Artificial Intelligence

Published on August 5th, 2023 | by Madhudvisa dasa

13

AI (Artificial Intelligence) Is Stupid!

Today we read so much science fiction that is passed off as the “science.” We have been bombarded with movies and television shows warning us of the dangers of “intelligent” computers and now we are being told, “the intelligent computers are here now!”

However there are no intelligent computers, there is no artificial intelligence. Computers do not have the same abilities as people, or even animals. They are simply machines running code.

We are NOT Machines Running Code!

The big mistake of science is they don’t understand what intelligence really is. If you don’t understand something, if you have absolutely no idea at all how it works, it is impossible to imitate it. So if we were to create artificial intelligence we would first need to understand what real intelligence is and how it works.

The misconception of science is they try to explain everything on the basis of the material elements: earth, water, fire, air. That is all science knows. There is also ether, that is another material element, but science is ignorant of the ether. Ether is the space, it is the medium on which sound, light and all other frequencies travel. But scientists don’t know it exists…

Because scientists know so little they try to explain everything within there severely limited knowledge. Science will ferociously demand: “There is nothing but matter, everything exists because of random chance combinations of matter.” Science is a religion after all, trying to establish that there is no intelligent direction or cause or purpose of this material manifestation we find ourselves trapped inside.

Because scientists know about earth, water, fire and air, and they have some experience now with computers, computer programs, data and data processing scientists now imagine we are simply machines driven by data and some sort of organic computer system constructed and programmed randomly by chance combinations of earth, water, fire, air…

Of course this is ludicrous. Even if we were to entertain the idea that we are simply biological computers running code and processing data, then the obvious question would be “Who has created these machines, who is programming and maintaining them and why?”

So even if we were to go down this path it simply opens up questions about the creator and His creation. But scientists are so dull that these thoughts never cross their minds.

There is SO MUCH Science is Ignorant of!

Science goes on with its crippled conception that everything in existence is created from earth, water, fire and air and the creation just happed by random chance and everything is going on by random chance and there is no control or direction behind it. So this is science…

The problem for science is, of course, in reality we don’t see any practical organized structure developing as a result of random chance. If I told you, “That huge shopping center just developed out of random chance, one night there was a “big bang” and in the morning the shopping center was just there. There was no cause for it, it was not created for any purpose, and there was no management or direction controlling and supervising the construction.” Of course no one would ever believe it, “Shopping center just appears by chance after a big bang…” But that is what “science” expects us to believe happened in the case of the universe. The creation, maintenance and destruction of the universe and everything in it occurred simply by chance, with no control or direction or purpose…

So the point is science is stupid. They get up in front of the gullible public and speak complete nonsense and demand the people believe them. “The science is settled. Shut up…”

In reality scientists know nothing of intelligence and nothing beyond four of the five gross material elements.

Beyond the five material elements (earth, water, fire, air and ether) there are three subtle elements:

• Mind
• Intelligence
• False Ego

Science is totally ignorant of these subtle elements…

Mind, intelligence and false ego are also material elements, but on a more subtle platform, beyond the gross material elements (earth, water, fire, air and ether) and beyond the subtle elements is the spirit soul, the self, the “I”.

We are the spirit soul, we are not the material body. We, as the spirt soul, are eternal, and when this particular material body becomes old and useless we will be forcibly transferred into the womb of our next mother where we will develop our next material body…

Intelligence is NOT in the Brain!

The whole concept of Artificial Intelligence is to study how the brain works and try to imitate the workings of the brain. But unfortunately for the AI scientists, thinking, feeling and willing is not a function of the brain. The brain is simply a material interface and control system for the body. The brain is mechanically issuing the commands that control all the senses of the body and the unlimited activities going on within the body. And the brain also stores and processes sensory memories. So the brain is recording what we see, what we hear, what we touch, etc.

So the brain is only an interface to the other bodily systems and a recording device for storing and retrieving sensory memories. In the computer example the brain is not doing any processing of the data, at least not beyond what is necessary to control and monitor the senses and other bodily systems, and it stores all the sensory memories.

It is the mind and the intelligence, subtle material elements, which function to produce what we accept as intelligence.

Intelligence is not a result of the interaction of the gross material elements (earth, water, fire, air and ether), rather intelligence is a subtle element, beyond the detection and measurement of the scientists, who are only aware of the gross material elements.

The mind is very closely linked with the senses of the body. And the minds of most people and animals act almost totally on the basis of sense impulses. So when one of the senses of the body desires some object to gratify that sense the mind will tell the intelligence to get that thing. The stomach may send a message saying it is hungry, and that is a command to the mind to find something to eat. The eyes may see something beautiful so they send the command to the mind to go closer to that beautiful object so we can get a better look, etc.

So the mind is a very low-level sensory driven system feeding commands to the intelligence and, if the mind is uncontrolled, all these commands will be in regard to attempts meant to satisfy the senses of the body.

Above the mind there is the intelligence. The business of the intelligence is accepting and rejecting. So the intelligence is higher than the mind and has the ability of rejecting or accepting the demands of the mind.

So the intelligence has the power to control the mind. Stomach may say “eat”, but mind may say “no, you are too fat already, you don’t need to eat.” So the intelligence has the ability to process more advanced concepts. Mind works on the very basic animal instincts connected to satisfying the senses: eating, sleeping, sex and defense, but the intelligence is able to consider higher concepts and override the demands of the mind.

And above all this, and with the ultimate control and power of direction is the spirit soul, the self.

Although actually the soul should be in control, in most cases he is looking on as a horrified witness as the mind and intelligence act indecently of the control and direction of the soul.

So science is looking in the wrong place. The AI researchers are chopping up brains and trying to find out how intelligence works, but this attempt can never succeed, because intelligence is not in the brain.

The History of Artificial Intelligence?

Ever since man has had the hint of the possibility of computer systems, data processing systems, science has been telling us it is simply a matter of more processing power and more memory and inevitably computers will develop intelligence and capabilities similar to and ultimately exceeding the capabilities of humans.

We first started hearing this story in the 1950s and at that time there were huge computers, with less than a millionth the capacity of your phone, and in 1951  German inventor Dietrich Prinz created the first chess program of any type to run on an actual electronic computer. So this was the dawn of the age of Artificial Intelligence…

What followed was sixty years of disappointments for the AI researchers. With the advancement of computer technology they were able to write better chess programs and ultimately their chess programs became so good that they could defeat practically all the human opponents. But it is only code. It is just a set of instructions written by a programmer who has coded in the strategies and it is a “brute force” method that calculates every possible move and then possible moves by the opponent and then the possible moves by the computer, etc. So computer will calculate all possible moves by itself and opponent many moves ahead and then with probability and statistics will calculate what is the best move.

This is where computers excel. Data processing. No human could possibly calculate all the possible moves five or ten moves in advance, but the computer can. But this is not intelligence, and it does not in anyway whatsoever imitate what is going on in the mind of a human chess player.

So we had sixty years of failures in any attempt to create any sort of computer program that could act even remotely like a human in any respect at all.

Then, in the last ten years, something has dramatically changed.

The Discovery of Neural Nets

Serious development on the theories behind the AI (artificial intelligence) systems we see today began in the 1960’s and continued on. But, although the AI theories were taking shape, any meaningful AI model needs to process enormous amounts of data and at that time the available computer systems were not powerful enough.

The real breakthrough for AI only happened in the past ten years when someone discovered that the processing done by computer graphics cards was quite similar to the processing needed to train AI neural networks. The current AI neural net system is based on mathematics, particularly on matrix manipulation. So computer graphics cards are created for matrix manipulation at high speeds. A computer screen is a huge array of pixels that are stored in and manipulated by the graphics card as matrixes.

So the Nvidia graphics card became the AI neural net processor of choice.

How Does Artificial Intelligence Work?

When ordinary people type questions into Chat-GPT and the AI is gives them seemingly meaningful and relevant replies they are amazed. Because Chat-GPT interacts with us in a way that is similar to our interactions with people of course many come to the false conclusion that there is some intelligence in the machine, that it is conscious, that it is thinking, etc.

Nothing could be further from the truth. All the Chat-GPT AI is doing is, word by word, taking the words it already has and comparing those words with the billions of words it has in the database and, simply by mathematical calculations only, it determines what is the most likely word to come next.

So this is just a program using mathematical formulas and matrix manipulations and because of this development in the technology and software, the AI model is able to interrogate HUGE amounts of data and determine what the most likely next word should be.

So actually what Chat-GPT is doing is no different at all to what Amazon.com is doing, although the Amazon.com example may me more easy to comprehend.

In the case of Amazon.com they have a huge historical record of every interaction every customer has ever made with their website. This includes every product that every customer has viewed, what the customers have added to their carts and what they have ultimately purchased.

So these are sequences of events. Customer views a particular product, then performs some actions and then ultimately ends up purchasing a different product. And if a number of customers follow the same path, the AI will detect this when the model is being trained, simply as a similar sequence of numbers, and that will be stored in the AI model.

Then when a new customer views the same product, the AI program will do its matrix manipulation with that viewed product as the input and the AI will spit out a list of product numbers customer is most likely to purchase next based on the historical data.

Artificial Intelligence is Stupid

A fool may think, this Amazon AI computer is so intelligent. “It knows what I want to order before I do!” But this would be a mistake. There is no mystery here. All the computer is doing is looking at what products you have viewed and just finding in the database of billions of customer interactions the other similar customers who viewed similar products and who have a similar purchase history to you, then the AI simply suggests you should purchase the same products that other customers who have purchased similar products you have already purchased and have viewed similar products that you have viewed.

This is a very simple idea. If you have enough data and if you can quickly extract the data related to similar customers who are viewing similar products then you can quite accurately predict what this customer is likely to purchase next and you can make product recommendations that may be uncannily accurate. The Amazon AI can often recommend products that the customer wants to purchase that he himself had no idea he wanted to purchase!

So a foolish customer may be amazed by this. But it is just a mathematical formula and a huge amount of matrix processing to extract the predictions based on similar customers viewing similar items.

This is a basic idea, but in the past very difficult to implement. Because in the past interrogating huge datasets and quickly identifying and making predictions was practically impossible because the processing time would have been unreasonable. If you want to make a product recommendation you have to be able to get it quickly. Not that a week later you get the recommendation.

So there is nothing intelligent about AI. All neural nets are is a huge amount of data that has been “trained” according to a particular model.

There is no intelligence. Only mathematics and fast matrix manipulations made possible by Nvidia graphics cards. AI is completely stupid, not in any way at all conscious or sentient…

AI is Not New

So AI neural nets are not a new discovery. The idea has been around since the 1940s and the mathematics was solved in the 1970’s and 1980’s.

It is only recently, in the past 10 years, the discovery was made that computer graphics cards were far more efficient at training AI models than normal CPUs, that the training and interrogating of AI models containing enormous amounts of data became possible.

The Limits of AI

Although AI based on neural nets is a very powerful tool for analyzing and finding relationships between elements in huge datasets and AI enables processing and applications that were previously impossible, what AI can actually do is VERY LIMITED.

Training a big data neural net AI model is VERY SLOW.

For example to train a model like the one used by Chat GPT takes many months and uses an enormous amount of computer resources and power. They are using around 1000 processors running for months and spending tens of millions of dollars to train the model only once.

All the actual processing and discovery of the relationships between the different data items is done at the time of training.

The model must be very simple. In the case of Chat GPT all the model is doing is looing at the arrangement of words in the data and calculating the relationships between the words. And after spending all these months training the model all that model can do is predict the next most likely word based on a string of words you provide as input…

Neural network AI models can not be updated. So, for example, Chat GPT is incapable of learning anything as a result of its interactions with the users. Because the model can not be updated. All they can do is store the user interaction data and try to use it somehow as part of the input for training the next version of the model.

To update an AI neural net model you have to retrain it, in the case of Chat GPT, that means months of processing time on a thousand or more processors and spending tens of millions of dollars…

AI is Really Stupid

So hopefully you can understand that all the hype about AI is nothing but science fiction. The dreams and statements of the scientists and the misconceptions of the general public about AI, are so far from the reality that it is astonishing.

All we have is two things. We have got traditional coding, that I think most understand. It is a list of instructions coded by a programmer and the computer just follows that list. With traditional coding IBM created Deep Mind and many have developed computer chess programs, for example. So by coding you can very superficially imitate some small activities like playing chess or answering game show questions. And we have the more recent addition, AI neural nets, which is a very powerful way of using big data to predict what is the most likely outcome on one very specific item [like the most likely next product customer will purchase on Amazon.com or the most likely next word on Chat-GPT] but this is so limited and so much processing is required to train a large AI model to answer only one very specific question.

So at the current state of technological development we have absolutely nothing that even vaguely approaches anything at all like human intelligence and there is absolutely nothing on the horizon. So there is no hope of anything more, except developing the coding and the neural nets.

But what is the use of an AI that can only make sense out of one very tiny aspect of a huge amount of data after a model is developed and then trained. So six months after you have the question you will get the answer from your newly trained neural net model.

So there are some very specific applications that are suited to big data AI neural net models like product recommendations for an online store or predicting the next most likely word for a chatbot.

And of course neural nets and AI models are very powerful if you have big datasets and you want to make predictions about a very specific action or deviation from a predicted action.

Certainly smart developers can find powerful ways to leverage big datasets for many applications using AI neural nets. But the idea that this in any way whatsoever approaches human intelligence or is even similar in any way at all to human intelligence is totally incorrect.

Tags: AI, artificial intelligence, big data, neural nets

About the Author

Madhudvisa dasa My first contact with a Hare Krishna was a most merciful Mataji in Oxford Street, London who sold me a "Higher Taste" cook book in 1984 while I was on holidays there. I started seriously reading Srila Prabhupada's books in Australia 1985 and by 1986 Srila Prabhupada had convinced me "Krishna is the Supreme Personality of Godhead" and "we should surrender to Krishna." I joined the Hare Krishnas in Perth, Western Australia in 1986. Since then I have been chanting Hare Krishna, Hare Krishna, Krishna Krishna, Hare Hare/ Hare Rama, Hare Rama, Rama Rama, Hare Hare, reading and distributing Srila Prabhupada's books and preaching as much as I can. That's my life and full-time occupation now really. I like it more than anything I've ever experienced before. Srila Prabhupada's books are so amazing... Even after reading them all many times they're still fresh and new. They are truly transcendental! That's it really. Now I'm just hankering to once again see the world chant Hare Krishna, dance and feast and float away in the ecstasy of Lord Caitanya's Sankirtana movement as it did in Srila Prabhupada's physical presence. Let the whole world drown in the ecstatic flood of love of Krishna!

Related Posts

Can a Computer Have a Mind? →

Are Computer Games Really Conscious? →

Are Artificial Intelligence (AI) Devices Really Intelligent? →

Artificial Intelligence–Getting Computers to Think →