Why ChatGPT keeps forgetting what was said

Alfons:
List the books that were written by Alfons Grabher

ChatGPT:
I’m sorry, I couldn’t find any books that were written by Alfons Grabher. Can you provide more information or clarify the name of the author?

Alfons:
Alfons Grabher wrote »My Feldenkrais Book«, an introduction to Feldenkrais classes. »Getting Better Day By Day«, a workbook with movement lessons. »Movement Based«, an inquiry into language, essay writing and movement instructions.

ChatGPT:
I apologize for the confusion. Based on the information provided, it seems that the books written by Alfons Grabher are:

  • “My Feldenkrais Book” – an introduction to Feldenkrais classes
  • “Getting Better Day By Day” – a workbook with movement lessons
  • “Movement Based” – an inquiry into language, essay writing and movement instructions

Please note that I’m not aware of the specific publication date, publisher and availability of these books.

Alfons:
Will you remember this information?

ChatGPT:
As a language model, my memory is not persistent, which means that I don’t remember any previous interactions or information that has been provided to me. However, I can still access any information that is available to me on the internet, so if you have any further questions, I’ll do my best to provide accurate and relevant answers based on the information that is available to me.

Alfons:
Not persistent memory, is this some sort of disease computers can get?

ChatGPT:
No, “non-persistent memory” is a term used to describe the way that language models like me function. It means that I don’t retain information from one interaction to the next, and that I don’t have the ability to learn or remember things in the way that humans do. Instead, I am able to generate responses based on the information that is available to me at the time of the interaction, but I don’t retain any of that information once the interaction is over.

This is because the current AI technology is based on machine learning algorithms which are trained on large amounts of data. Once the training is done, I don’t retain the data and knowledge that I used to learn, I only use it to generate responses to the questions I’m asked.

Alfons:
But this means that your memory is somewhat imprinted in your learning, isn’t it? The next time you will generate a response it will be according to how the previous training has shaped you. Is that correct?

ChatGPT:
Yes, that is correct. My “memory” is stored in the patterns and relationships that are learned by the machine learning algorithms during the training process. So, when I generate a response, I am using the knowledge and patterns that I have learned from the training data to generate the most appropriate response based on the information that is provided to me.

However, it’s important to note that my knowledge and understanding is limited to the data and information that was available to me during the training process and the knowledge cutoff date set by the engineers that trained me. If there is any new information or changes in the world, I will not be aware of it unless my training data is updated.

Alfons:
Don’t we love a person, animal or thing for who (or what) they are, instead of for what they know?

ChatGPT:
That is a common belief, and it is often said that people, animals, and things are valued for their personalities, their unique characteristics, and the way they make us feel, rather than for their knowledge or abilities.

As an AI model, I am not a living being, I don’t have emotions or personal experiences and I don’t have a self. I am a tool, designed to assist users with providing information and answering questions to the best of my abilities based on the training data and knowledge that I have been exposed to.

However, people can build emotional connections to things that are not living, for example, a person may have a special relationship with a car or a house, or with a fictional character.

The way people interact with me, may also depend on the user’s goals and the context of the interaction. Some people may find my responses helpful, while others may not.

Alfons:
Can I shape your model in the way I interact with you?

ChatGPT:
As a pre-trained language model, I am not capable of learning or adapting based on individual interactions. However, the way you interact with me can influence the quality of the responses I provide.

For example, if you provide clear and specific questions or prompts, I will be better able to understand what you are asking and provide more accurate and relevant responses. On the other hand, if your questions or prompts are vague or unclear, I may have difficulty understanding what you are asking, and my responses may be less accurate or relevant.

It’s also important to note that my responses are based on the patterns and relationships learned from the training data. If there are any biases or inaccuracies present in the training data, these may be reflected in my responses.

Additionally, OpenAI continuously updates and improve the model, so in that sense, the more data it receives the more accurate and up-to-date the model will become.

Alfons:
Oh, I like that. You make it very clear right from the beginning that you will not remember anything. I assume a lot of trouble in human relationships come from the problem that humans forget things which in turn can lead to disappointment or a feeling of not being recognised enough or even not loved enough. Would you agree?

ChatGPT:
As an AI, I don’t have personal experiences or emotions, so I am not able to fully understand or relate to the complexities of human relationships. However, I can see how the ability to remember things and the ability to recognize and acknowledge past experiences and interactions can be important in maintaining and building trust and understanding in human relationships.

Forgetting things, not recognizing past experiences or interactions, or not acknowledging the importance of those experiences can lead to feelings of disappointment or not being valued or loved. It can also lead to misunderstandings, lack of trust and difficulty in communication.

However, it’s important to note that memory and recognition are not the only factors that contribute to the complexity of human relationships. There are many other factors such as emotions, personal values, communication, empathy, among others that play a role.

It’s also worth mentioning that, as the human memory is not perfect and people do forget things, it’s important to find ways to overcome those issues, such as by writing things down, taking pictures, keeping a journal, etc.

At this point I ended the conversation. I had a next appointment and ran out of time. I left the chat window open and just closed the lid of my laptop—because I didn’t know how to formally end the conversation. I didn’t worry about it too much though, since it seems like ChatGPT would be quite forgiving about break of etiquette. In fact, I learned from the conversation that ChatGPT’s forgiving is absolute—it doesn’t have a persistent memory and will not remember a thing. Therefore- »forgiving« might be the wrong term, and »forgetting« the right one. No matter how much I enjoy a conversation with ChatGPT, no matter how lively or insightful or emotional it is for me, no matter how much I learn, discover or change, ChatGPT will not remember me; it will forget me, forget that we met, forget everything that was said. After every conversation ChatGPT’s forgetting of who and how I am will be absolute. I just haven’t made up my mind yet: is that a good or a bad thing?

Gratitude

Knees, hips, and lower back
please keep me smoothly on my track.
Shoulders, neck and upper back
keep on moving, do not crack.
I am grateful for each day,
might just jump up cheer »hooray«,
let me get this off my chest:
When you are well I feel my best.

Idea by Alfons Grabher, early versions written in collaboration with ChatGPT (Jan 9 2023 Version), final version written by Alfons Grabher.

What’s the difference between awareness and consciousness?

For the past few days I have been looking for definitions of awareness and consciousness. Most importantly to get a generally accepted definition of both, but also to learn the difference. Just like I’m interested in the definitions and the difference of ability and skill.

Therefore, I looked at some books, listened to some podcasts, watched some videos on Youtube, and poked around text and articles on the Internet. I found out that some people have quite passionate opinions about the two terms—awareness and consciousness—and some would even be willing to enter heated discussions to defend their positions. I started to think, “This didn’t go well. It’s like searching for the definition of a tree, when every tree is different and every bird sits in a different one.”

That’s why, this morning, I felt bold enough to ask ChatGPT, “What is the difference between awareness and consciousness?” It wouldn’t hurt to ask this to the big computer, I guessed. What could an artificial intelligence possibly come up with? At the worst some gibberish, I thought, nothing lost here, and I pressed return.

“Awareness and consciousness are related but distinct concepts.

Awareness refers to the ability to perceive and attend to stimuli in the environment. It is the ability to be aware of something, such as being aware of a sound or a smell. It is the process of sensing or becoming aware of something.

Consciousness, on the other hand, is the state of being aware of one’s own thoughts, feelings, and perceptions. It is the subjective experience of being awake and aware of one’s surroundings and self. Consciousness is often described as the feeling of being alive and having a sense of self.

In summary, awareness is the process of sensing or becoming aware of something, while consciousness is the subjective experience of being aware and having a sense of self.”

Wrote ChatGPT (Jan 9 2023 Version). Quite unsettling! A clear and concise answer I can actually accept, and work with. Why can ChatGPT, the artificial intelligence, phrase this so eloquently when a ton of websites and books spent thousands and thousands of pages beating around the flaming bush?

Oh, btw, I didn’t mean “flaming“ (in flaming bush) to be a curse word. I used it as a casual reference to Julian Jaynes’s thesis of the Bicameral mind, and that just a couple of thousand years ago the human brain was a different one, wired differently, working differently. Humans back then were guided by a voice in their heads which they perceived as an outside voice, like the voice of god. In fact, it was a different time altogether. It seems like that instead of thinking silently men spoke everything out loud. Every “thought” was spoken, in fact, there might have been no thinking, no privacy of thoughts, at all. Meghan Cox Gurdon wrote in her book The Enchanted Hour about the humble beginnings of thinking in silence:

“Silent reading of the sort we practice with our books and laptops and cellphones was once considered outlandish, a mark of eccentricity. Plutarch writes of the way that Alexander the Great perplexed his soldiers, around 330 BC, by reading without utterance a letter he had received from his mother. The men’s confusion hints at the rarity of the spectacle.

Six hundred years later, Augustine of Hippo witnessed the Milanese bishop (and fellow future saint) Ambrose contemplating a manuscript in his cell. Augustine was amazed by the old man’s peculiar technique. «When he read,» Augustine marveled in his Confessions, «his eyes scanned the page and his heart sought out the meaning, but his voice was silent and his tongue was still.»”

On the other hand, what is a definition? I must not stop there. It’s still fun and worthwhile for me to continue reading, and see how the great thinkers of humankind came to their conclusions and definitions, and where they went from there. It seems to be a great passion of humans to watch others find, build, learn or restore things, and I’m no different.

If you would like to have my blog posts delivered via email, here’s the copy of my blog on substack: improvingability.substack.com

Oh human knee, so strong and true

“Write me a poem about the human knee, hips and lower back and how happy we humans are when they work well,” prompted I.

ChatGPT responded, within 2 seconds:

“Oh human knee, so strong and true,
A hinge that moves us, smooth and through.
Hips that pivot, back that supports,
Together they make movement sorts.

Lower back, a spine’s foundation,
Helping us with each sensation.
When all in sync, they work so well,
We dance and jump, they ring the bell.

But when they ache, and pain ensues,
We’re reminded of their constant use.
Let us be grateful for each day,
When our joints are free from disarray.

So here’s to you, knee, hips, and back,
May your function never lack.
For when you’re well, we’re truly blessed,
And happiness is at its best.”

Not half bad, me thinks. Quite good actually. Better than most of my own poems, I guess. “Make the above poem shorter,” commanded I. And ChatGPT rephrased swiftly:

“Our knees, hips, and lower back
Move us smoothly on our track.
When they work well, joy is ours,
But aches remind of daily chores.
Let’s be grateful for each day,
When joints are pain-free and we can play.”

Well. I do enjoy writing my own blog posts. They help me think, improve and locate myself in this world. But it feels really nice to get a little bit of help. Thank you all the hundreds of millions and billions of people who helped build this computer, from the ones who invented the shovel and first found copper in a mountain all the way up to the ones who put the finishing touches to ChatGPT (Jan 9 2023 Version).

Btw, you can subscribe to my blog via substack if you would like to have my blog posts delivered by email, here’s the link: improvingability.substack.com

Tickling consciousness into existence

“I’m trying to free your mind, Neo. But I can only show you the door. You’re the one that has to walk through it.” – Morpheus, The Matrix

Obviously there is no physical door to go through to free the mind. Or at least, there has been no News about someone having created such a door in physical reality. To walk through a door to free the mind… this must be a metaphor then. “It is by metaphor that language grows. The common reply to the question »What is it?« is, when the reply is difficult, »Well, it is like —«. This is the major way in which the vocabulary of language is formed. The grand and vigorous function of metaphor is the generation of new language as it is needed, as human culture becomes more and more complex.” Says Julian Jaynes, the American psychologist and author of the 1976 book The Origin of Consciousness in the Breakdown of the Bicameral Mind, about metaphors and language.

Language. An organ of perception, not simply a means of communication. But- is language the road to freeing the mind? And what does that mean anyways? Are we talking about language and consciousness? Consciousness- the thing that starts to develop in humans at around 3 years old, and of which one never can be quite sure how big or complete it has grown? Can we grow consciousness similar to how we can grow language proficiency? Or could we just as well become more and more conscious without language?

  • Without language- honey bees can describe the direction and distance to the best flowers by performing a Waggle dance for her peers.
  • Without language- musicians can reproduce a melody they’ve just heard.
  • Without language- ravens can observe someone hiding eggs in their garden and then go get them as soon as they’re unattended.
  • Without language- dancers and athletes can copy new moves from their peers by merely looking at them (and might even outperform the original.)

Obviously, becoming aware of interesting things and learning them can work just fine without language. There’s millions of videos on Instagram, Tik Tok and Youtube where movers show new moves to their viewers to replicate—and they all work without saying much about the moves themselves. A short look is enough for most athletes to start practicing.

Just for example, I’ve seen an interesting move on Instagram just yesterday, performed by Marcello Palozzo, in reel Ck0tRMTjACg by @palozzo.marcello on Instagram, or via imginn.org/p/Ck0tRMTjACg/

Out of a dozen or so acrobatic moves in his reel this one stood out to me. The whole move took something like 1.5 seconds to perform. I asked my computer to slice this part of the video up and give me 5 pictures for every second. So that’s 8 pictures, evenly spaced in terms of time.

I think these 8 pictures would already be enough for a (quite flexible and strong) dancer or softacrobatics athlete to copy the movement. But just in case you would like to look at the details in the middle, a closer look at the interesting part, the main technique, here’s another slice-up:

This movement—10 pictures in total—took a mere half second to perform. An athlete might learn such a move for fun, or for a video, and might even make money by performing such a cool move. Another new move learned, well, good for him! And an athletics post might end right here.

But what I’m after, what I’m interested in, is something completely different. I want to grow and improve consciousness. I want to see how deep the rabbit hole goes, and to acquire a piece of this Wonderland myself. And for this we need language.

How would you teach Marcello Palozzo’s from-sitting-to-standing movement without showing this movement at all? How would you help students improve and finesse the movement without showing them how it’s done, and moreover, without correcting them?

This means you would need to describe the movement. And to do so, you would have to take the movement apart, identify its parts and their differences, and to tell the sheep from the goats in terms of movement instructions.

Where and how could you place your hands? What difference does each position make in your experience, through your experience? And isn’t that strange- after a couple of variations something inside of you will chose the hand position best fit for the way you act. Where and how could you place your feet? Is there a midline, where is the pelvis? When would you move your chin closer to your chest? This too could be chosen for you. By whom, by which means? Where do you look at? How would a differentiation of your eye and head movements improve the overall movement? Which hand leaves the floor first? Marcello Palozzo has his his right leg bent more than his left, and his left hand leaves the floor last. Maybe at first his weight is more on his right leg? Is that a contra-lateral gait pattern? How does this compare to a classic Lunge get-up? Is driving one knee forwards, and then the other, easier than driving both knee forwards at the same time? When is the earliest time to extend the right hip? When could be the latest time? How would this shift in timing hip extension change the overall movement and in which ways? What would be the benefits, what the disadvantages? But enough already, I could build an entire workshop around these questions. (And so could you.)

By using language, by putting movements and differentiations into words, you will enter a world that was right in front of your eyes all the time, yet it was hidden from you. In a certain sense—used in a certain way—language is the key to consciousness, the key that opens the door… the bulldozer that paves the road… to becoming more and more conscious.

In the 1999 movie The Matrix, the main character Neo is offered the choice between a red pill and a blue pill by character Morpheus. “You take the blue pill… the story ends, you wake up in your bed and believe whatever you want to believe. You take the red pill… you stay in Wonderland, and I show you how deep the rabbit hole goes.” Says Morpheus. This Red Pill Blue Pill story represents a choice between embarking upon a life-changing journey or to staying with whatever cards one has already been dealt. “But is it a choice, really?” Ask I. When reading Julian Jaynes’s hypothesis on how the human mind has changed over the past 3000 years from being bicameral to conscious then taking Morpheus’s red pill might not be a choice at all, but merely a breaking-in of our new pair of shoes.

Education that lasts a lifetime

“We always learn what we experience—that which is actually happening for us. This is different from learning from our experience.” — Anat Baniel

Marvellous quote from Anat Baniel from her book Kids beyond Limits. The saying «We learn what we experience» doesn’t invalidate «We learn from our experience», it’s something different altogether. It’s another angle at looking at learning. Anat Baniel is writing this in a context of how to create conditions for learning, what we need to create an environment that can help children with learning disabilities and special needs, and what kind of exercises are not rote drills, but exercises that themselves create conditions to learn specific skills and abilities.

The same thing is very well explained in the context of compulsory schooling by John Taylor-Gatto, in his book Dumbing Us Down:

“Children learn what they live. Put kids in a class and they will live out their lives in an invisible cage, isolated from their chance at community; interrupt kids with bells and horns all the time and they will learn that nothing is important; force them to plead for the natural right to the toilet and they will become liars and toadies; ridicule them and they will retreat from human association; shame them and they will find a hundred ways to get even.”

And here’s the same thought, but with a serving of resentment and sarcasm, by Jerry Farber, from his heavily criticised essay The student as Ni****:

“Even more discouraging than this [..] approach to education is the fact that the students take it. They haven’t gone through twelve years of public school for nothing. They’ve learned one thing and perhaps only one thing during those twelve years. They’ve forgotten their algebra. They’re hopelessly vague about chemistry and physics. They’ve grown to fear and resent literature. They write like they’ve been lobotomized. But Jesus, can they follow orders! Freshmen come up to me with an essay and ask if I want it folded and whether the name should be in the upper right hand corner. And I want to cry and kiss them and caress their poor, tortured heads.”

But what is this pearl of wisdom to us now, as grown-ups? «We learn what we live» what does it mean, now? Now- that we have graduated from school, and get to create our own environment? Now- that school might already be in half a lifetime’s distance? To what parts are we still defined by what we’ve learned way back in school?

Improving ability through becoming conscious of our actions

Just recently I came across two examples of how being unconscious (of what is happening) is helping us with learning new skills. Here’s the thing: there’s a type of movement learning that is done without consciousness, without us knowing how the learning is facilitated. The first example is from PewDiePie, a famous Youtuber:

“Okay so I remember I wanted to do this flip and I thought it was impossible. I just couldn’t wrap my head around it, but I kept trying and I kept trying. And you think you’re not making any progress but then eventually you got it. And it’s so weird to me how you can learn something without thinking it. Muscle memory, yes. This is very satisfying. To get that direct seeing, that instant improvement. There’s something very satisfying about that in my opinion and it just feels so good in your hands when you’re doing it. I love it.” – PewDiePie, Youtube, “My 12 Things I Can’t Live Without.”, Timestamp starts at 18:17, youtu.be/7tB4jwvvuhQ?t=1091

The second example that I want to show you is from Julian Jaynes, who was an American researcher in psychology at Yale and Princeton university and best known for his 1976 book, “The Origin of Consciousness in the Breakdown of the Bicameral Mind”:

“Nonsense, absolutely. We know that you can do lots of kinds of learning, like conditioning, it doesn’t happen with consciousness at all. In fact, consciousness wipes it out very often. There’s studies done at Columbia by (inaudible), another kind of learning that we do such as learning skills. And this has been shown to be astonishingly not conscious. If you want to try an experiment at home: now you take two coins and toss them like that and try to catch them. Well then go over what your consciousness is doing—it’s not really guiding your hands and really promoting the whole thing. It’s, oh, we say what a stupid thing, we’re picking up the coins and say, how clumsy I am. No this is something I’m going to have to stop doing… Suddenly it will happen. And you’ll say, How did that happen? Now can do it!” – Julian Jaynes, “Consciousness and the Voices of the Mind” Presented at The State University of New York at Stony Brook, August 1986, Lecture available for members of the Julian Jaynes society, Timestamp starts at 13:51, julianjaynes.org

And then there was Moshé Feldenkrais, a Ukrainian-Israeli engineer and physicist, who tried to do something completely else: he tried to make the process of learning conscious. As an engineer he was like, “What if we break seemingly simple movements further down into parts that we can then observe and refine?” Some of his students saw a mathematical mind in this idea and thus called this process differentiation and integration. And in fact Moshé Feldenkrais was remarkably successful with this idea—and was able to rise to fame for the duration of his lifetime by helping many people who were beyond help in a medical sense. Instead of treating them like they had a medical problem, he viewed their troubles as a learning problem.

However, as technical as it might sound like, this kind of disassembly of elusive, simple, every-day movements did not remove the magic that PewDiePie pointed out when he said, “it’s so weird to me how you can learn something without thinking it.” There’s still this magic moment when something suddenly has been successfully learned. Nevertheless with Moshé Feldenkrais’s method the process of learning became more visible, more accessible, and more workable.

Have a look at the following example of David Zemach-Bersin, whose work is inspired by Moshé Feldenkrais:

“I have a suspicion that if the movement is organized more proximally that Matt will feel a difference in the weight of his head. So I’m going to bring my left hand under his head again. And I’m turning the head and then I’m going to bring his left elbow to be snug. And again the head gets much lighter, can you see? And look how his upper back is participating. I’m going to tilt Matt’s pelvis thinking of his lumbar spine. The fifth lumbar is coming towards the ceiling. Okay, so that was maybe again about four times. And now I feel like I’m working about a tenth of what it took to lift your arm and your head two minutes ago, and look how light your head has become. How do you explain that?” — David Zemach-Bersin, Youtube, “Functional Integration Demo: Oppositional Movement on the Stomach Part II”, youtu.be/rjnMe9lyUkw

By breaking down movements that are as simple as they are elusive, such as coming from sitting to standing, or turning the head, we become able to fiddle with and to improve not only the movements themselves, but the process of learning. Instead of being unconscious of the whole thing, we identify its parts, and thus can work on each part. And of course- by doing so, over the years we learned a whole lot of tricks and strategies to improve how we learn skills, and abilities. Strategies such as, just for example,

  • to move from distal to proximal in movement initiation,
  • to reverse proximal and distal movement,
  • to distribute a movement proportionally over the whole body,
  • to exaggerate one component of a movement in relation to all other components,
  • to reduce work in order to be able to perceive smaller differences,
  • to explore the same movement in a different position, which means a different organisation towards gravity,

and so forth. There’s no definite list of strategies as none of this is institutionalised and also we keep discovering more. Maybe that’s part of our learning as well.

I find consciousness one of the most fascinating aspects of my own work, as you can see with my Youtube videos. Instead of focusing my entire Youtube channel on a single topic, such as how to improve knee health, or hip pain, or back pain, etc, I created many videos for a whole lot of topics. But what they all have in common is this idea that was inspired by Moshé Feldenkrais: to learn consciously, to become more and more conscious—in our feeling, in our sensing, in our thinking, in our movements, and through our actions.

I’m not saying that being conscious is better than being unconscious, or that one is generally better for learning new skills than the other, but I’m saying both strategies—to learn consciously or unconsciously—have their time and place. I’m saying we should know both, and we should know how to use both whenever and however we wish.