumping into each other. Assuming that the brain depends on the minutest of interactions to work properly, any abstraction in our brain model will break it.

 

[Blagsta]

You can write software that would mimic a neurotransmitter, yes.

Can you?

If you really wanted you could build one out of wire. What makes you think you can't?

What makes you think you can?

 

[axon]

You assume that the electrical impulses are the only important part.

Well not the only important part, but the end product (at the molecular scale). A whole slew of proteins and spatial properties go define the elctrical characteristics of a particular bit of neurone. I think the important part for a brain is having a electrically connected BIG network.

 

[Blagsta]

Yes, we currently have little understanding of how the brain works. If you do understand something it's a small step to actually build it...

Maybe we will be able to build it, but it won't be with wires and solder.

 

[Crispy]

Maybe we will be able to build it, but it won't be with wires and solder.

silicon chips then. or whatever it is computers are made of in the future.

 

[sleaterkinney]

Can you?

Me personally?. I haven't tried.

What makes you think you can?

Because it's just another system, a complicated one, but still a system which behaves in a certain way depending on the input.

 

[sleaterkinney]

Maybe we will be able to build it, but it won't be with wires and solder.

Why not?

 

[Blagsta]

Me personally?. I haven't tried.

Because it's just another system, a complicated one, but still a system which behaves in a certain way depending on the input.

That's a very big assumption you're making.

 

[axon]

Can wires do the same thing as neurotransmitters? That's rather a big assumption.

Yes, very easily. A neurotransmitter acts on a protein to either make a neurone fire or the alter the probability of a neurone firing. For example when the neurotransmitter acetylcholine attaches to certain receptors in the brain it causes a flux of ions across the neuronal membrane, the net result of which is a change in the voltage difference between the cell and the outside. It's the voltage that determines whether a neurone fires or not.

The difficult bit is reconstructing billions of synapses, all at once, and in the right arrangements.

 

[Blagsta]

Why not?

I've already told you why.

Computers are very good at simulating things, but that is not the same thing as replicating.

 

[butchersapron]

Yes, we currently have little understanding of how the brain works. If you do understand something it's a small step to actually build it...

Is it? How can on earth you say that *now*?

 

[axon]

Computers are very good at simulating things, but that is not the same thing as replicating.

Interesting point. If you have a perfect simulation, is this the same as replication? Although in the case of brain/consciousness I think you'd need to simulate not just the wires in the head but the inputs (senses) and outputs (arm waving).

 

[sleaterkinney]

That's a very big assumption you're making.

In what way is it an assumption?

 

[Crispy]

For a start, a computer is digital and the real world is analogue.

 

[sleaterkinney]

Is it? How can on earth you say that *now*?

Say what, that we have very little understanding of it or that it would be a small stage to build it?

 

[axon]

Pah, analogue is just the same as digital but with better resolution.
Or to put it a slightly more logical way, digital is the same as analogue but at lower resolution.

 

[littlebabyjesus]

For a start, a computer is digital and the real world is analogue.

I'm really not so sure that's true. A neuron is either transmitting or not, isn't it?

*waits Axon to give confirm/ poo=poo.*

 

[sleaterkinney]

For a start, a computer is digital and the real world is analogue.

Computers can store and process analogue information, use fuzzy logic etc.