Computing a theory of everything

[tical]

Ray Kurzweil confidently states that artificial intelligence will, in the not distant future, “master human intelligence.”

David Gelernter: “We won’t even be able to build super-intelligent zombies unless we approach the problem right.” This means admitting that a continuum of cognitive styles exists among humans.

Two of the sharpest minds in the computing engage in one of the oldest debates around: whether machines may someday achieve consciousness. (NB: Viewers may wish to brush up on the work of computer pioneer Alan Turing and philosopher John Searle in preparation for this video.)


Video link: http://mitworld.mit.edu/video/422

Watched the video last night nice stuff. My overall impression were that Kuzweil got eat up by Gelernter..ultimately because i feel we simple dont know enough about the brain, consciousness etc for kuzweil to make some of his futuristic claims some interesting points:

1) The brain is an emergent machine....meaning in its synergy(coming together of neurons, i suppose) somehow consciousness arises. Why can't that be the case for other devices(Kuzweil Point)?

2) Does the medium for consciousness need to be biochemical(Gelernter Assertion)? If so what processes does the interaction of organic and chemical properties simulate that mechanical and software can't?

 

[sean69]

Watched the video last night nice stuff. My overall impression were that Kuzweil got eat up by Gelernter..ultimately because i feel we simple dont know enough about the brain, consciousness etc for kuzweil to make some of his futuristic claims some interesting points:

1) The brain is an emergent machine....meaning in its synergy(coming together of neurons, i suppose) somehow consciousness arises. Why can't that be the case for other devices(Kuzweil Point)?



2) Does the medium for consciousness need to be biochemical(Gelernter Assertion)? If so what processes does the interaction of organic and chemical properties simulate that mechanical and software can't?

1] What device has over 100 trillion functional parts? Actually, much much more that that because that's not considering the neuron's protoplasm--all the stuff inside the cell; organelles which are made up of macro-molecules which are made up of smaller molecules. Not to mention the laundry list of neurotransmitters. So we're really talking about a complexity of a gazillion functioning parts. All in sync not only with each other but with the WHOLE BODY. Like Gelernter said, we think with our EMBODIED brain. Brain function is integrated with all body systems, CNS, immune, reproductive, respiratory, endocrine, digestive, musculaskelatal, circulatory...you name it. This is a major difference between the brain and a compputer and a very significant issue that Kurzweil is either taking for granted or just doesn't understand.

2] You can do stuff like simulate molecular dynamics, map surface electron density, optimize a molecule's structure based on stability pretty routinely using software aps like SPARTAN as long as it's a relatively simple small molecule. Larger complex molecules like proteins are a nightmare. Even with the best software, you're looking at computational times of months with a"meh" result.
Now you gotta simulate the entire brain along with all the other "embodied" parts, and the changing environment and then integrate all of this stuf

 

[tical]

1] What device has over 100 trillion functional parts? Actually, much much more that that because that's not considering the neuron's protoplasm--all the stuff inside the cell; organelles which are made up of macro-molecules which are made up of smaller molecules. Not to mention the laundry list of neurotransmitters. So we're really talking about a complexity of a gazillion functioning parts. All in sync not only with each other but with the WHOLE BODY. Like said, we think with our EMBODIED brain. Brain function is integrated with all body systems, CNS, immune, reproductive, respiratory, endocrine, digestive, musculaskelatal, circulatory...you name it. This is a major difference between the brain and a compputer and a very significant issue that Kurzweil is either taking for granted or just doesn't understand.

2] You can do stuff like simulate molecular dynamics, map surface electron density, optimize a molecule's structure based on stability pretty routinely using software aps like SPARTAN as long as it's a relatively simple small molecule. Larger complex molecules like proteins are a nightmare. Even with the best software, you're looking at computational times of months with a"meh" result.
Now you gotta simulate the entire brain along with all the other "embodied" parts, and the changing environment and then integrate all of this stuf

In somewhat defense of Kurzweil i see him in the same plight as a man way ahead of his time. You see people like Gelernter and i suspect some like yourself at this stage, right now, can laugh at some of his ideas!....As a scientist you should be familiar with many past generation of Kurzweil's, men way ahead of their time. A few points:

1) At this stage its obvious his time frame which i guess he based on some sort of moore's law for technology in this particular field ~Cognitive AI, is off. However,

2) The field of Neuroscience is only now starting to break ground can you or anyone else say in the next 20~70 years...their won't be insights that make up say its not a question of its its possible? but rather when can we have our first true working model of such said platform? IMO it's inevitable because

3) From what i know about biochemistry/biology and the little i know about neuroscience its obvious that many of the "moving" parts are empirical and right now like a toddler we are only limited by our infancy.Its only a matter of time as we learn more that we eventually have something and then who knows after that?

 

[sean69]

In somewhat defense of Kurzweil i see him in the same plight as a man way ahead of his time. You see people like Gelernter and i suspect some like yourself at this stage, right now, can laugh at some of his ideas!....As a scientist you should be familiar with many past generation of Kurzweil's, men way ahead of their time. A few points:

I'm not laughing at him. I simply fail to see the scientific justification of his predictions. Moore's Law is what it is. It's a conjecture that's been, thus far, on point. But Moore's Law is absolutely impertinent to his claim. If it's just a matter of using Moore's Law as a predictor of future technologies then one could very well say that, in 10 years or whatever, we'd be able to create a perpetual motion engine. This is obviously impossible as it violates both the 1st and 2nd Law of Thermodynamics for which there's irrefutable scientific consensus. Moore's Law wont help you here. It predicts the trend in information computation using technologies that are based on the same science that's founded on the 2nd Law. The same applies to the notion of reverse engineering the brain by simply using "code" from DNA. This simply WILL NOT WORK. Not because I or Gelernter say so, but because this idea clearly displays an utter lack of understanding of the neurobiology of the brain. Specifically MORPHOGENESIS of the neuron.

To say that you can reconstruct a cell because you have a bunch of "the building blocks", the nucleotide base pairs that "describe the brain", is just plain stupid! Aren't you a molecular biologist? You of all people should know this. This isn't some shit i'm making up. Kurzweil on the other hand, is a computer scientist and "futurist". I can understand his ignorance.

As a scientist, i'm familiar with several scientist that stepped outside the box, but still remained within the bounds of the foundations of science. Kurzweil is not one of them, trust me.



1) At this stage its obvious his time frame which i guess he based on some sort of moore's law for technology in this particular field ~Cognitive AI, is off. However,

2) The field of Neuroscience is only now starting to break ground can you or anyone else say in the next 20~70 years...their won't be insights that make up say its not a question of its its possible? but rather when can we have our first true working model of such said platform? IMO it's inevitable because

Like Gelernter said to Kurzweil, show me some evidence or inkling of of an indication that reverse engineering the brain based on computer software is possible. Don't quote Moore's Law because all that's saying is that computation power increases exponentially over the years. It says nothing about HOW this said "platform" will be developed.



3) From what i know about biochemistry/biology and the little i know about neuroscience its obvious that many of the "moving" parts are empirical and right now like a toddler we are only limited by our infancy.Its only a matter of time as we learn more that we eventually have something and then who knows after that?


Please explain what you mean by "many of the moving parts are empirical"

....

 

[mike123]

Watched the video last night nice stuff. My overall impression were that Kuzweil got eat up by Gelernter..ultimately because i feel we simple dont know enough about the brain, consciousness etc for kuzweil to make some of his futuristic claims some interesting points:

1) The brain is an emergent machine....meaning in its synergy(coming together of neurons, i suppose) somehow consciousness arises. Why can't that be the case for other devices(Kuzweil Point)?


sean already answered that one but i wonder what is the minimum complexity needed to have some kinda consciousness to arise. Not something that would even be as smart a primate or parrot but maybe a cell phone that knows and seems to like to be charged or recall not liking to be dropped with out having to program the responses in them


2) Does the medium for consciousness need to be biochemical(Gelernter Assertion)? If so what processes does the interaction of organic and chemical properties simulate that mechanical and software can't?

If thats the case then we would probably have to design electronics that work will with and inside of water or wait for nano machines that work fine in water

1] What device has over 100 trillion functional parts? Actually, much much more that that because that's not considering the neuron's protoplasm--all the stuff inside the cell; organelles which are made up of macro-molecules which are made up of smaller molecules. Not to mention the laundry list of neurotransmitters. So we're really talking about a complexity of a gazillion functioning parts. All in sync not only with each other but with the WHOLE BODY. Like Gelernter said, we think with our EMBODIED brain. Brain function is integrated with all body systems, CNS, immune, reproductive, respiratory, endocrine, digestive, musculaskelatal, circulatory...you name it. This is a major difference between the brain and a compputer and a very significant issue that Kurzweil is either taking for granted or just doesn't understand.

2] You can do stuff like simulate molecular dynamics, map surface electron density, optimize a molecule's structure based on stability pretty routinely using software aps like SPARTAN as long as it's a relatively simple small molecule. Larger complex molecules like proteins are a nightmare. Even with the best software, you're looking at computational times of months with a"meh" result.
Now you gotta simulate the entire brain along with all the other "embodied" parts, and the changing environment and then integrate all of this stuf

seems like there are 2 main barriers to all of this
Understanding the role that biochemical aspects of the brain play in consciousness and having the processing power to integrate that with the complex mechanics that we are trying to understand. i think we'll have computers powerful enough but we could become so fascinated with what we can do on computers that we completely miss the biological half of the puzzle

....

we may already have some of those inklings already but we may not be able to recognize them for what they are until we build a spider robot that doesnt know its a robot or a macro version of a bacteria that behaves in liquid the way a normal bacteria would

 

[Count23]

You're missing the point Malaki is trying to make. If you truly understand the relationship between science/physics and mathematics you'll get it. Read this book...http://en.m.wikipedia.org/wiki/Where_Mathematics_Comes_From?wasRedirected=true

that's supposed to be next on my list

 

[malaki]

george lakoff / "philosophy in the flesh" explains why the hard AI guys are fundamentally misguided. the brain is not a digital system. a functioning AI will not result from digital computation. you would need wetware modeled on the brain not digital computers.

 

[ScorpDiesel]

Ray Kurzweil confidently states that artificial intelligence will, in the not distant future, “master human intelligence.”

David Gelernter: “We won’t even be able to build super-intelligent zombies unless we approach the problem right.” This means admitting that a continuum of cognitive styles exists among humans.

Two of the sharpest minds in the computing engage in one of the oldest debates around: whether machines may someday achieve consciousness. (NB: Viewers may wish to brush up on the work of computer pioneer Alan Turing and philosopher John Searle in preparation for this video.)


Video link: http://mitworld.mit.edu/video/422

[FLASH]http://mitworld.mit.edu/flash/player/Main.swf?host=cp58255.edgefcs.net&flv=mitw-00812-csail-creativity-pt1-gelerntner-kurzweil-30nov2006&preview=http://mitworld.mit.edu//uploads/mitwstill-00812-csail-creativity-pt1-gelerntner-kurzweil-30nov2006.jpg[/FLASH]

 

[sean69]

1) sean already answered that one but i wonder what is the minimum complexity needed to have some kinda consciousness to arise. Not something that would even be as smart a primate or parrot but maybe a cell phone that knows and seems to like to be charged or recall not liking to be dropped with out having to program the responses in them


2) If thats the case then we would probably have to design electronics that work will with and inside of water or wait for nano machines that work fine in water


3) seems like there are 2 main barriers to all of this
Understanding the role that biochemical aspects of the brain play in consciousness and having the processing power to integrate that with the complex mechanics that we are trying to understand. i think we'll have computers powerful enough but we could become so fascinated with what we can do on computers that we completely miss the biological half of the puzzle


we may already have some of those inklings already but we may not be able to recognize them for what they are until we build a spider robot that doesnt know its a robot or a macro version of a bacteria that behaves in liquid the way a normal bacteria would

1) This is exactly what i'm talking about. We need to understand complexity and emergence first. Scientist have generally avoided these fields for several reasons; it's immensely challenging and involved plus unless you associate the study with some practical application where you can "make stuff right now", you're gonna have a helluva hard time finding funding.

2) I really don't know whether biochemistry is a necessary requirement for consciousness but I know one thing, the biochemistry of the brain is about as complex as it gets in the universe. Imagine, trillions of neurotransmitter molecules biosynthesized in mere seconds each with it's own specific function all initiated by processes as "trivial" as say, blinking or an itch. We take a lot of stuff for granted.

3) Computers are simply information retrieval machines. That's it. You feed in the info, write up the rules for retrieval (the algorithms) and execute commands for given function. All that computers have on the brain is computational power and memory. Sure you can write all kinds of exotic algorithms (like IBM's WATSON) but at the end of the day they're just a set of rules. The brain doesn't function solely on a set of inputed rules. Like I said,The brain's function is integrated with the external environment, which is constantly changing mind you, and internal biochemical control systems. All this shit engineered over 4 billion fucking years. It's not about computing/processing power.

To me, it's so obvious how computers and software are insufficient ingredients in recreating a brain when you understand how the brain is created. The cells that make up the brain aren't just assembled based on a set of rules enoded in the genome. Personal experience factors in. Every single synaptic connection and neuronal association that eventually form the final functioning brain is influenced by individual experiences--from the fetus to the developing new born--which can NEVER be duplicated.
Think about that for a moment the suggest to me how a computer software/algorithm can be developed like this.

 

[sean69]

GREAT THREAD!

bump







Haven't watched this yet. I'm on it...

 

[mike123]

1) This is exactly what i'm talking about. We need to understand complexity and emergence first. Scientist have generally avoided these fields for several reasons; it's immensely challenging and involved plus unless you associate the study with some practical application where you can "make stuff right now", you're gonna have a helluva hard time finding funding.

2) I really don't know whether biochemistry is a necessary requirement for consciousness but I know one thing, the biochemistry of the brain is about as complex as it gets in the universe. Imagine, trillions of neurotransmitter molecules biosynthesized in mere seconds each with it's own specific function all initiated by processes as "trivial" as say, blinking or an itch. We take a lot of stuff for granted.

3) Computers are simply information retrieval machines. That's it. You feed in the info, write up the rules for retrieval (the algorithms) and execute commands for given function. All that computers have on the brain is computational power and memory. Sure you can write all kinds of exotic algorithms (like IBM's WATSON) but at the end of the day they're just a set of rules. The brain doesn't function solely on a set of inputed rules. Like I said,The brain's function is integrated with the external environment, which is constantly changing mind you, and internal biochemical control systems. All this shit engineered over 4 billion fucking years. It's not about computing/processing power.

To me, it's so obvious how computers and software are insufficient ingredients in recreating a brain when you understand how the brain is created. The cells that make up the brain aren't just assembled based on a set of rules enoded in the genome. Personal experience factors in. Every single synaptic connection and neuronal association that eventually form the final functioning brain is influenced by individual experiences--from the fetus to the developing new born--which can NEVER be duplicated.
Think about that for a moment the suggest to me how a computer software/algorithm can be developed like this.

1. I see what u mean
u'd need specialist in many different fields working together to build robot ants or honey bees that behave like the real thing and maybe even blend in with them. Then takin some of that an working our way up to what we presume are higher levels of consciousness among animals til we get to humans. Or we could go the other route like i mentioned before and try to make things like cell phones and doors that truly know something about their environment, react when needed and even change ( software or hardware ) in unpredicted ways to deal with changes to their environment. But turnin any of that into something that could be profitable, which inturn leads to funding, is either to hard to picture or to far in the future to see anybody botherin to make the investments needed


2.Using something like the robotic equivalent of neurotransmitters is a whole nother can of worms that i havent seen anybody in the AI community address. And besides that unless we build a robot with a body and nervous system similar to ours we may have a hard time even recognize if its consciousness or if it just using clever programing and advance hardware to trick us. But i dont think having those biological components ( real or man made ) automatically means that thing is consciousness. Fruit Flys and organisms with only thousands or just hundreds of neurons probably arent consciousness in any measurable way. Just a bunch of obstacle/ object evasion software and food detection software for the most part. the neurotransmitter problem may be a piece in the puzzle that makes consciousness possible but not the piece

3.No arguments from me on this one
The work done creating WATSON is incredible but at the end of the day its just the software that will replace google some day and not the lead up to some type of AI. We'll need to build something that can be creative and interact direct like with its environment, not to mention the whole memory and subconsciousness stuff also. And All of this has to be built in such a way that it can change when it needs to or even wants to without a programmer or a technician needed to come do the upgrades. I still feel like its doable but the end result would only be a computer in name cause it probably wont be anything like we've ever seen or used before

 

[sean69]

3.No arguments from me on this one
The work done creating WATSON is incredible but at the end of the day its just the software that will replace google some day and not the lead up to some type of AI. We'll need to build something that can be creative and interact direct like with its environment, not to mention the whole memory and subconsciousness stuff also. And All of this has to be built in such a way that it can change when it needs to or even wants to without a programmer or a technician needed to come do the upgrades. I still feel like its doable but the end result would only be a computer in name cause it probably wont be anything like we've ever seen or used before

Why?

 

[mike123]

Why?

cause the problem seems to be a matter of approach, motivation and tech. The motivation is there and the tech is gettin better everyday. Most of the approaches being used may lead elsewhere or to just some deadend but thats usually how it goes when people are tryin to invent something new. but just cause somethings doable doesnt mean it'll happen. Most of the scientist who are around now that are going about the whole thing all wrong or have already made their minds up about what isnt possible probably wont be the ones who figure it out. It'll be those kids that are choosin what college to go to now who will make those breakthroughs

 

[ScorpDiesel]

GREAT THREAD!

bump







Haven't watched this yet. I'm on it...

Watched this some time ago. Shit's 3hrs long but, pretty interestin'.

 

[mike123]

watchin now and hellofa vid
gonna have to find dudes book


edit
i have to say ur vid may top everything i've post in this thread so far
alot of the stuff dudes been speakin on have been poppin into my head for the last couple of weeks
and that stuff on love and entropy sounds like the shit gurus and sages have been sayin for ages

 

[mike123]

bump

 

[sean69]

cause the problem seems to be a matter of approach, motivation and tech. The motivation is there and the tech is gettin better everyday. Most of the approaches being used may lead elsewhere or to just some deadend but thats usually how it goes when people are tryin to invent something new. but just cause somethings doable doesnt mean it'll happen. Most of the scientist who are around now that are going about the whole thing all wrong or have already made their minds up about what isnt possible probably wont be the ones who figure it out. It'll be those kids that are choosin what college to go to now who will make those breakthroughs

But mike123, I'm actually saying that based on the science (biology, neuroscience, etc) it is impossible to reverse engineer a brain -- especially, using computer software and algorithms! And understand that when I say impossible, i'm not alluding to the issues you brought up -- approach, motivation and technology. I'm alluding to the conceptual understanding of how computers are created and how they work, and how the brain is created and how it works.

 

[Olosi]

Fractal. That's what it all boils down to.

 

[sean69]

Fractal. That's what it all boils down to.

Care to elaborate?

 

[sean69]

Mikke123, remember this?
http://www.bgol.us/board/showpost.php?p=9255246&postcount=13

this is basically my point.

 

[sean69]

BRAIN BASED DEVICE
http://www.webofstories.com/play/15785?o=MS


WHY I DON'T THINK THE BRAIN IS A COMPUTER
http://www.webofstories.com/play/15775?o=MS

 

[mike123]

Care to elaborate?

i think i know wht hes gettin at but i'd rather let him fill in the blanks

Mikke123, remember this?
http://www.bgol.us/board/showpost.php?p=9255246&postcount=13

this is basically my point.

i get what ur sayin
our approach to the problem top-down, no variability and doesnt use anything resembling a body
So building something thats supposed to be like a human brain wont happen under those conditioned, i'm wit u on that one.
And chances are causality is just an illusion but puttin that aside as far as the brain is concerned i agree with u and the doc. But i still think going down these paths, even if they turn out to be dead ends or something else altogether, may lead to something that can and does work.

 

[mike123]

something cool and dangerous i've found



this one is more cool than dangerous

 

[John_Gault]

Architects and engineers build models of the real thing every day. Intelligent idiot ... indeed.

What up MK,


Have you ever heard the phrase . . ." The map is NOT the territory"?

It's one of the most profoundly elegant quotes (to me) that sum up our 'current' limitations in describing physical existential phenomenon via physics and maths.

But like many success in science, we get better and closer over time.

JG

 

[sean69]

CONSCIOUSNESS: A PROCESS NOT A THING
http://www.webofstories.com/play/15791?o=MS

 

[ShotzSho]

YOU won't like this........

We want to live forever.

The "conscious machine" is a bi-product of life.

It will happen.

Get over it.

The debate fuels the inevitable.

Stars moving like marbles.

Keep it simple stupid!

Reality Check....... I'm out THX for the gas.

 

[sean69]

this was a good thread. BUMP!