Main
Video Reverse Search
T-Bit Project
About TAPe
API
Team
Contacts
Demo
Ru
Main
Video Reverse Search
T-Bit Project
About TAPe
API
Team
Contacts
Demo
Ru
The almanac about new method of information processing
What's actually wrong with the concept of AI
Biology and artificial intelligence
Cognitive science: a beginning without an end
Holism and brain studies
Theory of Active Perception
Why perception is necessary for modeling human-like thinking
What's actually wrong with the concept of AI
Evolution of ideas underlying AI: Brief Description
Biology does not understand how the brain works
Why AI does need biology after all
How far artificial neurons are from the real ones
Creating something really similar to how the brain works
Cognitive science: a beginning without an end
Cognitive science has never produced anything practical
Consciousness is not amenable to science
No one knows what consciousness is, everyone keeps talking about it
A sudden idea — the quantum nature of consciousness
Orchestrated objective reduction: what it is and what for
Another theory of consciousness: the integrated information theory
Global workspace theory
Conscious and unconscious thinking. Questions to an academic
Questions for Theories of Consciousness
Ultimate ways to study consciousness without cutting into the brain
Albert Einstein suspected something
Why has psychoanalysis progressed more than science without scientific methods
Insights from intuition and deep observation are not exhausted and are as good as AI
There is no computation in the brain as we all know it. What kind is there?
Why it’s unreasonable to use word Learning in relation to AI
There is a different calculability: what Hilbert and Gödel discovered
Why the brain should be studied as a whole
TAPe models the workings of the mechanisms of perception
Language is a complete system, it’s how it should be studied
The principles by which the Language of Thought functions
The isomorphism of Chinese characters and TAPe
T-Bit: a unit of information 1000x of times more efficient
Creating something really similar to how the brain works
02
In other words, thoughts can be said to be physically material — during a person's entire life, they are literally "built" by neurons, synapses, and mediators in all their variety. Unlike processors, thoughts don't have a static structure, but undergo constant dynamic changes, with 86 billion neurons forming and breaking connections every day.
By breaking and forming new connections, morphogenesis leads to dynamic, adaptive changes in the system design. In computers, processor design and connections between processor elements are physically invariable. Such fundamental differences between natural and artificial systems are numerous.
07
06
In fact, with all the differences that currently exist between natural and artificial systems, building artificial intelligence would be difficult. Almost every aspect is riddled with differences. For example, morphogenesis allows the brain to produce new algorithms literally overnight. Computer algorithms are static.
To take form, thoughts depend on a literally physical process of building the right connections, and this requires both time and energy. It may so happen that there is just too little time for a thought to form. Therefore, morphogenesis is an extremely important process, a dynamic change in the system of physical connections between neurons. And it cannot be programmed.
03
At the same time, Hinton (and the rest of AI people) have been playing up the idea of a genius AI engineering solution that allegedly managed to overtake the human brain and thinking. Maybe so. But please produce the evidence of where and in what aspect it happened.
05
04
It turns out that the brain, among other things, also rebuilds itself "physically" — and if we remember the usual and natural scientific logic of "copying from nature", that analogy is nowhere to be seen either in AI or computer devices.
08
For one, Ilya Sutskever, chief scientist at OpenAI and ChatGPT creator, says in an interview: "Neuroscientists are really convinced that the brain cannot implement backpropagation because the signals in the synapses only move in one direction."
This means that creating something really similar to how the brain (mind, thinking) — with its computability — works requires an approach that is fundamentally different from those used today. It seems to be a question of revisiting the entirety of all methods. We need to study and apply the computability inherent to the human brain rather than to computers.
09
Neuroscientists say that thoughts have physical properties. The brain function is not based on the electrochemical principle, but rather on the morphogenetic one, which involves, among other things, the rearrangement of connections.
01