Amazing demonstration. Now I get it and your opening post too!! Like in big 
text data, you want to count all occurrences of all Byte Pairs Codes ex. 'e' 
x46364 'the' x5735 'thank you' x39. Here, you can only count the occurrence 
'the' if all 3 letters are present together, and in a certain order in this 
case. Similarly, if seeing the past 3 Code contexts means the Next Code is 
'and' ex. "but [if we go] _" = "but [if we go] and" = "but if [we go and] _" 
then we MUST see all 3 at the same time, and in a certain order in this case. 
However James, my computer can do this, and therefore so can a Turing Machine. 
Any sensor in the real world can scan over a solid surface and collect 
stationary parallel data. If there's only a flash, a camera can pick it up. The 
data of looking at a grassy lake with your eyes enters in parallel and checks 
for co-occurrences but so can a Turing Tape machine if given extra time.

I see your point with the Universal Turing Machine. It forces it to be serial 
and isn't 'natural' as 3 dimensions, only 1D. I think it is the simplest 
computer though! You could have a 3D computer parallel, even the brain is a 3D 
Turing Tape computer however only does 2D I think (models 3D using 2D). But if 
you consider the skin as a 2D manifold wrapped around the body, maybe its 3D 
occurrences parallelly felt!! So the brain is a sort of 3D Turing Tape that is 
parallel. I suppose you could have a 3D serial Turing Tape, oh no you can't, 
only a 1D Turing Tape that mimics 3D co-occurrences.

Our 1D Turing Tapes today (and our 99999D GPUs) are able to store/compute 1s & 
0s from the real world, and output them to the real world, it works and can 
control the I/O of a robot tracker following a fast moving target like we seen. 
The real world data makes it work, its outputs affect the world correctly.
------------------------------------------
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/Tc33b8ed7189d2a18-M048d4401f429947c7ec3266a
Delivery options: https://agi.topicbox.com/groups/agi/subscription

Reply via email to