-- A blog maintained by a pessimistic over-confident High-School kid.

Sunday, July 12, 2020

Recursion Performed in Neurons - AI

It is Sunday, and I lost my urge to work again. I think the very fact that I know today is a Sunday made that happen. But nevertheless, this could give me an opportunity to refine a thought I have since 2 or 3 years ago. And don't mind the title, as I am just trying to criticize the running joke about people on the internet using big words to assume their authority; there is an unhealthy amount of fake mediocre gurus on the internet that flash terms and sell courses. I digress, the thought that had been following me had been the psychology and the workings of the human brain. Consciousness as I see it is anything that makes decisions out of order. In human, that order is the strict rules of our neurons. In other animals (in this post will be abbreviated to just animals), it is a smaller network of similar neurons. And in a computer, it is the combination of transistors and resistors that creates patterns from rhythm. 

The Absence of Language

The distinction between the decisions made by human consciousness and animal consciousness is their access to language. We are able to voice our consciousness into words and verbal thoughts. Animals while proven to have all the tools to memory, sight, and some, even imagination, do not have language and could not articulate and systematically expand upon their existing thoughts. In result, they only rely on abstract thoughts1 of things that may be pictures or other things. Lastly, computer consciousness does not even have this abstract thought, and they operate on the basis of 1s and 0s, similar to an unorganized group of neurons.

We are not too different from computers:
  1. We have our transistors in the form of neurons
  2. We have motherboards, ram, power supply, a whole computer in the form of our brain
  3. We have programs in the form of the pre-existing algorithms in our brain that help us understand sight (into outlines of objects), filter different frequencies of soundwaves (to hear dialogues in a noisy room) and much more
  4. Lastly, we also have an internal clock in the form of the Cereblem to maintain our heartbeat and breathing.
Our brain is a computer that is built from pure trial and error through evolution and the rule of "survival of the fittest". Computers just lack the language and efficient algorithms for sight and movement.

Necessary Algorithms

Aside from integrating sight, touch, hearing, smell, taste together in the frontal lobe of the brain, (hypothesis) our brain has algorithms that simplify data before it reaches the "consciousness". We see outlines of objects and shadows, not light rays of different frequencies. I say there is a "pre-wired" algorithms written in neurons that parse light rays into outlines that could be understood by our brain. The list of algorithms that I have thought of are as below: 

  • Vision
    • Outlining algorithm
    • (imaginative) 3d system (turn outline into objects in 3D, similar to Tesla auto-pilot (without predefined 3D models)
  • Movement
    • Balance - sensing the muscle used to hold the body upright and figure out the relative gravity
  • Hearing
    • Distinctly separate different sound waves into voices, music, hums, etc.
  • Storage
    • Algorithm of short-term memory to long-term [things in long-term are naturally deemed more useful as its repeated] -> maybe just as the type of storage used, similar to ram, and is not an algorithm
  • ?
    • Smell
    • Touch
    • taste

In the current state of AI, we mostly have a premature state for each one of the algorithms. The problem is that it does not have a generalized output, and one algorithm cannot communicate with others. Not to mention that each one of the algorithm is not polished yet. While AI can differentiate between objects, their input is of bits, and it is almost impossible for us to manipulate the neural net to our will.

The solution I prematurely suggest is to write the algorithm using Math ourselves. Not only will it allow us to manipulate it to our heart's content, but we could also tweak it to maybe "ensure" the power level of the AI. These kinds of algorithms are easy to conceptualize but difficult to "create". And I am not encouraging the development of it, but only illustrating its need in making a "conscious" (to the human level) system.

These notes are from last year, so if you are going to nitpick
I am just going to say these are not my full thoughts now.

Plus, this is just part of the story anyway. 


My own purpose in attempting to understand the brain is to create a structure for computers to process data. Yes, it is very far-fetched and definitely out of my league, but what is the harm in trying. While it is a guaranteed failure, this thought process had been going on for years, and there is no reason to stop it now.

After knowing that I want to understand the brain to build J.A.R.V.I.S, the consequential question is why. My imperfect argument is as follows, correct me if I am wrong. To the trend of global education, there is a pattern of selective teaching, where the people with a head start is given proprietary treatment as schools want bulletin examples. This is building up towards a generation with a very distinct gap. I put it roughly, a youtube (shows how discredited I am) video illustrates "education now is about no one left behind. But just a minute ago you were talking about giving kids a head start. Head start, left behind. Head start left behind. Someone is losing ______ ground here." As annoyingly biased viewpoint it illustrates, it is true to a certain degree. While everyone is provided with education nowadays, the favouritism of the students with a head start creates a huge gap between groups. (I'm going to skip to the point, I want to write about the brain and not why AI is needed). An AI can fill in jobs that are undesirable. The missing ingredient to communism is that people would not to work, and with AI, people won't need to work. And if Ultron is created, then is that not just evolution.
    ^^ There is an incentive for the government to dumb people down, to continue producing the people needed for the cycle of buyers and sellers. If people in the United States decide to save money, the debt of large companies and a line of many other things would fall. The debt, and the structure of business consisting of pure design strived upon the buyers with "BS jobs" (that do nothing but earn and spend to keep the cycle going).

Other arguments would be comparing it similar to a nuke. If a nuke is feasibly possible, then we would probably want to unravel it and demand for regulation before any one evil entity gets to it first.

Another argument is that when we do hit a roadblock in the struggle of our existence (In movies, that would be figuring a way to become multi-planetary), a AI that is multiples of magnitudes better than humans would become useful.


I have many thoughts, hypothesis and attempts that I had not shared. But I have been writing this post for 2 hours now, and I would stop this post here as I am getting tired. Because of the size of this goal, this conception would fill out my boring times indefinitely. It is the ingredient that allows me to sit in a chair silently alone for hours without end. Avoiding boredness had not been easier.

Other notes:

 It had occurred to me that I had not alluded to the title any bit. The title of this post is illustrating how we imagine scenarios from memory. My hypothesis is that the inputs to our vision not only includes our eyesight but our memory and logic (frontal lobe) as well. We take the memory of the inputs required for a certain image, and process that back into the input of our visual cortex, otherwise known as the occipital lob. This occipital lob would then transfer the output back into the frontal lobe to process the image and make it look real. 

There is a possibility that it skips the occipital lobe all together and wire the output from the frontal lobe right back into it, but this is my hypothesis.  

Other Parts of the brain:

  • "clock"
    • Breathing
    • Heartbeat
    • Reflexive response


  • The human brain while are composed of a network of neurons like a neural net, has "hot wired" aspects that are evolved from evolution
  • Our brain is more hard-wired then it is adaptive

Same thing, this is from last year


P.S: The reason that I am using headings in this post is that this idea is too messy in my head, and I can't articulate it easily.

PS: My eyes are dying from staring at the screen to much. This problem always takes my breath from me.

1: The abstract thoughts in animals are hard to explain. I would describe these abstract thoughts in humans as the knowledge that do not have explanation, as we ourselves had not evolved a language for it yet. Some examples of this include, the knowledge of riding a bike, touch-typing, shooting a basketball consistently, a photographic memory, etc. All these knowledge cannot be explained and articulate from one individual to another, as our existing language would not allow us to describe our muscle memory (it might not be muscle memory, this is only one way to see it).
    All of these are of speculation, and one could argue that there is nothing that would be able to articulate the knowledge of touch-typing. But it is also true that it's difficult to imagine the unknown. (it is like how a person from a 2D world can't understand 3D. And we ourselves in the 3D world cannot imagine what exactly is 4D).


  1. WoW, you are evolving, from Aldrin 1.0 to Aldrin 2.0.

  2. PUBG slots - 2021 bonus codes - Casino-Site
    Check out the latest PUBG slots from PUBG now! a wide range of slot machines and scratch cards, and 바카라사이트 they have a massive range 우리카지노 of bonus features!