• Log in with Facebook Log in with Twitter Log In with Google      Sign In    
  • Create Account
  LongeCity
              Advocacy & Research for Unlimited Lifespans


Adverts help to support the work of this non-profit organisation. To go ad-free join as a Member.


Photo

"Artificial" intelligence: no programming Necessary?


  • Please log in to reply
5 replies to this topic

#1 fizzionz

  • Guest, F@H
  • 32 posts
  • 0

Posted 27 February 2008 - 10:54 AM


ok so im no physicist, cell biologst or computer programmer.. yet.. but i have do have imagination.. so shoot this idea down if you like.

Firstly I know we currently dont have the computer power to achive this idea...yet.. but here we go. I have read in a book i have been reading (The Cell: a molecular approach) that they have the ability to scan a cell at a specified depth using photons that meet at the desired location giving off a signature of the atom or atoms present there. so lets say one day we get the technology to scan an entire cell very rapidly, with all inforation going directly into a computer model. it may take a massive amount of information to make a model of an entire human (like a trillion cells?) making it a feat bigger then the human genome project. but i have thought alot about this, and have decided that this may be overcome being that alot of inforamtion is repetitive inside the body. we dont have to scan and model every single chromosome of every cell, and why even scan an entire human.. when humans grow from tiny embryos, how much information or yet how many cells are there in a fertilized embryo? if this information was put into an atomic replica and then lets say we apply real world laws of physics and chemisty to the model... what then.. (im assuming we have a grand unified theory of physics at this stage).. obviously if we made the model work exactly like in real world physics the embryo would die due to the lack chemicals that are givin by the mother of real world embryos, so we would have to simulate the in/out put of the virtual embryo e.g food , oxagen and other organic processes that maintain the growth of the cells. imagine that you were to make a tamagotchi human... if you were to replacate the every atom of a human into a working model, as soon as the model was "turned on" into the emulated physics of our universe, the model would die from a lack of oxygen.. so the designer would have to include an enviorment for the model. I know this whole idea would take a huge leap in technology, but lets say we do get a working model of an embryo to grow, the computer wouldnt have to remember the past, only the present, the same as the universe doesnt have to remember the big bang to support life today. infact the universe doesnt have to compute any information to make my consciousness, every atom has a place and a job to do by itself, the universe does have to produce hydrogen and oxygen for me to use as a system. so the universe supplys 103 differnt building blocks, all of which are combinations of hydrogen atoms fused under nuclear energy. so do i have to only undersatand hydrogen in this embryo emulator of mine...

ok so im going on and on here.. but what im getting at is that if i were to emulate atom for atom in real world physics an embryo, it should grow a human consciousness that would only require a computer that can compute all the repetitive chemical reactions in the system..

what really interests me about this concept is the ability of the tamagotchi human, to build a computer that can do this very task out of the atoms you provide him. kinda like the matrix... spooky :-) hope you enjoy my blaber, maybe ill make it into a screen play..

#2 maestro949

  • Guest
  • 2,350 posts
  • 4
  • Location:Rhode Island, USA

Posted 27 February 2008 - 12:53 PM

Good thinking but "build a computer that can do this..." contradicts the "no programming necessary" as what a computer is does is based strictly on what you program it to do. Even machine learning algorithms must be taught how to accomplish tasks. The atomic replica idea is quite possible and there are already software tools that can simulate quantum behavior of chemical reactions with great precision, Qbox is an example. The problem is, as you mention, is scaling these applications. Even with petaflops, only a few thousand atoms can presently be simulated.

A more tractable solution for our generation might be to replicate brain functionality by studying a developed brain and simulating as much genetic, expression and neuronal functionality as possible.

sponsored ad

  • Advert

#3 fizzionz

  • Topic Starter
  • Guest, F@H
  • 32 posts
  • 0

Posted 29 February 2008 - 01:37 AM

ok, what i mean by no programming, is that the task of understanding a learning thinking computer, or brain.. and then making a program to emulate this may be alot harder then building a computer that can desribe an atom or many atoms working together to build a brain.. the blue prints for a thinking machine are written in the DNA, we emulate the DNA and let the atoms build it for us..? the only real hurdle other then raw computational power will be an accurate desription of the forces within the atom, and to my knowledge they are closing in on this idea. hopefully they will prove or disprove string theory in the next few years with the new accelerator turning on in may! woop.

#4 cyborgdreamer

  • Guest
  • 735 posts
  • 204
  • Location:In the wrong universe

Posted 29 February 2008 - 06:33 PM

There is the same problem of memory space either way. It would take the same (enormous) amount of storage space to simulate the atoms in a human whether we program it ourselves or grow it from a simulated embryo. In fact, assuming you wanted to grow it past the newborn baby stage, it would become even more taxing. You'd have to simulate the external world in great detail to give it the life experiances needed to develop intelligence. Finally, if you succeded in all this, you would still only have an ordinary human on a computer (which would still be pretty darn cool). It wouldn't help much in developing super human intelligence because we wouldn't understand how its brain worked any more than we understand our own.

#5 Cyberbrain

  • Guest, F@H
  • 1,755 posts
  • 2
  • Location:Thessaloniki, Greece

Posted 29 February 2008 - 07:35 PM

This was my initial interest and it's also what I want to base my line of research ... to create AI through (organic or inorganic) hardware, without the need to program AI like software. Kinda like the way DNA replicates it self.

sponsored ad

  • Advert

#6 fizzionz

  • Topic Starter
  • Guest, F@H
  • 32 posts
  • 0

Posted 01 March 2008 - 05:36 AM

i like to think of what it might mean to make a person who is alive inside of a computer... i mean.. the person would never be real, his intire atomic structure would be pure information. even the space he occupied would be information. and what if we find a way to make an interface with him. it would be the best game of sims ever.. to the argument that it might take aot of information to process.. well then lets take some ideas from the universe... i am a complex organisim.. but all the "work" of the processing does not get done on a single prcessor. every atom is its own processor only dealing with itself. but when we look at the macro scale of the structure, it appears to function as a single complex entity. the same could be done with this idea.. a vast network of processes.. a single atom of say hydrogen in my leg does not have to worry itself about a sulfer atom inside my eye..

the only thing i wonder is, what if we did get this idea working one day. but the it didnt work.. as if the virtual human had no soul.. or ghost in the machine... what would that mean for your universe?




1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users