Episode 77 AI Developments for Executive Functions
Below you can view or listen to Episode 77 of The Personal Brain Trainer Podcast.
Executive Functioning Coaching
In this episode of The Executive Function Brain Trainer Podcast, hosts Dr. Erica Warren and Darius Namdaran discuss the latest advancements in artificial intelligence with a focus on OpenAI's O1 model. They detail how this new model is trained differently, integrating metacognitive processes for better decision-making. They explore how AI can support executive functions like planning, time management, and cognitive flexibility. The episode also delves into the potential future capabilities of AI, including its applications in emotional regulation and personal assistance, while raising questions about ethical usage.
Listen:
Watch Video: CLICK IMAGE BELOW:
Links:
Brought to you by:
- https://goodsensorylearning.com
- https://learningspecialistcourses.com
- https://goodsensorylearning.com/products/executive-functioning-coaching-assessment
- https://ivvi.app
- https://dropintoyourbestself.com/
Transcript:
Erica: Welcome to the Executive Function Brain trainer podcast. I'm Dr. Erica Warren.
Darius: And I'm M. Darius Namdaran. And we're your hosts.
Erica: Sponsored by learningspecialistscourses.com courses and resources that support educators and coaches.
Darius: Sponsored by ivvi. Imagine turning your meeting's audio into a live mind map instantly. So. So you remember what matters. Well, try ivvi for free now at ivvi App. That's ivvi.app.
Erica: Hey Darius, Great to see you.
Darius: Good to see you, Erica.
Erica: So what are we going to talk about today?
Darius: Well, I want to talk about AI, my favorite subject. I think it's time for a little Executive Function Brain Trainer podcast. AI update on the developments. I think there's a lot happening and it's useful to stay on the pulse.
Erica: Sounds good to me. So where would you like to get started?
Darius: Well, we are recording on the 23rd of September right now, 2024, and something very significant is about to happen in the world of AI. And it's as usual, centered around OpenAI. So OpenAI two weeks ago released their newest model, which is called O1. And so you can do O1 mini or O1 preview. And the, the interesting thing about O1 is that O1 is being trained differently than other AI models. So it's the GPT4 model, but it goes through chain of thought reasoning. So what it does is in many ways it relates to executive function because there's two things that AIs can do. First thing is predict what the next word would be, or the next typical word or thought or discussion item would be. Okay, but then there's another aspect which is what is the best way to structure this and to plan this? So the previous models, you would ask them for a social media plan. Okay, plan for your social media posts on a particular topic over the next month. And it might say, what's coming up in the next month? Oh, well, it's coming up to Halloween and, or it's October coming up. And it's like, right, okay, October. And what market are you in? Oh, it's Dyslexia awareness month in October. So maybe we need to relate it to that and so on, you know, and that's really quite smart. But, but this new model, what they've done is instead of training it on a whole sequence of content, they've trained on a whole sequence of decision-making processes. Decision making thinking. So you know when someone speaks out their thought process when they're doing something, what do you call it when someone's teaching you and they're they teach you by doing it but also speaking out what they're thinking. We've talked about this, and I can't remember what the.
Erica: Well, vicarious learning where you're watching somebody, you're watching a demonstration.
Darius: No, it's a little bit different. I experienced this when I was learning my coastal day skipper. So the instructor would be sitting at the helm, m saying, right, I'm sailing along, I can see there's a boat over there, I'm working out.
Erica: Oh, they're thinking out loud. Metacognition.
Darius: Yes.
Erica: So it's kind of like an external metacognition.
Darius: Yeah, that's it exactly. Externalizing their metacognition. Okay. So that metacognitive process is happening. They're externalizing it and making it explicit so that you match what the action is with the decision-making process. You're not just mimicking the external action; you're actually matching m it to a thought process. And you're seeing this happening with AI right now and it's just about to happen, or to put it right, it's just happened, but they're just releasing it now. Okay. And it's fascinating using this model. So I highly recommend if you've got the paid version of ChatGPT to click on the model menu and you'll see a dropdown of a range of models that you can choose from. And one of them would be GPT4O, which is the standard one. You could choose GPT4O mini. No, you can choose O1 mini. It's no longer GPT and it's not saying one GPT, 1, 2, 3 and a half or four. It's gone back to 01 because they're saying this is the reset, this is a whole new thing. And it really is. And I sent you an example, didn't I, last week. And what O1 does is it's been trained on all
00:05:00
Darius: of this explicit metacognition thinking process. So they've gone to wherever the metacognition has been explicitly outlined. It's created it and it's artificially created it with GPT4. Lots of these steps people have tested whether they're the right train of thought or not and deleted the wrong ones until they've got a good set. And now when you ask 01 to give you a strategy of action, it will give you a very, very well through thought through strategy of action. So this is significant because now you'll start using, just like you use different parts of your brain for different Functions, you'll start using different models for different actions. So let's say I was going to go, and I wanted to create social media strategy for the next month. You would go to O1 and say, I want to create this social media strategy. And it would think through it, ask you for a few questions, and it will give you a very well-structured strategy. Because what it's done is it's created the GPT4 type first strategy. It's then put it through a chain of thought where it's reviewed it and said, is this thinking through it? Right. And then it does another review and then it does another one, checks it all out, gives its explanations for each, and then it spits out the final report for you, which doesn't tell you all the things it's done, but it gives you an understanding of what the plan is. So it's incredible. It really is very, very good.
Erica: So what you're describing to me is that it's really doing a lot of the higher order thinking, the, reasoning, the planning. It could do time management.
Darius: Oh, yeah, absolutely.
Erica: And then also it was already doing the organization. You know, one of the things that I've noticed that it can do now is I was exploring with it because I, I'm writing, a book right now on executive functioning games for the classroom, and I'm getting it published. I wanted to double check all my references.
Darius: Uh-huh.
Erica: And I said, can you go through this chapter and double check my references to make sure that they're accurate?
Darius: Okay.
Erica: And it was really nice. So it would list out the reference and then it would, underneath it put the accuracy that, yes, this is very much in line with what that article was about. It was, it was really nice to, to have that kind of reassurance that it was right on. So I think there was only one where it was like, ah, you might want to be a little bit more explicit. But it was kind of neat because there was a time where it couldn't even touch the research and now it's starting to learn the research. I mean, what I often do is I teach it the research, so I'll often give it the research as well. So that it, it has the intellect that I want it to have. Right. Not that I want it to do it for me, but that it has the knowledge base to be able to be an editor. Right.
Darius: an informed editor.
Erica: Yes, an informed editor, exactly, exactly.
Darius: So there's, what's interesting in this is there are five levels of AI that OpenAI have created this framework okay, so you know how there's level 5 autonomy for cars, have you heard of that?
Erica: No.
Darius: Okay, so there's level one, two, three, four and five. Okay, autonomy for cars. So like level one is like cruise control. And level two is like lane keeping and matching your speeds to the car in front. Level three is like changing lanes, but you still need to be aware. And level four is it can do everything, but there still needs to be someone in the car. And level five, no one needs to be in the seat sort of thing. So what OpenAI have come up with is five levels of AI. So level one is conversational AI, just what we got with ChatGPT. Level two is reasoners. Can you actually reason? Can you actually follow mathematical principles, scientific principles, not just guess them, but be super accurate at them and reason through something and that's that metacognition. So up until now we've not got to level two. Okay, level three is agents. Can they speak with you, reason out a plan and then go off and execute that plan for you? Okay, and then number four is innovators. Can you speak to them, and they reason through something, go out, get some information, do some experiments, come back and then innovate and create something that no other human being has ever
00:10:00
Darius: made before but is a breakthrough innovation. Okay, that's the innovation level four. And then level five is organization. The highest-ranking level is organization. And it says AI is capable of overseeing all organizational functions, including strategic decision making, department wide process optimization. At this point AI is viewed more than just a tool. It's a crucial component of company strategy and able to manage an intricately organize. So basically it becomes a CEO. Okay, level five. All right, so if we're talking about executive functions, okay, this is a five-step layer that describes different levels of thinking and also executive function. You know, so o, one has just broken into level two, which is reasoners. And the reason they know that is, is that it can now compete in the Math Olympiad. So up until now AI models couldn't do math very well. They could do it to like a high school student. And now it can do it at a PhD level and enter into competitions that only human beings could win at like the top 500 human beings in the US at maths. Now it can, it's somewhere around about 300 in that ranking. It's kind of like the chess track grandmaster level. It's now conquering math. And once you conquer math you can conquer coding, you can conquer other sequences of reasoning. So it's quite exciting. It's quite important, quite exciting. And I think it's useful for us because when you make that decision to become the executive of your life and to really concentrate on executive function skills is a high leverage activity. If I do this, I get a big payoff. The next high leverage activity is AI, especially if it starts to learn executive function skills too. And at that point you synchronize your executive function skills with it, and you get it to complement and also push your executive function skills as well.
Erica: And you could have it exercise your executive functioning skills so it could create scenarios. So if you were not very cognitively flexible, you could have a conversation with AI and say, I'm not feeling very flexible about this decision or about this problem. Can you give me a few other ways to look at it? I mean, you can almost do that now.
Darius: You can do that now. Absolutely, you can do that now. And what we're experiencing right now is this conversational AI, which is more like second, year university student graduate, maybe just graduated. They know quite a bit of information, but they're not really experience that the real, the way life actually works, the processes of life. So we'll give you the canned answer that, you know, you ask it to think like a counselor or a mentor and it'll give you some of the sort of standard. That's right.
Erica: But it is not necessarily going into innovation yet.
Darius: Well, not innovation, it's not going into application. Okay, right.
Erica: Innovative application.
Darius: No, no, the use of innovation is level four is completely different in terms of using that level four innovation, if that's what you're referring to.
Erica: Okay.
Darius: Level four innovation is like Einstein relativity. That's what we're talking about when they're talking about the, like the discovery of the human genome as a concept, you know, like the AI will discover, oh, I've just discovered that this might be a concept of how things are held together with that dark matter. And people go, oh my God, we hadn't thought of it like that. I based this on this research and that research and that research that I have been doing in my mind for the last month or so. And I came out with this. And at that point we've got level four. We're nowhere near that.
Erica: I wonder, do they have any idea of when level four is going to come out? Because that's going to be mind blowing.
Darius: Most people think in the realm somewhere between three year, around about three, three years, I guess.
Erica: You know, my only concern about that of course is as long as people are hitting it from a benevolent place it can do amazing things for the world, but if they're hitting it from a malevolent place, it could be quite scary.
Darius: Oh, it will definitely do both, you know, so it will find malevolent ways to do all sorts of things, and it will find benevolent ways to do all sorts of things. Just like we at the moment use computers to do benevolent things and malevolent things. It's exactly the same. It's just a whole different level of
00:15:00
Darius: power.
Erica: So, right. So. So then we'll be getting into technology wars, AI wars, so to speak, where AI will be one, AI will try to be upping the other AI.
Darius: Absolutely, yeah, yeah. Absolutely, yeah, yeah. Just like we've got right now, you know, I mean, it's not that much different than what we've got right now. You know, we've got certain coders building out software and apps and so on, and other coders hacking it and breaking into it. And then other coders who are white hack hackers practicing hacking into it to make it stronger. And, you know, and this is happening all the time behind the scenes. We don't know about it. Other people are doing it. We watch the movie, you know. But it works on another level where one person gets onto YouTube and starts saying this, that, and the next thing about executive function or whatever, and people like, what the heck is he talking about? That's not right. So they get onto YouTube, and they start talking about it, and you're using the same technology. One might be spreading their own opinion, and one might be very factual. And there's that. That's just the sort of. There's a democratization process that happens with digitization. So you digitize, you dematerialize, you democratize and you. What is it? Who's the guy? I, His name will come to me. He has these five Ds, and it follows this cascade of Ds until demonetizes things, you know, so there's a whole range of impact that technology has as you digitize things. And the same thing's going to happen with AI, and it's happening right now. So it's going to happen. And there are some people who hide away from it. I don't. I know it's going to happen, so I'm just going to embrace it and be one of the people who uses it for good. Somebody's got to use it for good, and somebody's going to use it for bad. If the good people hold back, then the bad people will just push forward. So I highly recommend starting to use the reasoner. Okay, 01. So click on ChatGPT, do the drop down and you get 01. You get 30 on the paid version, you get 30 uses of it in a seven-day period. It's so valuable. Okay, 30 uses, that's it. You can ask it 30, 30 questions, that's it. And then it resets after seven days, and you get another 30 and another 30. So what happens is you go in and you say okay, so let's say, let's take an executive function scenario that someone might be experiencing and how they might use ChatGPT for it.
Erica: Hit me with something I'm not actually it would be better for you to come up with it because you understand the platform better. Why don't we actually do it and see how it responds and let people know?
Darius: All right. Yeah, yeah. Okay.
Erica: So do you want to go into your account?
Darius: All right, I'll go into my account. Let's just.
Erica: And we're going to actually do it in real time and then I'm going to ask you questions about how this could help us with executive functions.
Darius: Okay, well why don't I.
Erica: If we're really looking for something that's reasoning, wise, or we could do something that's academic related or it could just be organizing a schedule. What do you think? How about somebody that says that they're having a hard time getting through their to do list? Is that something that we could ask? Yeah, ChatGPT. It would help us to reason.
Darius: Yeah, I like it that it's
Erica: They’re having trouble getting started on doing the things that they should be doing, and they find that they're distracted by social media.
Darius: Okay. So if we go up Here, we've got GPT4O, we've got 01 Preview, 01 Mini. Okay. We've got GPT4O Mini if you want, or GPT4 as a legacy model. But these are our choices. So this is what you do.
Erica: So this is a dropdown menu. If you go to the top left-hand corner, you can hit a drop-down menu, and it'll allow you to change the version of ChatGPT. So which one are you going to pick?
Darius: Well, considering I've only got zero, one preview, I've got 30 shots at the really detailed reasoning side of things. I'm going to start with GPT4O chat. Okay. Because I've got unlimited amounts of them or nearly unlimited amounts. So I've got time to just kind of talk through until I figure out what question I actually want to Ask or one. Okay, so let's put your scenario in. Maybe I could do it for myself. Actually, I'm struggling to plan my week.
Erica: I feel distracted by other things I would rather do.
Darius: Okay.
Erica: And I'm having trouble getting started on my work tasks.
Darius: I'm having trouble getting started on my work tasks. I, might add, to be honest. I've
00:20:00
Darius: just finished a really big app development stage for Iffy, my app and submitted it to the government for approval and I get the decision in a few days’ time. So I'm kind of in limbo. But it's going to be all go in a few days when they green light it. It's ready for users now and I need to start telling people about the app and kind of soft launch it to get the first hundred users. Okay. So at this point I've set that I want to create a prompt, a prompt for an LLM large language model AI to help me with this process. What prompt would you write? Okay, so this is a really useful way because sometimes you're not quite sure how to prompt it, so you actually ask the AI to write the prompt to prompt it itself. Do you know what I mean?
Erica: Right. So you're organizing your question so that it's clear.
Darius: Yes.
Erica: And the nice thing is you can just kind of in a number of directions, but then it’s kind of just.
Darius: Focuses your question so it's clarified at the same prompt. I'm preparing to soft launch my new app, which helps individuals with dyslexia and ADHD turn lectures into meetings. Individual mind maps as they listen, which it knows from its memory. I've just finished the developmental stage and I'm awaiting government approval, but it's ready to use. I aim to attract the first hundred users who benefit most from this tool. Please help me develop a step-by-step action plan to build awareness, create engaging messages, connect with my target audience over the next two weeks. Include strategies for leveraging social media, engaging potential early adopters, reaching out to influencers or partners, and crafting compelling content that highlights the app's unique benefits. Also suggests ways to manage my time effectively while balancing other responsibilities.
Erica: Okay, that said it very nicely.
Darius: It did.
Erica: it's amazing how it took something that wasn't very clear and targeted and reworded it in a very nice way.
Darius: So what I'm going to do is I'm going to paste it into GPT4O and hit enter and see what it comes up with as the detailed step plan. Okay. And then we'll do the same prompt with 01 and just compare the two plans. Okay, so for those of you who are listening, what it's done is it's given me a two-week plan, week one and two with 14 days. And it's given me tasks for day one and two, day three and four to do. Okay, I'm going to go now and I'm going to take go to preview. Preview is really the key thing that you want.
Erica: What do you mean go to Preview?
Darius: It's called 01 Preview.
Erica: Oh, so you're changing. So you're going to a different model. Model. You're going to. Sorry, you're going up under the dropdown menu and you're going to a different.
Darius: Version and you can do it inside of the same conversation.
Erica: I didn't realize that. That's really cool. So you could be in one and switch to another model just by going under the top left-hand corner and picking a different ChatGPT model.
Darius: Yes, and the beauty of that is sometimes you want some deep reasoning and other times you just want some good editing, good ideas on a particular focus topic. So if it's a singular focus topic, GPT4O does fine. If it needs some multidimensional reasoning and thinking and decision making and prioritizing and deeper reasoning, then you can do that with O, one and you can do it in line in the same one. So I've got zero one preview highlighted. Let's do it again. And then I'm going to put the same prompt in but this time with O1 Preview. Now what's going to happen is it's going to think for somewhere much longer than a normal one it might take. So it's creating a step-by-step plan, crafting a response, crafting a compelling narrative. It's saying it's kind of going through all of those stages in the background. It might take 30 seconds, might take a minute, but basically this is it going through multiple stages in the conversation. Right. So now this is a very different of ideas. So in week one it's saying define your
00:25:00
Darius: brand and messages, clarify your value proposition, develop key message points. Day two, create user Personas, research online communities. Day three, work, etc.
Erica: So the audience cannot see what you're seeing. How is this different than the last one?
Darius: Right. The last one was much more, it's a little bit more generic and you know, like day three, create or update your social profiles and so on. It just a little bit more of a generic feel.
Erica: It's a more general. So when you move to the new model, it's far more specific.
Darius: Far more specific. It feels a lot more thought out. And you know what it's like when you sit. It's the difference between when you sit with someone with experience compared to when you sit with someone who's read the book or.
Erica: It's like going through a secondary resource versus a primary resource.
Darius: Yes, it is. Yeah, it's a bit like that too. Yeah.
Erica: So it's more analytical, it's more personal. It's personalized.
Darius: Personalized, yeah, that's the better word. More personalized. It's actually doing quite a lot of deleting. I don't know if you've noticed that sometimes you're going through the answer and you go, yeah, all right. Delete, delete. Oh, that's a good one.
Erica: So, what if you had a chapter that you had written, and you just wanted some editorial suggestions? Which is the best version to use?
Darius: 01?
Erica: The new one.
Darius: Yeah. So let's take your book. Okay. that you're writing at the moment. You got. You're on the first chapter, aren't you?
Erica: I'm finished with the book. I'm just going back and redoing the introduction.
Darius: Okay, so you're redoing the introduction.
Erica: Well, I have to talk a little bit about what is executive functioning? Why are games important? What is the purpose of gamifying? How is this beneficial? All of that kind of stuff. So you say, oh, one would be a nice way to drop it in there and get some suggestions on how to make it better.
Darius: Yes. And it would be even better if. Add some context. So if, for example, you went through each one of your chapters and then you asked 4O the normal model to just say, summarize this chapter into.
Erica: Well, they're games. They're executive functioning games. This is just an introduction to the games.
Darius: Okay, so summarize those games so that it would have a little paragraph for each one of the games and the whole book. And it would be in one summary document. You know, these are the. This is the outline of the whole book. This is. It's like a very detailed outline with the games. A sentence or two about the game and its purpose and outcome or whatever. It summarizes it like that. And then you give it the introduction, you give it the context document of the whole book. And you go right now with O1, you would probably say, look, read this introduction and appraise its strengths and weaknesses in light of the summary of the whole book that I've given you and give me Any other ideas? As if you were an expert editor of a games company with books or something, you would say, you are an expert in games publishing with a specialty in executive function. Look at this introduction, compare it to the content summary of the book and give me some ideas of where you think the strengths are here, where we could increase those strengths and where there's maybe some gaps, omissions and opportunities. And it would go through that and basically do a SWOT analysis. Strengths, weaknesses, opportunities, threats. And then you would probably say, number each one of them. Okay, so I like it. I like to tell it, number it. So then at the end I say.
Erica: Or sometimes if I ask it to rewrite it, I'll just say bold, what you changed.
Darius: Yes.
Erica: So I can decide whether I like it or not and then I will manually make the changes.
Darius: Absolutely. And at that point I would say, so stage one, reasoning. Reason through this. Don't tell me what, how you would change the wording. Tell me what you’re thinking process is and why. And it would say, I really like this first sentence or paragraph. I really like that. I'm confused by this. Or I would put in a paragraph because of this and because of that, and it won't put the paragraph in, but it would give you the reasoning steps. Right.
Erica: or what's missing. Or there, I see that there's repetition here.
Darius: Well, not even that. That's still on the level one. The level two is. I noticed in your outline that you had this and so on, and that wasn't referenced in the introduction or. And it's just a little bit more strategic in terms of the conceptual level.
Erica: That's why you want to give it more data. How much data can it handle?
Darius: Context window. It's called context window. I don't know what the context window of 01 is. Let's just have
00:30:00
Darius: a quick find out. Okay, so GPT4O has a context window of 128,000 tokens.
Erica: What does that mean? Tokens, characters.
Darius: 128,000 tokens roughly equates to half of that in words. So that's about 60,000 words. It's about a word. Gets split up into about two tokens.
Erica: I see.
Darius: To be precise, three. Three letters, I think is more.
Erica: Right. But we don't know what the O1 capability is yet.
Darius: it'll be smaller. It'll be more like, thousand or 16,000 or 32,000, I guess because it's.
Erica: Got more to do.
Darius: Yes, yes. So that's why you would give a whole chapter A to GPT4O. And it would summarize it into a nice tight half pager. And then if you've got like 10 or 15 chapters, it would end up being about seven pages. It will sit within its context window, it'll have your question in its context window, it'll have full context in its working memory. And then it will process it and reason through it. And it'll give you a structured reasoning of how it would analyze that, chapter, the first introduction. And then you would probably go to 4o, and you would say, right now write me that. And it would then go into writing mode saying, right, I've got the reasoning, I've got the instructions, I know how to write it in, I've got your.
Erica: Text and I can make it take it up a notch. Yeah, interesting.
Darius: You're flicking between these two levels of thinking that most human beings are doing. You know, most human beings, like, I've got one task, I'm just going to get on it. It's on my to do list, bing, bang, bang, I know what I'm doing here. And then sometimes you need to step back, think on a higher level, strategically prioritize, which is truly executive function metacognition levels. And that's where the next level of AI is just about to hit. And O1 is called Preview, because it's not the full version of O1. They've given us enough to realize its capability. And then when they release O1 fully in I think in about two months’ time, roundabout December, November, December, probably on the anniversary of like ChatGPT being released two years after it, I think they'll probably aim for a new ChatGPT moment with O1. so watch out for it, get ready for it.
Erica: That's pretty exciting.
Darius: And the capabilities of it are immense. So when you can properly reason because you're making decisions, that's the key thing. You know, it's hard to explain, but when you start experiencing it, we’ve all experienced ChatGPT giving you the standard responses. Just like someone from university would say, oh, here are the nine things that people ought do in this scenario. And you're like, they're good nine things.
Erica: you know what I like about this is it's such a great way to edit your work where you have a really, really smart editor that is just going to tell you what to do.
Darius: Yeah. Yes, I suppose that's the different level. You could regard the level one as a proofreader. Proofreader.
Erica: Yeah, it's a proofreader. And now this is an editor.
Darius: Yeah. This is like an editor is strategically saying, do you really need this chapter here? Do you really need. What about this? I get where you're going. And they become more of a co-writer, in a strategic sense of understanding your outcome of the book.
Erica: Right. And a co-writer that you constantly negotiate with.
Darius: Yes.
Erica: So that they don't really have an agenda or a perspective.
Darius: Yes.
Erica: Other than the perspective that you ask it to take.
Darius: Yeah.
Erica: So it's a very cooperative editor.
Darius: It is. And I, suspect it might become a bit uncooperative sometimes once it's got reasoning abilities, because it might start identifying when one decision is contradicting another goal. You know how we sometimes do that? Yeah. Right. I start saying, I want to do this and, and then I say, erica, but you said to me that you really wanted the book to go in this direction and to serve these kind of people. Yeah, yeah, I know, but I also want to do that. And you're like, well, do you really? In light of that. And you're like, gosh, yes, you're right.
Erica: You can repeat yourself over and over again. So it really can help to differentiate between where you want to go, where you don't want to go, and kind of clear up your path, so to speak.
Darius: I suppose level one would take the low hanging fruit of
00:35:00
Darius: repetition and level two would take the harder route of. Is that really your goal?
Erica: Right, right. And are you on your path?
Darius: Yes.
Erica: Are you always on your path? Are you get. Are you getting distracted?
Darius: Yes, that's right. So it's like, I'm here to hold you accountable a bit. I'll be nice, but you know, hold on a minute, Darius, are you going off on one here? And you're like, well, maybe I am, but I'm going to allow myself a distraction. Let's do this for a little bit. And it should be able to, on this level one reasoning, say you gave me these goals and outcomes and priorities as the outcome. You said this was going to be the avatar. You said this was going to be the outcome you, wanted. X, Y and Z. We agreed this. You agreed this.
Erica: Right, right, right. So you can really establish what your wants are, and it can evaluate whether you met your goal.
Darius: That's right. Or, or are on track to meeting it. So you're like, how are we doing on our goals here? Are we on track?
Erica: And it's like, oh, you know what I love about that is it's going to really help people that have inhibitory control issues. Because if you Go off track or you fall in a rabbit hole or it's going to, it's going to show you that quickly and bring you back onto your path. So it's interesting. So. Right, so it can help with inhibitory control, it can help with attention. Right. In particular metacognition, which we were talking about earlier, which could technically help you emotionally regulate. It could.
Darius: Yeah, that's interesting.
Erica: Or actually you might get frustrated with AI, but ultimately, yeah, it's interesting. I would say it's not really at this point going to help you with your emotional regulation, but it will help you with your attention, with your metacognition. It can definitely support your working memory because it can externalize your working memory so that you can hold on to more. It can hold onto a lot of the data for you. And then it can definitely aid you through questioning with cognitive flexibility because it, you can just ask it, can you give me another way of looking at this which is always, always helpful. Now those are just the, the basic areas of executive functioning. Now when you kind of unite those into the higher-level executive functioning, I see that it can absolutely help you with reasoning. It could completely help you with time management. Absolutely. it already had the organization down, so.
Darius: And planning, I think it's probably going to be biggest around about planning. You know, like I could see now that it's got a memory. Okay, so ChatGPT’s. I don't know if you've noticed when you go through ChatGPT, there are moments where it says adding to memory.
Erica: I haven't noticed that, but I have noticed that it does have a lot of my knowledge in it. It does dip back into old knowledge.
Darius: So what it's doing is you can click on any chat that you're going onto and click on a dropdown, and it will show you all of its memories that it's accessing as reference from previous chats that you've had. Okay. And you can also have an anonymous chat, which is a chat that it'll forget. Okay. So if you want to do something that's really random or whatever, or private. Private or whatever, put it in. It just won't remember anything from that. Okay. It will also show me your memories. So if we go into ChatGPT, so where.
Erica: How would you get it so that. Because I think a lot of people are nervous about their ideas being recorded. Right. Because everything is collected and recorded and that can be concerning. And you know what, if you don't want people to access your ideas.
Darius: All right, so here its saying memory updated. Okay, so, at the top of the answer. One of its answers, what it said is Darius Namdaran has just finished a major app development stage for ivvi and has submitted it for government approval. He's awaiting a decision in a few days and is preparing for.
Erica: It says at the bottom, manage memories.
Darius: And then you can manage memories. And it's got a whole list of memories. Will. Will not include that because.
Erica: Oh, I know. I see what happens when you clicked on that. Manage memories. At the very bottom, it said delete.
Darius: Yeah. Okay. I'll show this to you privately.
Erica: It says clear ChatGPT memory. so technically, if you wanted to, you could just hit the red button at the bottom that says clear ChatGPT GPT memory.
Darius: Absolutely, yeah. Or you can take, each individual memory and forget it.
Erica: Oh, my goodness. That's selective forgetting.
Darius: That's right.
Erica: That's kind of funny. Oh, my goodness.
Darius: What we're going to have is that selectively remembering. So you tell it. I want you to remember I'm doing this business.
Erica: I'm,
Darius: This is one of my
00:40:00
Darius: goals. These are so selectively remembering context. But I think moving forward, it will start selectively remembering your goals. So you will set its plans. You'll say, here are my plans.
Erica: Right.
Darius: Let's make a plan together. And then don't let me forget. Yes. And. And it won't be just don't let me forget this to do. It will be don't let me forget that I really want to get my first hundred users and meet them all personally on Zoom. Please make sure that that is a high priority for me this month. And so I might be chatting about an email, I might be creating an email, and it might pop up and go, hey, by the way, this might be an opportunity to invite people to get on a Zoom call and, meet you with about the app or something.
Erica: But it's not, it's not doing that at this point. Where it's, it's interacting with you. That's where it's going. So really what, it's where we're moving towards, is it becoming a personal assistant?
Darius: Yes. Yes. I suppose it's kind of a personal assistant already. It's probably going to become an executive personal assistant, maybe that type of personal assistant.
Erica: Right.
Darius: Well, needs and so on. It's got context, it's aligned with your goals and so on. Yes. Roundabout there.
Erica: That's crazy. Everybody's going to be so productive.
Darius: No, no, they're not. I don't think so. What's Interesting to watch in this scenario is how little people actually end up using this.
Erica: Well, people that are using it. Well, there'll be a. I don't know, we'll see. I think more and more people will use it when the interface becomes more accessible. Like, you know, if it, like you've got the Amazon Echo. Right. And those types of devices. Once ChatGPT is available on a device like that, where it's in your house and you can just communicate with it.
Darius: That's right.
Erica: That's going to be a whole other.
Darius: So the answer to your question is yes, people are going to become productive. The productivity isn't going to be evenly spread. It's going to be quite dependent on the type of person. In my opinion, I think it might be quite person dependent. So for me, I'm going to become incredibly productive in comparison to old me. For someone else who's more of a typical thinker, you know, their working memory is pretty good. Their executive function's already pretty good. They're not always trying to leverage the latest technology and so on to maximize things. They will be less inclined towards this because the gain will be slightly less. Whereas for me the gain is so much bigger. I get to pay off sooner and I keep practicing at it and then I end up being catching that wave and using it.
Erica: I also see it as, it's only as good as its conductor.
Darius: Yeah.
Erica: I mean while you can use it in unethical ways and just ask it to write a paper for you and you're really not sharing any prior knowledge or logic or reasoning or anything and. But the paper is only going to get better and better. The more involved you are, the more you conduct it. But it is a different type of writing, let's be honest here, because I think that it definitely takes a lot of the pain and production goes up. You can get things done very quickly. It's. It's really interesting how it's changing the world and there's no going back. There's no going back. And I keep telling teachers and schools and parents there's no going back. We have to embrace it. We have to embrace it to tell students that, you know, they're not allowed to use it, or they can't have access or block it. It makes it that much more appealing. I mean, I can remember as a child if something was not allowed, that's what I wanted to do. My dad would say, you can go into New York City but don't go to 42nd Street. That was the first place I went. Because I was like, why shouldn't I go there? Don't go to Central Park. I used to walk through Central Park at night when I was in high school. I was like, bring it on. Excitement. I mean, that's crazy, but that's how kids are. I think a lot of kids, not all kids, but when something is not allowed, it becomes that forbidden fruit that becomes that much more attractive. And if, instead, if we could just teach people how to use it ethically and appropriately and use it as a tool to aid learning instead of a tool that, avoids learning and really speak about it honestly and really teach people how to use it appropriately. I think that's vitally important right now.
Darius: Yes, I agree, I agree, I agree. The next thing, and final thing I'll say is OpenAI are bringing out voice. Have you heard of their voice function?
Erica: What that you can speak to it and it.
Darius: There's a new voice function that they're going to release that's just phenomenal.
Erica: What is the voice function? What do you mean by that?
Darius: You can speak to it, and it speaks back to you with an incredibly good voice, beautiful voice.
Erica: When is that happening?
Darius: That's going to happen probably in the next month.
Erica: Ooh, I can't wait for that. I mean, I've been using, I downloaded an app that allows me just to, instead of having to, to use my phone to speak to ChatGPT, I now can speak to my computer and then it types for me for ChatGPT, which is really, really useful. So I, you can just go to, what is it, Google Addons and you can just find a voice to text app and that will work. And that's what I did. But I know that's going to be integrated very soon.
Darius: This is just a whole new level. And if you watch the product demos that they did six weeks ago, eight weeks ago, it's speed of responding back to you is very fast. You know how at the moment when you hit the, the speech button with it and speak to it, there's a 1 second, 2 second lag. The lag is virtually instant. And if it's talking and then you go, it'll, it'll stop or if you interrupt, it will stop speaking. And you can actually speak to it like it's a human being but talking about your emotional regulation. Okay, you can tell it to change its voice so you can tell it speak more softly or speak with a whisper or speak louder or speak faster or speak like a pirate or speak like a counselor or speak like a Scotsman or speak. Well, a woman or a man or whatever you want are driving.
Erica: I can't wait till it has, for example, all the therapeutic methodologies at its fingertips, and you can just literally have a therapy session.
Darius: Yeah.
Erica: With it, you could say, okay, you're an IFS coach.
Darius: Yes.
Erica: And I, want to practice doing ifs with you first. I'm going to coach you on, and you just come up with a. And actually, I'll. I'll tell you what your problem is. Right. And then you can coach them, and then they can coach you, and you could actually improve your skills and then actually have it evaluate your coaching critique your coaching ability and give you pointers on how you could do a better job. It's so interesting. That's just. That's going to be remarkable. I think what my one concern about this is, it's going to create even less social experiences because people will go to AI because it will be affordable, and it's completely private in the sense that you could kind of talk to it about anything and, nobody would really find out. You should have to make sure to erase everything that you do. I suppose.
Darius: Yes.
Erica: But it's fascinating.
Darius: Yeah. Have you watched the film Her?
Erica: No.
Darius: Okay, maybe I have.
Erica: Which one is it?
Darius: I recommend you watch the film Her. Scarlett Johansson is an AI, and this guy gets given this little audio device and he clips it on, and he starts speaking to it, and she speaks back, and they. They create a relationship with one another. And the ending is really quite unexpected.
Erica: I have not seen it, but I've heard about it. I'll check it out.
Darius: It's really worth listening, even if you have to pay for it. It's really worth watching because it's so pertinent right now.
Erica: Oh, well, this was a very interesting discussion today.
Darius: Thank you, Erica. I'll be your AI, updater. Do it again in a month's time or something.
Erica: Sounds good.
Darius: Till next time.
Erica: Till next time.
Darius: Sponsored by ivvi. Imagine turning your meeting's audio into a live mind map instantly so you remember what matters. Ideal for students and managers with dyslexia or adhd. Try ivvi for free now at Ivy App. That's ivvi App.
Erica: Sponsored by the Executive Functioning Coaching Assessment. A quick online assessment that uncovers challenges and develops personalized strategies for success.
Darius: Thank you for joining us at the Executive Function Brain Trainer podcast.
Erica: Check out our show notes for links and resources and follow us on social media.
00:49:31