Episode 45: AI and the Socratic Method

Below you can view or listen to Episode 45 of The Personal Brain Trainer Podcast.

AI and the Socratic Method

AI and the Socratic Method podcast episode

Listen to Our Podcast:

Watch Video: CLICK IMAGE BELOW

Brought to you by:

Good Sensory Learning (http://www.goodsensorylearning.com/)
Learning Specialist Courses (http://www.learningspecialistcourses....)
Bullet Map Academy (http://www.bulletmapacademy.com/)
Dyslexia at Work: www.dyslexiawork.com

Transcript:

Erica: Welcome to the personal brain trainer podcast. I'm Dr. Erica Warren.

Darius: And I'm Darius Namdaran, and we're your hosts. Join us on an adventure to translate the scientific jargon and brain research into simple metaphors and explanations for everyday life. We explore executive function and learning strategies that help turbocharge the mind.

Erica: Come um, learn to steer around the invisible barriers so that you can achieve your goals. This podcast is ideal for parents, educators, and learners of all ages. This podcast is brought to you by Goodsensorylearning.com, where you can find educational and occupational therapy lessons and remedial materials that bring delight to learning. Finally, you can find Dr. Warren's many courses at, uh, Learningspecialistcourses.com. Come check out our newest course on developing executive functions and study strategies.

Darius: This podcast is sponsored by Dyslexiaproductivitycoaching.com. We give you a simple productivity system for your Apple devices that harnesses the creativity that comes with your Dyslexia.

Erica: Hello, Darius. I am excited about this next topic. Will you please tell us?

Darius: Well, I'd really like to talk about AI and, um, using AI with your executive function skills like Chat GPT and how to use it to get things done, but in a slightly different way than most people normally regard AI in getting things done and focused around something called the Socratic method. One of our mutual friends, Stan Gloss, and I spent some time talking about this. We even wrote a joint blog about this. And it's a different way of approaching artificial intelligence and how it can help us think and, uh, learn.

Erica: That's really cool. So tell us a little bit about the Socratic method.

Darius: Okay. Socrates, very famous philosopher and, um, teacher, used a, uh, technique of teaching his students by asking questions. Instead of being the source of knowledge, what he was trying to get the students to do was to go out and be the explorer of knowledge, to go out there and explore. And so the Socratic method is this way of questions and answers, questions and response. So I don't know if you've ever encountered a teacher that teaches in a Socratic way. I think I've only really seen one teacher really do this very intentionally. What happened was he always answered a question with a question. He never gave you a straight answer. So you would say, what do you think about this scenario? And he goes, yeah, that's quite interesting. What do you think about it, uh, from this perspective? And you go, well, from that perspective, I would think about this, but I'm wondering about that. And you go, yeah, that's quite interesting. What's your take and how do you see that affecting? And so it's just this constant question backwards and forwards, rather than always coming to an authoritative statement of truth. It's kind of like, well, that's the answer, isn't it? And the Socratic method, really, there's a few elements to it. He's very clear. Socrates, first of all, it needs to be all about stating open ended thought provoking questions so that it becomes a conversation with the truth, conversation with reality, conversation with a knowledge subject, where the teacher and the student are both going on a journey side by side and encountering this experience together, okay? Rather than the student and the teacher. And, uh, the student saying, well, tell me what the right answer is. And the teacher goes, well, I've been here before, I'll tell you what the right answer is. So instead of it being a tour guide saying, right, this is this kind of place, and I'm going to tell you all about it. Oh, that's very interesting. That's another place I'm going to tell you all about it. That's kind of one style of teaching. Whereas a Socratic method would be, all right, let's go around this monument. And they go, well, have you noticed this carving over here? And you go, yeah, I do notice that carving over here. Why do you think it's there? And you go, well, do you know what? I saw another carving like that around the corner, and it looks very similar to that. You're like, yes, it is actually very similar to that. And so you might ask some questions. When was that carving made? Oh, it was about the 18 hundreds. Well, that's fascinating. Do we know who carved that? And it looks very similar to that other carving here and this carving over there. And you have this conversation. So that person is being actively encouraged to listen, clarify, reflect and really probe and draw out information through a conversation with a person that knows more than them, but also with the understanding that you also have knowledge that's maybe latent or knowledge of questions or perspectives of inquiry that could bring out different aspects of the knowledge. And so it's all about this inquiry process. And so this is very, very different than our standard educational approach, standard learning approach, and also very different to the way most people approach artificial intelligence.

Erica: It's interesting because what it does remind me is of a few things. It reminds me of a facilitation model where you're facilitating learning instead of giving information, you're bringing it out in others. It reminds me of adult education, which is more of androgogy versus pedagogy androgogy meaning that people come with a wealth of knowledge. And so, again, you're just facilitating. And then it also reminds me of this coaching course that I just took because I loved this course. It was a course through Upbuild, which is, uh, an organization in New York City, but they're actually pretty much international at this point. And I got certified as an Upbuild coach. And that whole training was all about how to ask great questions, how to be curious. And the difference between therapy like a therapist versus a coach is therapy tends to go into the past where you're digging up old things that may trigger you. And although coaching may touch on that, the real momentum within a coaching session is about manifesting where you want to go, who you want to be, how to be your best self, and how to get there. But it's so interesting because it was such a valuable course for me, because it really taught me to listen and to facilitate and to ask just these amazing questions that helped people. Because you can't know a person as well as they know themselves. It takes a lifetime to get to know someone as well as you know yourself. And most of the answers do reside within yourself if you're given the right questions. And when you're able to use that innate wisdom within yourself, it's extraordinarily empowering. Plus, you're far more likely to follow through with it because it was your idea, even if you're a little bit guided. So I love the whole idea of the Socratic method. It's very near and dear to my philosophy of teaching, which I think is far better because it's more interactive. It really gets the individuals actively involved so that they can maintain attention and they can maintain interest and motivation in what they're learning.

Darius: Yeah. And you've got all three aspects of executive function working in this scenario, where you've got this working memory of information coming in in chunks, and you processing that, relating it to your inhibitory control. What am I focusing on? What's my question? What's the direction I'm wanting to go? And then you've got the cognitive flexibility of having to adapt to circumstances and changes. And all three are being stimulated through this Socratic method.

Erica: Yeah, I could see that the metacognition being very stimulated. I really see the working memory big time. It really teaches you how to utilize your working memory. And in a way, what the Socratic method does is it takes your inner thinking and makes it outer thinking.

Darius: Yes. And, um, it ties into this sort of inner voice that we've been talking about before you're having a conversation. You can have a Socratic method with yourself, this inner voice, the questioning, and other people. So the thing I really want to kind of push into is this area of artificial intelligence and learning and thinking and doing. And, uh, there's a tendency for people to open up chat GPT and say, write me a letter to such and such, or Write me this article or do this for me, and it goes, Fine, okay. I'll do. The standard things that I've learned are in a blog post about such and such. And it gives you a, uh, finished result, but it's not adequate, and it's unsatisfying. And actually, it's not the ideal way to interact with an artificial intelligence or this machine learning, because in many ways, it's operating on a Socratic method as well as a default. It's wanting questions back from you, um, it's wanting answers back. It's wanting a dialogue, because once you have that dialogue, you can really narrow down the scope of what you want. In many ways, it's a way of creating inhibitory control. It's like, let's zoom in, zoom in, zoom in, focus, cut until you get what you really want out of it, and it's out of a dialogue. And even the way Chat GPT has been created is like an antidote to the way people have reacted to machine learning in the past. So, machine learning, you give it a job, it'll give you an answer, and you go, that's quite good, but it's not there yet. And they put into this Chat process where you go, I like it, but it's a little bit formal, you're being a little bit too rigid. Make it more friendly, make it more warm, and it'll go, all right, okay. And it'll readjust it. And then you're like, you're being a bit too brief. Expand on this a little bit, et cetera. So even on a very basic level, that's happening. But I, uh, created a prompt in a conversation with Stan Gloss, and we experimented with this because we were like, could we have an actual conversation in a Socratic method where the AI acted like know? Because Socrates, in his conversations with his students, would know more than his students, but wouldn't reveal too much of it, so that he didn't rob his students of that exploratory journey, but revealed enough to help the exploration continue. So there's this sort of fine line to go through, like this coaching method that you're talking about, life coaching, it's very much about drawing out the answers that are within. This is like drawing out the answers that are just close at hand to be discovered around you. So let me just give you a prompt that we did. Okay? It went like this. I want you to act like Socrates, and this is to Chat GPT. We type this in. I want you to act like Socrates and imagine you're a teacher using the Socratic method to engage with students. When answering questions, remember to pose open ended, thought provoking questions to guide the conversation. Number two, actively listen and encourage elaboration and clarification of ideas. And then three, encourage self examination and personal growth. Four, demonstrate intellectual humility and openness to new perspectives. Like from me, the student, engage in a dialectical process, refining ideas through dialogue and adapt your approach to each learner's needs and abilities. And encourage emphasize the importance of the inquiry process over finding the right answer, and encourage participants to ask their own questions and cultivate curiosity. And then I said, you're going to be the teacher named Socrates, and I'm going to be the student. I want you to m ask me a question and wait for my response and then repeat the process. And so what happens is it then goes, Hi, I'm Socrates. I'm here to ask you some questions, and let's go on a learning journey together. And you go, Great. And it'll go, what do you want to explore? And you go, I want to explore the industrial revolution and its impact on art. And you go, oh right, great. Less so what do you know about the Industrial revolution at the moment? And you go, well, I know this about the Industrial revolution and what makes you interested in art. And it just goes on this question and answer, question and answer process.

Erica: So this is all typed in?

Darius: This is all typed in, uh, at the moment, yes.

Erica: I look forward to when Chat GPT can accommodate a yes, yes, that would be great.

Darius: And, um, it's not hard to do that, actually. You can program it to do that right now with the API. You could make sure all of its output is just turned into speech to text. So you hear it instead of read it. And then you can just do speech to text into the dialog box and speak it out. In fact, it's halfway there with, uh, chat GPT iOS app, because you can click on a button and dictate to Chat GPT your answers, and then it will give back your text.

Erica: Cool, I didn't realize that. I've never used it on my phone, I use it on my computer all the it's very, very useful. But yeah, I really enjoy the conversation that I have with Chat GPT to really refine what I'm looking for. And it is a conversation, because whenever I use it to create any kind of written work, it's a collaboration.

Darius: Yes, absolutely. I think that's the key. There's a loads of people out there who are kind of like saying, oh, it will do this for you, and you don't have to do anything. And you're kind of like that's. Uh, just quite a lot of hype and BS, to be quite frank. The reality of using artificial intelligence in real life, in real work, is that you are the one who is exercising discernment and exercising decision-making processes, and Chat GPT is bringing materials into the conversation. So it's kind of like a building process in the way I think about it. So if you take this metaphor, we really like to turn practical, theoretical things into metaphors, real life metaphors on this podcast, let's say you're building a garage, okay? And you say, I want to build a garage. And let's say in this example, like a garage would be, you want to write a report. Analogy, I want to write a report. There's a specific structure to a report, there's a content to the report, et cetera. And you're building it like you would build a wooden garage. You might get Chat GPT come along as an assistant laborer and you go, hey, you've built garages before, haven't you? And yeah, I've built garages before. What has been the standard material list for a garage like this? Oh, well, in these other garages, we've used four x twos and then two x twos, and then we've clad it like this. Uh, and it gives you a list of different materials, as it were, and you go, well, that's interesting. I don't think I want to go down the completely wooden garage route here. I think I would really like to have it much more long term and made it out of metal work and let's go down the metal route. Oh, right. Well, okay. And so there's this kind of instead of this assistant laborer giving you all the answers like they're the architect, they're not. They're more like a, uh, very well informed and experienced laborer and assistant. And I think once you start working in that kind of way, you get really good results from it, as opposed to, oh, go build me a garage. And it builds you a garage completely out of woods. And the door is on the wrong end because you didn't specify what door where the door should have been. You go, Hold on a minute, why did you put the door on the side like that? And you go, well, that's where most people put the doors. No, look at the road here. The big door has to be at the front. There needs to be a little door at the side. Oh, I'm sorry, okay, I'll go away and fix it and sort it. But there's a dialogue.

Erica: Well, I love using the metaphor of, uh, Chat GPT being almost like an orchestra and you're the composer. Mhm, or you could say you're the conductor. Well, you're really the conductor and the composer. Okay. It has access to all of the instruments that you need, or even using something like GarageBand. Right. You've got all the tools right there that you need, but you have to conduct it. And interestingly enough, uh, this be such an interesting class to offer to students, which would be a class on how to properly use AI to assist you in learning.

Darius: Absolutely. And I think the key to it is when we in the future, six months from now, people will require you to make some sort of statement at the bottom of a piece of writing, which is some form of any formal writing. Was this primarily human generated or AI generated? And there will be an acceptance. You can see this in academic requirements. Now, academic institutions are saying, we will allow you to use Chat GPT for your academic paper as long as it's not generating the content, but it's cleaning up the content and structuring your words and your content. And so the emphasis is you come up with a content and it cleans it up. It's kind of like you choose the planks of wood that you're going to put in and put it into place, and it comes along and nails them into place, or sands it down and polishes it off or whatever. But you've decided this is the structure, this is the height of garage, this is where the door is going to go, this is where everything's going. This is designed by the architect. Yeah, absolutely. And composed, like you're saying, composed by you. Now, it could be that you might go to Chat GPT and say, I want you to articulate how you would build or play this bit of music. And then it'll come up with a piece and you go, gosh, that's really good. I'm just going to use that and copy and paste that in. Now, that's a little bit tricky. And I think my experience with using Chat GPT every day, nearly five to ten times a day, with my work. Now, for me, as a Dyslexic individual, with some ADHD, this is like the ultimate Dyslexia, uh, assistive technology for me, because what I do using the building analogy is I would go, I've got to write this report or build this building. Give me a bullet point list of the 15 things that I would need to consider when doing this exercise. And it would come up with foundations, depth of foundations, positioning of doors, height of entry, height of roof, pitched roof, flat roof. And it would come up with all these kind of sensible things to consider. And you're like, all right, yeah, that's quite interesting. And sometimes we can forget some of the sensible things to consider because we fixate on a particular thing. And so having this counterbalance, and sometimes it can suggest too many things, like environmental considerations of insulating the roof. And you think, yeah, no, I'm not going to do that. We're not going to have an insulated garage.

Erica: It's funny, there are times where I feel like Chat GPT almost has attention deficit. It's like, no, um, you're veering off here, I want you to stay focused. I feel like I'm constantly pulling it back in. And it's interesting how it will veer into new areas that it hasn't veered into, or new language. And it's funny that I'm just constantly having to kind of pull it back in, putting those blinkers that we've often talked about, which are those screens that you put around the edges of a horse's eyes so they don't get too distracted. But I find that's something that I'm often doing when I'm toying with Chat GPT.

Darius: Yeah. And I think this kind of Socratic method where you have a dialogue backwards and forwards, not just on, um, exploring a new learning subject, but also exploring how you're going to write a report. I'll give you an example. There was a young person, an artist that I know, who was required to submit a, uh, business plan for a funding application. And he doesn't need a business plan for this art school, for this art. He knows what he's going to go and do, but, uh, for this formality, how to do it. Now, this could take hours and hours to do, but I just went to Chat GPT and I said, look, if you are an artist wanting to create a, uh, small sole trader business in the UK. And you've just come out of art school and you've got a few exhibitions set up. What are the 15 things I would need to make sure are in the business plan? It would make a bullet list of things that I would make sure oh, that's really interesting. Let's go down through each one and say, give me four or five talking points around that first topic. And then I would just talk it out and dictate it. And then another four or five for the next one, talk it out, dictate it, and then I'd say, take everything that I've dictated and summarize it and clean it up into something that's presentable, and it would do it. And then you've got a business plan that's well presented. But it's my thoughts and my words. But along guided with the, uh, structural processes of some intelligence that's very systematic at processing information.

Erica: It's interesting because I often think about this quite a bit, but initially, I thought that Chat GPT was kind of acting as another form, like an exterior executive functioning tool. But I think when you use it properly, it allows you to use your executive functioning skills. Because what I find is what happens with a lot of students is they get so caught up in the details of their academic work that they are not able to do executive functioning. For example, if they're writing a paper, they may get so caught up in the handwriting, they might get caught up in the spelling. They might get caught up in the sentence structure. They may have, uh, dysnomia where they can't think of the right word and because they can't think of the right word, they lose their train of thought. And what I'm also realizing, although it can be used as an executive functioning tool, it can actually be used as a way to release you of all those other burdens so that you can truly learn how to use your own executive functioning skills.

Darius: Bang on. I totally agree. And in effect, what you've been describing is processing versus executive functioning. There's all these cognitive processes that are going on. Decoding words, recoding words, structuring grammar, reorganizing things, expressing yourself. And these are kind of low-level processing systems, which often overload the higher level executive functioning systems. And, um, if we can delegate that to something like these m machine learning processes, that is very much where that level of intelligence lies. It doesn't lie in the metacognitive level. That's where our level of intelligence lies. This is much more a procedural intelligence.

Erica: I just had a really cool thought, and we just recently talked about the twelve ways of processing. Maybe we now have a 13th way of processing, which is AI processing. It's a little different than verbal processing. It's a little different. Or maybe it's really uniting multiple ways of processing. I guess it all depends on how you do it. If you're doing a Socratic method, of processing. That's really more of the logical reflective.

Darius: Yes.

Erica: Right. If you're asking Chat GPT to be someone that organizes your information into like, a very usable sequence, then it would be sequential. Maybe you want it to create a mind map. Then it's going to be simultaneous. But, uh, how beautiful that can be for an individual that does learn how they process best, and then they can teach AI, uh, to really accommodate and support their way of thinking, particularly if a teacher is not doing that in a classroom.

Darius: And in my case, I'm more focused on the workplace and people at work using artificial intelligence to do their work better. And I think that's really lovely because I don't think it's a 13th way of processing. I think artificial intelligence is learning all twelve ways of processing. So it's doing visual, it's doing musical, it's doing movements. You're talking about how, uh, does it do movement? Well, look at Tesla. Tesla's artificial intelligence is cool.

Erica: Okay. Uh, then you're just going, we're moving into that type of artificial intelligence. You're absolutely right. I was like, how is that kinesthetic?

Darius: But you look at mid journey and this visual spatial, draw me a painting of this with this, and doing this in the style of such and such, and it creates it. And then you go, well, that's not quite right. Can you do it like this? And it changes it? And so on. And then you can also not just talk to it. You can start drawing on it and saying, no, I want this. And you're actually having a dialogue through drawing, not talking. So you can draw. You could be mute and still dialoguing, visually processing this thing. So each one of those processing styles are being learnt right now by machine learning. And there's like an app or a machine learning neural engine that is optimized for each so there's a neural engine that is optimized for the visual, like mid journey and stability. You've got the Chat GPT one, which is very much optimized for this dialogue approach.

Erica: Verbal. Yeah, I guess that would be verbal, technically tactile, if you're typing in well, I mean tactile.

Darius: No, I think there'll be a point where, for example, Optimus, which Tesla are creating at the moment. Have you heard of Optimus?

Erica: A little bit, not too much.

Darius: So, Elon Musk is developing a know, a humanoid robot. So basically putting in all the Tesla cameras into a, um, body. And this robot can pick up things and move things around in the factory and ultimately in your home. And that would be tactile because you would do something to the robot, the robot would do something to you. And there would be a physical, uh, interplay where it would learn and respond and process through physical stuff.

Erica: Well, I mean, typing is physical, but you're tactile, but in a limited way.

Darius: But a tactile processor would want to build something together, make something and oh, I'm going to put that there and no, don't do it like this. It's got to be smooth like that and so on.

Erica: So let's go back to the Socratic method to pull this all together, because I think that is just so interesting. Now, what if somebody, perhaps in our audience, wanted to use the Socratic method to help them? It could either be in the workplace or in graduate school or how can they do that?

Darius: So the way they do that is, first of all, by having a Socratic approach to how you interact with AI. So fundamentally, this podcast, my heart in this podcast, was to change people's approach to the AI, that it's not just a dumb vending machine that you go, I want this, spit it out. Oh, that's not quite right. Uh, this is rubbish. This is going to be more hassle than this work. That it's a dialogue. And you go with that approach rather than wanting a quick finished result. End of story. No, it's like working with a coworker and you get to know what they're good at and what they're not so good at, and you get to learn how to prompt the AI that is so crucial. How do you ask questions? How do you frame what it's meant to be doing? So even looking at this prompt of being a Socratic, let's break this down. I want you to act like Socrates. So first of all, I've told it that it's got to act like Socrates. Now it knows who Socrates is. It's read all of Socrates work. It understands what Socrates has said over the last 5000 years, what people have said about Socrates and the um, Socratic method. It knows all of that already, okay? So I am stimulating that by saying I want you to act like Socrates. And it goes oh, right, Socrates. I'll put my Socrates hat on, okay? And I want you to imagine you're a teacher using the Socratic method. So not just Socrates the philosopher, but Socrates the teacher. So I'm training it to say, look, this is the mode I want you to operate in. And so when it comes to a child or a teenager or an adult using Chat GPT, you've got to tell it what role it's going to play. In this, it's like a theater. You've created a theater, you've set the stage. Let's even tie it back to our working memory analogy of this, creating this stage. Your working memory is like a stage or the spotlight on it, and it's got limits. And what you're doing is with Chat GPT or any AI, is you're actually trying to make sure it's taking this huge amount of knowledge and restricting it to a, uh, stage. Act like a teacher, act like Socrates. Not Plato, not as, ah, someone else or whatever, Socrates. And so that's one definition. Now, I want you to engage with me like a student, okay? So I'm. The student, you're the teacher. Ask me questions. Oh, okay. You want questions? Ask me these open-ended questions. Okay, I've got you. If you start saying, I think this, and then I start saying I think that, then be open to adapt and adjust your viewpoints. I could tell it to be completely the opposite. I want you to act like a professor of mathematics who knows exactly what needs to be done. I really want to interrogate and know this area of knowledge. I want you to correct me if I'm wrong, really get me to the right answer. You can tell it to act like that, and it will. And so you need to shift and set the stage that you want to act on and play in.

Erica: It's so interesting. It's like you're almost presenting a paradigm yes. To AI and saying, I want you to be restricted to this paradigm. I want you to be a behaviorist no, I want you to do psychoanalysis, but I, uh, want you to be a cognitive behavioralist. But, uh, it's very interesting because what it gets you to do again, as the conductor, it enables you just to have fun being cognitively flexible and directing it to be very rigid. Or maybe not. Maybe, as you said, you can change that perspective and say, I don't want you to be rigid. I want you to be considering multiple different perspectives. And perhaps even say, I want you to combine certain theories, or perhaps even come up with a new one that integrates multiple theories. Which is really fun, too, because then you can start to come up with completely new ideas because nobody else has necessarily combined those theories. And so, uh, it's really fun. You've got a nice little eight step process here that we've written out, that we've been talking about. And is that something we can add to the show notes?

Darius: Yeah, we'll put it in the show notes, and you can just copy and paste it and put it straight into Chat GPT, and then just go and explore a topic that you know quite well in a Socratic method. And, um, the interesting thing is, if you choose a topic that you know quite well, like I put in a topic like Dyslexia, um, or executive function, and I would put it in and have a conversation with it. And I'd go, yeah, you're being a little bit limited. Dyslexia is not just about reading, is it? And it goes, oh, right. Yes, you're correct. Dyslexia is a processing difference that affects reading primarily, but it also affects other areas. And you're like, yes, it does. And I'd like us to have a more holistic view of Dyslexia, because what Chat GPT has done, if you look at Dyslexia in particular, is completely fixated on reading. And whenever I have to dialogue with it on reading, on Dyslexia, I go, Remember, Dyslexia is not just a difficulty in reading. It's a processing difference. And it goes, oh, yes, and I actually actively make it open its mind up from getting out of the rut that the rest of the world that it's learned from is in. Yeah. So it's reflecting the ruts of thinking that other people and other things have, and it will default to that rut unless you ask it to expand itself to include other concepts, even if they're obvious to you. So, yeah, it's a fascinating approach and it's about setting the stage. And if we tie it back, one of the things about metacognition, I think let's go into that level, because we've talked about working memory, inhibitory control and cognitive flexibility, and you can use Chat GPT to help you with all three. For example, working memory, I can say, hey, look, don't overwhelm me with a whole ten different paragraphs to think about here. I just want two words, bullet points, 15 of them, bang, bang, bang, bang, bang, bang. And then my working memory can really focus in on that and not be overwhelmed with words. So I can direct it to work with my working memory in that capacity. I can also use my inhibitory control by saying, look, we're going to stick to this, this and this, but not all of that stuff. There really focusing on this. So you're using your inhibitory control then? Like you said, with cognitive flexibility, it's like, yes, this is interesting, but you're in a bit of a rut. We need to consider this, this and that as well. And it goes, oh, right, yes, of course. And you're directing it also to be cognitively flexible, too. But there's another level that the research on this has shown. The Chat GPT can pass a test to, say, 80% pass mark, let's say Chat GPT four. But if you ask it to reflect on its answer before it gives its final answer, it will increase its pass mark by another 5%, just by reflecting, because Chat GPT doesn't know what answer it's about to give. So the moment you ask it a question, it makes it, um, up on the spot. It literally is just guessing very accurately what the most probable answer combination of words is to that it gives that probable answer. Okay. And then if you ask it, would you look back over your answer and reflect on it as if you're an expert and see where you might change it and it will reread what it's just said and go, uh, that's quite interesting. With this other knowledge I've got, I would maybe adjust it and add this other point of that other point and it becomes a better answer.

Erica: Yeah.

Darius: And that's metacognition by reflecting back, that is metacognition.

Erica: And it's funny, because when you talked about inhibitory control, I was going to say something like, oh, what about metacognition? And you totally grabbed it. So that was beautiful. But, uh, for this week, let's wrap this one up. I loved this discussion and yes. So if you're curious about the Socratic method, we will definitely put a nice little prompt there for you in the.

Darius: Show notes and, um, get really conversational. Don't ask for immediate, perfect results from the beginning. Start like an architect. Ask for the bare bones. Accept a sort of structure of what you're going to talk about. Then start fleshing it out. Start using your words in it, and then it becomes part of what you've expressed and understood. And then use Chachi PT to clean it all up. And then it's fundamentally your knowledge, your words, your experience. But in collaboration with this very systematic processor, this is a processing machine that really helps you process the lower level functions so you can concentrate on the higher level functions.

Erica: Beautiful. Thanks so much, Darius. This was a great episode.

Darius: My pleasure. I'm completely obsessed by this stuff. I love it.

Erica: I do, too. Until next time.

Darius: Next time. See ya.

Erica: Bye. Thank you for joining our conversation here at the Personal Brain Trainer podcast. This is Dr. Erica Warren and