I am no expert on Artificial Intelligence, and I am not a fan of it either. I use it sometimes, mostly when there is no other choice. And I listen to all the hype and all the disaster stories that say the end is coming. But more than anything, I am a fan of being alive. And Artificial Intelligence is not alive. It may replace life, but it itself is not alive. It may replace life in the same way that the automobile replaced the horse. Horses are alive but cars are not. But they are both useful and efficient means of transportation. Just as a human who calculates a sum is alive but can be replaced by a calculator which is not alive.

But there’s something about horses that cars can’t replace. And there something about humans that AI can’t replace. There are still horses today, but not as many as before. And it seems the assumption with AI is that there will still be people, just not as many as before. And perhaps the open question is, who will be the people that remain and who will be the people that are gone.

But not even this is my main obsession about AI. My question pertains just to myself – not to the general population or to the future of humankind. My obsession is this: What is it about me that can’t be replaced by AI? And how can I hold on to that for dear life, while all around me changes?

The answer that I found is this: The part of myself that AI cannot and never will be able to replace is the part of myself that can’t be remembered.

Wait. Say what?  Please let me explain.

First, AI is made up of two very human things – talking, and memory. AI requires memory. Computer memory, yes, but still memory. Computers have several types of memory, which include memory of the programs that I am running, and memory of the data that the programs use to do their work. Of course this is simplistic, but the important point is that computers require memory – built-in memory and memory that is input as data.

The memory required by AI computers is astounding. We all know this. A random internet search produced the following from October 2024: “The first widely-based public version of ChatGPT … was trained on 175 billion different parameters.… [T]he enhancements over the past two years required many more parameters – some estimate as many as 1.8 trillion.”[i]

So AI requires memory. But AI also requires talking. It runs on talking. This is not obvious at first, but a quick glance might convince me that this is so. I interact with AI using talking. I ask questions, and the answers are more or less in talking-form. They could be in diagram or image form, but diagrams and images are really part of the broader human world of talking.

But there is more to it than that. AI itself is based on talking. It is, very simply, very, very accurate talking. AI works by predicting, with more and more accuracy, the next word that will appear in a string of words. And this simple process brings about all of the astounding results of AI. Again, this is a simplistic explanation, but it is essentially true. “When you prompt a Large Language Model with a sentence or a question, it uses what it has learned to predict the most likely next word or words to follow. This isn’t just a wild guess; it’s a calculated prediction based on the patterns and rules it has observed during its training.”[ii]

This part of AI is extremely and unexpectedly powerful, just as the ability to talk to each other made humans capable of such feats as building the Tower of Babel and going to the moon. And everything in between.

So AI is talking and memory. Memory and talking are also fundamental to how I see myself as a human being. Sometimes I feel and act as if that is all that I am. The power and danger of AI is that one day it will be much better at both of these than I am. As a human, my memory and my talking are not going to protect me from AI. My memory and my talking are not the “Part of Myself That AI Will Never Replace.” They are exactly the part of me that AI is capable of replacing. Talking and memory are what I use to define my individual human identity. AI threatens to usurp that identity.

So my question is this:  Is there a part of myself that AI will never replace? In other words, is there a part of myself that is not made up of talking and memory? I don’t mean my body – I mean a part of my psychological human existence that is not talking, and is not memory.

Again, say what?

For over 2500 years, we humans have been trying to get in touch with the part of ourselves, and the part of our world, that “can’t be talked about.” Beginning with Lao Tsu on down, the list is long and tedious:  From the early Taoists who said “The Tao that can be talked about is not the Eternal Tao,” to Plato’s Unwritten Doctrines, Pyrrho and his suspension of opinions, Plotinus and The One about which calling it “The One” is already saying too much, to the Neoplatonists including Pope Gregory who in 600AD wrote “anything I say about God is wrong, simply because I can say it”, to Zen Buddhism, to the Apophatic Theologians starting with Pseudo-Dionysius the Areopagite who described God not by saying what God is but by saying what God is not, to the Medieval Sufi Poets such as Rumi who wrote: “There is a voice without words – Listen”, to the Medieval Christian Mystics such as Meister Eckhart and Saint Francis who taught about “Beyond-God”, to the Pure Existence of Kant and Hegel and Nietzsche, the Collective Unconscious of Carl Jung, Be Here Now, Mindfulness, and even “The Unsayable” of Ludwig Wittgenstein.

These luminaries and many, many more have taken very seriously the idea that there is a part of ourselves, and a part of our world, that can’t be talked about. That is “outside of talking.”

Many humans have spent a lot of time puzzling over and seeking out the world that is outside of talking. I myself have been one of these. I spent 50 years, and full-time for the last decade, trying to figure this out. I just could not accept that it is simply a mysterious realm that is unrelated to my physical existence and experience. The problem is that it can’t be talked about, so my rational mind is kind of left out of the equation. Or is it?

I finally discovered that there is a way to rationally approach the part of the world that can’t be talked about. This is kind of in the tradition of the Apophatic Theologians who approached God by talking about what God isn’t, instead of talking about what God is. My version of this was to approach “the part of the world that can’t be talked about”, not by trying to talk about that part of the world, but by talking about talking itself. And in particular by talking about the limits of talking.

If I understand the limits of talking, I may be able to point myself in the direction of a non-verbal understanding (Heinlein used the word “Grok”) of this part of myself that so many of my fellow humans have believed exists, for so many thousands of years throughout human history.

What are the Limits of Talking?

First, talking requires many, many gaps. In fact, there must be a gap between any two words, so there are at least as many gaps as there are words. And that’s a lot of gaps!

Next, talking is one-dimensional. The rules say that words must be set out in a single line. Even a book that I can hold in my hands is really just one long line of words. If I wanted to, I could write out War and Peace in a single line down the middle of the roadway that would extend for miles! So talking is linear. It is one-dimensional.

Talking is also one-directional. I read words one-way from the beginning to the end. Reciting a sentence backwards like Mozart or saying the words at random is just confusion. It is not talking.

Without these rules, there is no talking. And because Artificial Intelligence is created with talking, it is also bound by these limits. It must have gaps, it must be linear, and it must be one-directional. But reality itself is not bound by these limits. How can talking, with its narrow and inescapable limits, fully reflect my remarkably complex and multi-faceted reality?

The very specific and rather narrow limits of talking make talking smaller than reality. And if talking is smaller than reality, there is a part of my reality outside the limits of talking. There’s a part of myself that can’t be talked about. It might be a world without gaps, with many dimensions, and with multiple directions. In other words, the real world!

Since AI is built on talking, and talking has these limits, I wondered whether “slipping around talking” might also be “slipping around AI.” Whether it might be a way to “grab the tiger by the tail” – and to hold on for dear life to a part of myself that can never be replaced by AI.

Then one day, in all this process, I realized that going beyond talking was not enough. There remained a part of me beyond talking that was still vulnerable to AI. And now I know what that is. AI is not only talking. AI is also memory. And so my question broadened to, what is the part of myself that is both outside of talking and beyond memory? Now that’s a question for the ages!!!

But I was ready for it. I had in my holster a weapon that was effective against talking, and now I wondered if it could also be effective against memory. That is, I conquered my dependence on talking by understanding the limits of talking, and by broadening my imagination beyond those limits. Perhaps I could also understand the limits of memory, and broaden my imagination beyond those limits. It was worth a try!

When I say “the part of myself that can’t be remembered,” I don’t just mean what we don’t remember. I literally mean what can’t be remembered. It is outside the limited reach of human memory. It is beyond the “limits of memory.”

What are the Limits of Memory?

It didn’t take much thought at all to start to tick off the limits of memory. They are familiar, they are always present, and they are simple to understand.

To start with, Memory is one-way. I remember my past but I don’t remember my future, much as I would like to. Memory is a lot like time, which has a past that has already happened so it is set and can be known, and it has a future, which hasn’t happened yet, so it is not set and can’t yet be known fully. So memory is tied to time. Time is a limit of memory.

Another obvious limit of memory is that it requires separate things. If nothing else, it requires a remember-er, and what is remembered. And it requires separate things that can be remembered. If there was just one thing to remember, it wouldn’t really be memory. And separate things are required for us to make comparisons and judgments. Separate things is a second limit of memory.

And finally, memory is intensely personal. I remember my own past experiences, and not those of others. We can “share” memories by talking about them of course. But It would be odd and a little scary if I could directly remember someone else’s past memories – perhaps it would be a good plot for the Twilight Zone! The main problem with remembering someone else’s past memories, is that I would not know who I am. Am I me? Or am I that other person? I would lose my identity. Individual identity is a third limit of memory.

I’m sure there are other limits of memory, but these will do well enough for the present. The important point may be that, as with talking, these very specific and rather narrow limits make memory smaller than reality. And if memory is smaller than reality, then there is a part of my reality that can’t be remembered. There is a part of myself that is outside the limits of memory. And if that part of myself is outside the limits of memory, it is also outside the reach of Artificial Intelligence!

Another aspect of going beyond the limits of memory, and therefore slipping past the sticky fingers of AI, is this: Memory accounts for much more than just my own personal memories. Memory is also responsible for what I have come to call “the products of memory.” The products of memory are the elements of human experience that I rely on for my daily existence and interaction, but that would not exist except for memory. For example, trees and mountains and birds would still exist if I had no memory. But houses and cars and jobs and books and bank accounts would not.

I call these “products of memory,” because they are made of memory but become identifiable “things” in my human life, things I rely on and use and am subject to (such as the account balance at my bank). The products of memory are the tools that we use to create, control and benefit from the amazing edifice of modern human life.

Perhaps the most important product of memory is talking itself. Talking could not exist without memory. In order to talk, I have to remember individual words, their sounds and their spelling, their definitions, and the rules for stringing them together in a meaningful way. In other words, all the things that I was tortured with by my grade school English teachers! And not only that, but you and I have to remember the same things about talking – words, definitions, rules –  if we are to be able to communicate with each other. Then I have to remember what I said before, and what others have said before. And remember what I am going to say next! Without the ability to remember, talking itself would simply not happen. So talking is a product of memory. 100%

But talking is not the only product of memory. The list is endless, and these are just a smattering:  Writing, thinking, symbols, mathematics, rules, plans and procedures, coordination and agreements, contracts and laws, art, music, culture, institutions, religion, trade, money, ownership, technology, institutional identity, indeed the very agreed rules and complex cooperation of Civilization. Not only this, but the things of human intelligence all require Memory. Things like knowledge, learning, theories, definitions, descriptions, explanations, opinions, categories, comparison, analysis. And even things related to emotion such as expectations and knowledge of feelings, and things related to perception itself, such as recognition.

These are things that I would not have without memory. They are also things without which I could not participate in civilized human life. This shows the incredible importance of memory to human existence. Without memory, I would have none of these things. Because these things are built by memory, they are also things that can be created by AI, which operates by controlling talking and memory.

And this is the main worry I have. Not only that AI will take over creating these things, which is a part of AI that is kind of enjoyable right now — it does a good job, certainly better than I could do on most things. But AI can not only create the “things of memory.” It can also control the things of memory. And if AI starts to control my own products of memory – from my bank account to my car to my job to my communications – then AI will control me. AI will “take over.”

So my important worry about AI is this: That AI will take over control of the “things of memory” that make up my human life and my participation in human civilization. What can I do about AI taking over my things of human memory? Well, just as with talking, perhaps I can find the part of myself that is beyond the limits of memory, and so cannot be controlled by AI. If I as an individual can relate to the part of myself and the part of the world that “can’t be talked about”; If I can do that by understanding the limits of talking, and then expanding my ability to perceive myself beyond the limits of talking; If I can use metaphor to open up imagination to the part of my world that can’t be talked about; Does the same go for memory?

Unlike talking, there is no ancient tradition of moving beyond memory. Of reaching for the part of myself and the part of my world that “can’t be remembered.” The idea seems to have been first mentioned in public in the late 1900’s by J. Krishnamurti. He however thought that the idea was “too radical,” and he refused to talk about it even when prompted.[iii] To my knowledge, there has been no investigation of this idea since. So this is really a brand new area of exploration, and one that may be quite ripe in the context of Artificial Intelligence and human consciousness.

The new idea is this: The part of myself that will never be replaced by AI is the part of myself that is “outside of talking and beyond memory.” It is beyond the limits of memory. That we can experience this part of ourselves, the part that can’t be remembered, is not a given. We have to explore it. We have to realize that we experience it every moment, and so remembering it is not important. It may not even be possible since current experience may override any memory we might have of the past experience.

In other words, the true “higher” part of myself is always with me. I am continually “experiencing” that part of myself. If I am continually experiencing it, then I don’t have to remember it. The fact that it “can’t be remembered” is not a liability.  But it does make it difficult to recognize, and it makes it easy to ignore, since we are so addicted to memory and to talking about things. And we mistakenly believe that is all we are as humans. But we can open to it, with a whole-body and a soft heart, just as we have been taught from time immemorial.

What might the part of myself that is beyond memory be like? How can I possibly relate to it? Well, I sat down and had a little think about that. And here is what I came up with:  We’ve already seen that the limits of memory are at least these — they are required for memory, and they are also required for Artificial Intelligence: Memory requires time. Memory requires separate entities. Memory requires individual identity.

These don’t have to be seen in a spiritual way, but they can be. In my life-long search for the completeness of human experience, I have heard many times the metaphorical descriptions of the higher world of human experience, the world of my own higher self. Among those descriptions I regularly found these elements: That part of the world is Timeless — it is Eternal. It has no Beginning and no End — it is Infinite. It is the Oneness of All Things — it is Non-Dualistic. It has no comparisons and no judgments – it is Universal Love. It contains no individual identity.

These exactly describe what I might expect to find in the part of my world that is beyond the limits of memory. In other words, I may have come to understand that the world of my higher self, the world that humans have sought for millennia, is simply the part of me and the part of my world that is not only outside of talking, but is also “beyond memory.” It is beyond the limits of memory, and beyond the “products of memory.” And I may have come to realize that this part of myself that is beyond the limits of memory, may in fact be: The Part of Myself that AI Will Never Replace

So there is a part of myself that can’t be remembered. It can’t be remembered because memory has limits, and so who I am is bigger than memory. It extends beyond memory. It is who and what I am, and I can’t do anything but experience it. It just can’t be remembered. But remembering it is not so important, because I am always experiencing it. Why would I want or need to remember it, if I am constantly experiencing it? And because it can’t be remembered, my true self is the part of me that Artificial Intelligence will never replace.

If I get in touch with the part of myself that is beyond memory, even if that part of me is beyond the reach of Artificial Intelligence, that still will not protect my own personal “products of memory” in my daily life from AI. What it might do, however, is to encourage me, and perhaps encourage others, to remember that “who I am” as a human entity is beyond my personal products of memory. And that the effect that AI has on my memory and its products does not affect the true essence of who I am.

I know this all sounds pretty wild. And that is why I say it is grabbing the tiger by the tail. It promises to take us for a memorable ride! But perhaps that is the lesson for these times. We have been so lackadaisical for so long about letting ourselves identify with the things that are not truly part of who we are as human beings. Only a few of us have put in the time, effort and sacrifice necessary for moving our deep identity over into the part of us that is the true core of who we are.

Perhaps now we are all being forced to do that, since we are in the process of turning over our talking-selves and memory-selves to the machines. If we worry that there is nothing left after we do that, perhaps we only need to be reminded again of what we have been trying to remind ourselves for millennia. That our talking-selves and memory-selves are not who we are.

And perhaps it is time that we, as a species, will finally come to terms with who we truly are as humans. And to realize that the tools that we create, including talking, memory and Artificial Intelligence itself, are meant to serve and nourish our true selves, and not the other way around.

…………………………………………

[i] James Myers, “Is Data-Intensive AI Facing a Memory Limit? New Approaches Might Provide Solutions”, Science News, October 18, 2024, https://thequantumrecord.com/science-news/is-data-intensive-ai-facing-a-memory-limit/

[ii] Code Signal,” LLMs Are Next Word Prediction Machines,” https://codesignal.com/learn/courses/understanding-llms-and-basic-prompting-techniques/lessons/llms-are-next-word-prediction-machines

[iii] J. Krishnamurti, “There is nothing to learn about yourself,” Krishnamurti Foundation Trust 1978, https://youtu.be/FWuD1Sh1GYY?si=1BmBn4LcIznQXy-C, minute 2:50, 4:38.

……………………………………

“Alexander Talby” is the nom-de-plume of a former Director of International Affairs, and international treaty negotiator. He has engaged in a lifelong study (full time for the last 10 years) of the human “higher self” as an integral part of the everyday life that we know and love.

Artificial Intelligence, Consciousness, Identity, Mind, Robotics