Loading summary
Transcript65 lines
- [00:00]
A
I could use your help. Every year we put together our AI Marketing industry report and thousands of marketers use it to help determine where they should focus on their AI tasks for the next year. And I need your help making this year's the best one we've ever done. We've created a very short survey that will literally only take you a couple of minutes to take. So if you could pause this podcast and head over to socialmediaexaminer.com/AI survey, we would love it. It would just take a couple of minutes. Again, social mediaexaminer.com AI survey and by the way, it's going to inform the kinds of topics we talk about on this podcast. Welcome to the AI Explored podcast, helping
- [00:46]
B
you put AI to work.
- [00:49]
A
And now here's your host, Michael Stelzner. Hello, hello, hello. Thank you so much for joining me for the AI Explored PODC podcast brought to you by Social Media Examiner. I'm your host, Michael Stelzer, and this is the podcast for marketers, creators and business owners who want to know how to put AI to work. Today we'll explore how to train AI to think more like you. My special guest is an AI strategist who helps marketers and entrepreneurs systemize their knowledge to scale their businesses. He's the founder of Cognitive Fingerprint, an AI speak. He's also an AI speaker and trainer. Max Bernstein, welcome to the show. How you doing today?
- [01:27]
B
I'm doing so good, Michael. Thanks for having me. Super excited about this.
- [01:30]
A
I'm really excited to have you on the show today. So we're going to get into obviously how to train AI to think a lot like you. And before we get into that, I'd love to hear a little bit of your journey. Where did you start with AI and bring us up to the present as you share that story.
- [01:44]
B
Sure. So I spent the first 16 years of my career in marketing on the corporate side, so working on some big brands inside of pet healthcare, in cpg, consumer packaged goods on the baby side side, and some baby brands. And then during COVID I started branching out and doing some consulting for some smaller businesses, doing some marketing online. And I got absolutely hooked to helping smaller businesses, helping medium sized businesses, helping one on one with consultants, coaches, et cetera. And then around part of that journey. It was a little bit after Covid, I went ahead and I got certified in a messaging framework and I had spent $10,000 to get certified in this messaging framework. So at the time that was, I mean, it's always a lot of money, but at the time, that was a lot of money and I was super excited about it. I came home, was ready to hit the ground running, and it was right around that time that ChatGPT 3, 5 or 4, 0 had just come out. And right as I got home, someone sent me a prompt that pretty much recreated everything I just paid $10,000 to do and to learn and get certified in. And so that was the immediate, oh my goodness, I better figure this stuff out, especially in marketing, or I'm in big trouble.
- [02:56]
A
Was that right when ChatGPT kind of rolled? Was that when it first came out, like right around November 22nd or whatever?
- [03:02]
B
Exactly around that time, yep.
- [03:03]
A
So what happened next?
- [03:04]
B
I started making the transition. I went in head first thinking, hey, you know, I'm going to use this to supplement my marketing journey and with my clients. And what I realized is I was just so passionate about it. All of my conversations when I was talking to people, started leading with AI and I just went in head first. And so for the past two years, I am literally in this stuff, 10, 12, 14 hours a day, seven days a week. I mean, I am so lucky. I literally feel like I have one of the best jobs in the entire world. I wake up excited every single day to go to work and just explore new tools. I mean, it's like Christmas morning for me. Just, you know, people talk about overwhelmed, but I get overwhelmed. But I also get so excited when there's something new released. I get to dive in, explore, play, see how we can apply it. So it's a so much fun for me.
- [03:47]
A
Tell us a little bit about. So, okay, we know that you saw ChatGPT and you saw these models and you're like, holy cow, this can actually do exactly what I've been trained to do. How did that lead to you ultimately doing what you're doing now? Like, I know you said you used to work in the consumer package goods space. I'm assuming you started some sort of a consultancy or something along those lines and kind of tell us a little bit of that journey and lead us up to what you're doing now.
- [04:12]
B
So part of that journey was doing consulting. So I started doing AI consulting, AI implementation. And I was having a lot of these calls and they were going really well, but I couldn't figure out exactly why they were going so well. And so what I ended up doing is taking all the transcripts from the calls and really digging in deep and exploring, you know, what was so special about these calls, what was so different about these calls that made it stand out or that they were going so well when I actually didn't have much of an outline or much of a structure for them. I was just kind of going in and kind of winging it at the time because it was so new. And the reoccurring theme kept coming up and it was all around how I was thinking about AI and my mental models around AI. And so I was like, oh, that's super interesting. And so I just kept going deeper and deeper and seeing how far I could push it and what all was available inside of a transcript and how much it could uncover about how I made decisions, how I thought about certain things, what blind spots that I have on and on and on. And as I started to develop that out and as I saw how useful it was for me, I was like, this has to be useful to other people. This is so useful to me. And I don't see anybody else talking about it, at least at this level, especially back then, you know, about close to a year and a half, almost two years ago. And so I started some small cohorts where I started bringing people on, teaching people how I was doing this and really just building this as a full methodology. So cohorts doing lots and lots of one on one, lots of transcript extraction and seeing what resonated with people. And the more I did it, the more I saw people just light up when I would work with them, go into their transcripts, pull out their thinking and repeat back a lot of the stuff that I was seeing. Just watching their faces light up and just that, that moment of almost validation, of just seeing their expertise come to life was just so incredible.
- [05:52]
A
So now you've got this thing called cognitive fingerprint, which effectively my understanding of it is it takes a lot of insights that's maybe available through transcripts or podcast interviews, and it effectively helps people understand what makes them uniquely them. Is that, is that the essence of it?
- [06:09]
B
Yeah, that's exactly it. Transcripts are one of the richest data sources out there as it comes to, you know, extracting information about you and the person that you're talking with. So, you know, as I started thinking about how could I build this out, I really went deep into the different data layers that were actually involved in this and breaking down. You know, if I was going to go look inside of a transcript, what are, what is all the data that is available in here, which led me down the path of a lot of the cognitive science. And how do you start uncovering this stuff or seeing what is actually there to extract and led me down, of course, another massive research rabbit hole and seeing a lot of the science behind this and a lot of the science and what people were talking about has actually been around for 40, 50, 60 years. We've just never had the technology at this level to go in and extract it ourselves. It used to take huge experiments or huge data and research case studies being done, and now we can just do it from our homes. So it's really, really cool.
- [07:10]
A
One of the biggest challenges today faced by a lot of people that are using the AI is that it doesn't really sound like us, right? It just sounds like AI, you know, and it's got this obvious fingerprint, if you will, you know, as people are thinking about trying to TR AI to properly represent them. What is one of the biggest misconceptions that you see out there that maybe a lot of people might not see? You know, what do you see people believing is the path to this that really isn't maybe the path in your professional opinion?
- [07:40]
B
So one of the, the biggest pieces of advice I always see people give is have AI interview you to learn more about you. And that that works well. It is very still surface level, right. And the reason that stays very surface level is because it is capturing only what you are able to articulate. Right. And so there is so much more that happens, and I think we'll get into a little more about this of a lot more that happens when you are in your zone of genius, when you are in problem solving mode, when you don't know the camera is on, or when you're not taking a, you know, an interview or a test, so much happens below the surface in your, you know, it's called unconscious competence. So when you are able to go in and train the AI on how you actually showed up in real situations, again, when the lights aren't on, when you don't know you're taking that test and pull that out, it is going to able to get such a deeper, richer, more accurate picture of again, how you work and how you think. So the interview style, really good to get started. If you don't have anything, absolutely do that. It's going to get you much further than most people out there. But if you really want to get it to know how you think, how you show up, how you make decisions, this is definitely the way to do it.
- [08:54]
A
Love it. So we're going to get into how to do it in just a minute. But if people pay close attention to what we're about to talk about today and they can get AI to actually think like they think and talk like they talk metaphorically. What are some of the benefits that come with that, that maybe we don't realize?
- [09:11]
B
One of the things, and one of the reasons that I am so passionate about this right now is because so much of the narrative is all around, you know, it's all around the fact that AI is going to commoditize people, right? It's going to commoditize jobs, it's going to commoditize knowledge workers, it's going to commoditize expertise as a whole, right? So when we're able to use AI like this, you go beyond, you know, your basic job titles, your basic roles, your basic surface level, how you think you do something, when you put it into, into AI or when you explain it to somebody and you are able to explain, and more importantly, you're able to articulate what makes you so unique and so valuable. Because as it starts pulling out, these patterns over time, these patterns in how you make decisions comes from your experience, right? From the things you've seen, from the problems you solved in that order. And so there's going to be nobody out there that is going to make these series of decisions and these choices and think about things in the exact same way that you do. And so then when you are able to go in, extract that, package it up, and clearly articulate that you could stand out in a way that nobody else in the market is going to be able to stand out. You're going to be able to just be a true market of one in how you talk about yourself.
- [10:30]
A
Okay, so let's now dig into how we can actually get AI to sound and conceptualize and think just like we do. Let's start with any kind of basic foundational concepts that we need to explain here, because you've mentioned some science and stuff like that. I don't know if any of that matters. But before we get into the how, let's just talk about anything that you feel like we should talk about first.
- [10:54]
B
So whenever I approach any problem, and especially with AI, and I know everyone's talking about this, and they are, because it's important is you want to think about context first, right? And whenever I think about AI, I like to get down almost to the root of the problem or first principles of that. So if, if you know that context is important for solving this, then it's first about identifying, okay, what is all the context that we need to provide in order for AI to go in and extract this, right? And so where is all the places that our judgment is showing up, where our expertise is showing up? When are we in our zone of genius? And so many times, especially for knowledge workers, for coaches, consultants, business owners, et cetera, this is happening when we're SOL problems, when we're brainstorming with our team, etc. And again, going back to the transcripts, so much of this lives inside these transcripts. It lives in, in your consulting sessions. It lives again when you're, you're talking with your employees. It lives when you're just doing coaching sessions. Any of that stuff, any of the especially non scripted type of stuff, again, you're not literally thinking, you're just almost on autopilot because this is, this is you operating out of your expertise and getting those transcripts and starting to collect all of those together.
- [12:07]
A
Yeah. When we were prepping for this, you introduced a concept called tacit knowledge. Maybe you can explain. Is this what we're talking about or is this something different?
- [12:14]
B
Yeah, yeah. So tacit knowledge is again, another one of those scientific terms that goes back, goes back years, talked about and invented by Michael Pollony. And it's the whole idea about, you know, we, we know more than we can say. And I'm sure anybody with any level of experience has, has felt this, you know, that, that, that feeling where you just, you know, you, maybe you walk into a meeting room, you could just tell something's off or something that I'm sure would resonate with you, Michael, you know, as a, someone who's really good at copy, is how you can just look at a headline sometimes and you could just tell that something around that headline doesn't quite feel right. You're not sure why, but you're able to rewrite it or, or correct it, but have a hard time explaining that why. And so, so much of this is that feeling that experts have of we just know more than we could tell. It's tacit knowledge. Right. It's this unconscious competence. You, you, you really can't articulate what you can't see or say out loud. And briefly touching on the science behind this. It's just how our brain works, right. As we start over time, start accumulating all these, these seeing these patterns, building this expertise, our brain starts chunking these patterns in the back of our head and it again works in almost a unconscious. We start working.
- [13:28]
A
Love it. Okay, so just to summarize some of what I'm hearing, you say everyone who's listening, no matter whether you are someone who sells your knowledge as a consultant or a coach or an author, or you're someone who just wants AI to kind of properly represent you so that when you use it and it writes on your behalf, it kind of thinks like you think. These are all important things to, to process, which is that, hey, your ability to perform, quote unquote during an interview, your ability to write, you know, a detailed description so that the AI gets what it needs. Those things are not who you really are, right? Who you really are is the totality of how you operate when there is no camera on you, when there is no microphone on you. And I've said this a lot, like I know that when I'm one on one with people, I feel like I'm in my little zone and I feel like some gold comes out of it. And sometimes I wish I had a camera on me because I knew there'd be really good content. And really what I'm hearing you say is that, hey, actually that is kind of when you're probably in your zone of genius and if there's a way we can capture some of that and use that information, then all of a sudden I can understand you at a deeper level. Is that kind of what I'm hearing you say?
- [14:41]
B
Yeah, that's, that's 100%. In that example that you just gave is, is perfect because I, and I bet if we went back and you took, you know, a bunch of those transcripts over time of those one on ones, I can almost guarantee you there's going to be certain patterns that are going to be showing up every single time that you'll kind of know you do, but you didn't know you do, if that makes sense. But you're going to be doing, making these decisions when you know, it's almost like a bunch of if then statements, right? When the person always does this, you always slightly let, let them talk a little bit longer or you always set up the, you know, interviews like this. You always show up with a lot more passion or empathy or what have you. And then you lead like this instead of like this. And so like you're almost, you're just able to codify your thinking so much better. And so if you can imagine as you're able to codify these thinking, these decisions, these, these, you know, almost your unique decision DNA, if you want to think about it like that, that is exactly what AI needs to understand you and to replicate how you work.
- [15:41]
A
Love it. Okay, so let's start with gathering up all this data, for lack of better words, like where in the world do we get this stuff? How much of it do we need? Like, let's, let's get into this a little bit.
- [15:53]
B
Sure. So any meeting transcript tool will work for you. They're obviously a lot more prevalent than they probably were, you know, three, five years ago. So, you know, Google Meets will have them. Any of your Zoom, Fathom, you know, any of the more popular ones. I love a tool called Granola. Granola has a ton of really good preset options and ones you can customize so that it goes beyond just the basic summary, you know, spitting your summary right back at you. It goes super in depth.
- [16:20]
A
Yeah. So what is granola just for people that don't know what it is? Is it like. Yeah, just explain what it is and what it does and why it's different.
- [16:28]
B
So Granola, I would say the biggest difference is a lot of the customization options that it gives you. It doesn't do, it doesn't record video, so it's audio only. A lot of times it'll sit in the background of the meeting. So you don't have that kind of annoying meeting bot that joins a lot of meetings and takes up space. And so it'll sit in the background there for you. And then it has a ton of connections now. So if you need to plug it into, you know, sync it with your notion or sync it with any of your different AI tools like Claude or ChatGPT, Codex, any of the ones you use, it's got really good API connections and then it has really good team functionalities too, where you could do shared work folders and shared spaces and share different, almost prompts on the inside and workflows. So some really good stuff.
- [17:07]
A
I know that I have some friends who wear these pendants. I don't even know what they're called exactly, but they're always recording and sometimes they're like walking and they're just talking to themselves, kind of documenting it. Is this another sort? I mean, are you using these things? Are you even familiar with some of these things?
- [17:20]
B
Oh, yeah. There used to be one called Limitless that Facebook bought up. There's Pocket. Now I'm drawing blank on the other one that I have too. Yeah, like any of the ones, the pendants, you can put them on the back of your phone, you can wear them around your neck. I think they have them. You can clip them to your shirt people, especially if you're doing any kind of live events or you're doing a lot of in person meetings, they're super, super valuable. They can actually stick to the back of your phone now and be voice activated. So when you get a phone call you can actually do some recordings. Obviously I would check what's legal about recording people in your state, but yeah, plod plot is the one. Yes, I have plot as well. So I test a lot of these out obviously as a transcript enthusiast.
- [17:59]
A
Okay, so talk to me about like the kinds of situations where we ought to start recording because there's plenty of people who aren't recording anything and then there's some people who are automatically like, like generally speaking we are at Google House here at Social Media examiner, but we, we don't automatically record every meeting. We have to remember to turn it on. So talk to me a little bit about like what we should record and a little bit of best practices because obviously this is a data gathering exercise we're doing here, right?
- [18:30]
B
Yes, 100%. And you know, moving forward data like we talked about the beginning context is so important. And so the more data you are able to collect just about everything, the more valuable it's going to be to AI later. But for best practices, almost all of these tools now you can set to auto record your meeting. So like you said, you're a Gemini house, you can select it to always record so you never even have to think about it. So there's you know, little settings in your admin panel or what have you. So you could select hey, I just want transcript or I want transcript and video. So you could have the always on type of settings. You know, that's one of the ways I do is try and make it so you don't have to think about it. You know, remove as much friction as possible. Which again why those, those pendants type things or things you wear on your neck can make it a lot easier. And then the types of meetings, again it's where wherever you're, you're showing up and problem solving because usually when you are problem solving or doing a lot of this deep thinking where you, you're, you're not actively planning stuff or you're, you're not reading from a script basically right. That is where you want to capture what's going on there. And so again it, depending on what you do, that's going to change. So you know, a coach or a consultant, anything you can do with those one on one conversations with your clients can be so beneficial not only to you to learn your thinking, but to help you show up better for them later too. You know, I, I've done it with financial advisors so you know, the just Helping financial advisors really at the top of the, you know, the, usually how those are set up, there's a couple advisors at the top, they bring on some junior level advisors. So to have all of their sales conversations recorded and to be able to really show people how they, how they sell and how they close. So you can think about that in any sales situation, any prospecting situation. I also like to do it in any brainstorming or any kind of meeting situation too. So how you show up and how you make decisions, you know, when you're just brainstorming new ideas for a campaign is also going to be different when you, how you show up just in your one on ones when you're helping someone. You know, if you're a manager and you're helping your employees, you know, as part of their, you know, development, just coaching them through that process, anything you could pull out of there is good too. And then when you go and you give it to AI, just making sure it has the proper context of the situation that it's pulling from. So when you see it on the back end, you realize, okay, this is how I'm showing up in this situation. And then you can look at it from an aggregate level.
- [20:52]
A
But since we're dealing with transcripts, how in the world does AI know when it's me talking versus another person talking? Are the tools getting sophisticated enough to know the speaker, their voice and identify those. Do you understand what I'm saying when I'm in a situation where I'm.
- [21:04]
B
Yeah, I guess the nerdy term for that is the diarization of these. And so you'll see some of them like zoom and everything. When the people register, come on and have their emails, it's really good because it will, it will put the names next to the person, but if it doesn't have that, it's good to have AI beforehand adding little snippets to your prompt at the beginning of any of the extraction prompts that say that pretty much ask you, hey, who's, who's all in this meeting. So it has the proper context of who's in the meeting. And there's pretty easy ways to do this in prompt. Have it actually send you a couple of almost like test sentences of it trying to identify who the places are. It can go through and label that. So you want to go through and label all the speakers before it goes and starts pulling it apart and really diving in.
- [21:50]
A
You know what most marketers are doing manually? Ad creative, writing the copy, coordinating with the designer for the images, waiting on Video edits, managing revisions, every campaign, every cycle. I wanted to fix that for our members. So on July 16th, I'm bringing Caleb Cruz to teach inside the AI Business Society. He's going to show members how to build personalized AI pipelines using Claude code that runs the entire creative process on its own. Copy, images, video revisions, all customized to your brand and you don't need to be a coder to do it. I'll be in the room for the entire session. I'm in every session and I can tell you members are going to walk away with a clear, buildable system that they can put to work that very same week. If you can't make it live, the full recording and resource kit will be waiting for you. Transcripts, slides, key takeaways, the whole shebang. This is what we do every month inside the AI Business Society. Come and see for yourself. And don't miss this training. Join@SocialMediaExaminer.com AI Again. SocialMediaExaminer.com AI Whisper Flow is something we talked about as well. I use whisper flow instead of typing, but I would, I also use it sometimes to diary too because I have a journal that I have every day and I'll just speak into the journal about like what happened during the day. How are you using tools like whisper Flow, which are really one, it's just you and your tech and you're just talking in those kind of situations. Do you have any tips on how we can just free flow? Because most of these examples you gave are us giving someone else guidance. But what about us just talking to ourselves? How valuable or maybe not valuable is that depending on what we're doing with it?
- [23:36]
B
No, I mean it's still super valuable. Again, it's all about making sure you're just properly labeling what those conversations are, you know, and why you're having them so it could be processed in the right context. And I'm sure you've seen this too as you're using Whisper Flow. There's something about just being able to talk like into a whisper flow, into a mic, versus typing. That AI ends up giving you such better results. So if you just think about it as just the prompting and like, like you were describing using it, use it all the time for that. So I spend any time, I go to my, my computer and open it up, I'm, I'm starting everything I do inside of codecs or inside a cloud code, right. And so the amount of time it saves me just in there of just, you know, getting a stream of thoughts out to start a project to think through something that's all super valuable in there and then again taking it on, you know, in the car with you or on walks or whatever in. You know, I listen a ton of podcasts and so, you know, a lot of times I'll get these ideas when I'm listening to a podcast. I'll pause the podcast and just, you know, just go through again, just a whole brain dump of everything I thought about the idea, which gets then saved and then you can go back later in and extract and process it. So it is still interesting. Then when you go back and see, hey, how well do these, even some of these patterns or some of these mental models of how I'm thinking about this stuff, how do they overlap with when I'm, you know, meeting with other people or is, you know, more of a one on one conversations versus by myself.
- [24:57]
A
We're going to get into some of these different ways that, that people think, you know, the different layers of thinking in just a minute here. But how much information is enough information? Because obviously if you're a podcaster or a solo podcaster, you've got lots of information. It may be scripted, it may be unscripted. If you're on someone else's show and you're being interviewed usually I know that's potentially a good source because you're probably free flowing a little bit here. But how much do we need to do what we're about to do next? Right, Because I would imagine this could be never ending, you know, depending on how you look at this. Right. Because you know, if you do a lot of talking, is, is a week's worth is we're talking a couple dozen files here. How much are we talking about?
- [25:38]
B
Well, like any personal context file, if you think about it, it probably should be never ending, right? Because you're always going to be changing, you're always going to be developing, et cetera. So you should always be tweaking. But you know, in order for, for it to be a pattern, you want to see it show up, you know, at least three more, three or more times. So at the very baseline you want three to five transcripts and you want them in different situations, right? That is as baseline probably as you're gonna get and just to kind of get started. So if you have that good, I'd go ahead and use it. But the more, the more the merrier. I mean, I work with, I work with people where I'll process. I think the most I've done is probably anywhere between a thousand and twelve hundred transcripts to do it. So you get some, some really good. It's great to see a progression in a timeline, I think. Yeah, one guy was way over a million words, so that got pretty, pretty intense. But really fun and really cool. You could pull out. So the more data the better. But we'll see. As you start doing this and you start pulling in more data, a lot of times all it's going to be doing is just validating what's already there. If it's already seen these patterns show up 3, 5 plus times, chances are it's going to start showing again. So it's going to be just revalidating what you already did. To me at that point, the biggest benefit is doing a check in, you know, like once a quarter and seeing, you know, if anything has changed. But when transcripts become valuable at that point, it's for taking. What we'll talk about later is taking this context file that we built, taking in the new transcripts and then using those to produce different pieces of intellectual property, content, et cetera.
- [27:10]
A
Let's say we've got like 10, let's just say hypothetically there's 10 files, right? Maybe there are they text files and where are we putting them exactly? You know, that's just one of the things that people are, are probably going to wonder like what the heck do I do with all this? Does this go into like a cloud project or a custom GPT or a gem? I mean, is that effectively where we're going with this?
- [27:31]
B
Yeah, that's exactly it. So anywhere where you can essentially put it in a ChatGPT project or a cloud project or a Gemini gem, anything that could pretty essentially simulate some type of rag folder or data retrieval folder, right. And so putting them all in this one place where it's able to have the proper context contain it, and then once they're all in there processing these one at a time and pulling out the insights in a very linear way as you go through these and so it starts building your context file. So again, ChatGPT Project, Claude Project, at a very basic level, those work fine. And it's how I got started with this. As I started doing a lot more of these and get into the thousands, you're going to want to start using something a little more agentic, like a clog code or a codecs or something that can handle much larger files. Like a lot of these things, markdown files work best, but any kind of text file works great. If you have a PDF, sometimes it's better to get it and go ahead and convert it to a markdown or a text first to save it from having to use up tokens as part of that process. But any basic txt txt file or a markdown MD file works really good.
- [28:45]
A
Okay, so at this point, every one of these files, let's just say it's a text file, hypothetically, because that's the easiest thing I'm assuming at the top of the file you've got a context. Like this was the situation where this happened, right. Like this was a coaching call with a client, or this was me being interviewed, or this was me problem solving. Right. And then you've got the rest of the text and we've got all these files. Now let's get into this thinking stuff we were prepping for, like, help people understand, like what's happening, what are we doing in order to get at the hidden wisdom, if you will, inside of these files.
- [29:16]
B
When I was going into this and doing all my research, I started breaking down, okay, what are all the data sources that are available? How can we put this into different thinking models? So I could actually start categorizing this stuff and really understand it not only for myself, but then I can feed it back to AI in a way that it understood it as well. And so there are four different models. And again, this goes all the way back to, you know, 60 years of science that I use every time that I do this. And I start with level one, which is declarative. And declarative basically is pretty much what you'd say if somebody asked you what you did. So it's very surface level answers and it's honestly what you're going to see probably on a lot of job boards, a lot of LinkedIn type posts at the beginning, you know, just very surface level of what you do, what you think you do. Right. Level two is going to be the how. So this is called the procedural level and it's how you actually do what you do. And so it goes in there and it's going to, you know, again, going in, this is more of the step by step, almost like SOP type stuff. So a lot of, a lot of stuff with people when they'll go in and they'll ask for an SOP out of a transcript or something is done. This is a lot of the procedural type stuff that's coming out of that. Right. Level three is when we're going to start really getting into the rich context and the stuff that's going to be most unique to you. Right. So this is going to be what's called conditional. And this is the why. This is when I said decision DNA. It's a lot of, a lot of that. So you know, what is your decision criteria when you do what you do right, like how you make decisions and why you make decisions. So you always think about it as a lot of these if then statements. When this person shows up like this, I'm always responding like X or you know, you're really digging in deep there to understand the why, why it is what you do and your decision model behind that. And so that's number three, that's conditional. And then number four is the layer that has the most valuable data. Especially in this day and age with AI. You're seeing it out there on every social platform where everyone's pretty much posting about level one and level two type stuff. So now when you get into level four, this is going to be the richest, it's the metacognitive layer. So it's how you think about thinking. This is your mental models like this, this is the stuff that's, that really defines, you know, what makes you unique here.
- [31:36]
A
So so far we've got declarative, which is just the very simple matter of fact stuff. Then we've got procedural, which is how you do what you do. Right? Here's the steps I take when I do things. And then we've got conditional. You said this is what's most unique to you. This is the how, how you make decisions and why. And that's really powerful and highly unique to each person. But then this meta cognitive thing, meta means about, right? And cognitive means like the mind, right? So about the mind, this is like a layer above it, right? It's the most valuable stuff. It's how you think about thinking which is really kind of crazy. Your mental models, right? So talk to. That's the part that's really hard for people to self diagnose, right? I mean am I right or am I wrong 100%.
- [32:24]
B
Like especially if you go and you look at a lot of the research and the studies that people just have the hardest time describing because I think, because it just sits at such a meta level, you could ask people like what are your mental models? And a lot of times, and I've done this too, they've tried to give me some pre assessments of what they thought their mental models were. And then when you give them back and you showed them how, what their mental models actually were when they showed up in certain work situations when they were making decisions, they were Completely different. Right? And so it's. Again, the reason it's so hard is because this is your expertise operating from almost a unconscious level. And so until you have a, another, you know, a third party source like AI, an unbiased source in this case like AI, that is just telling you what it is. Like, this is how you're showing up and this is how you are thinking in these situations. It is just, it's super, super, super powerful. And again, this is the stuff that you want to talk about. You know, we talk about in the age of AI, when so much of the execution is becoming so much easier to do. You know, anybody can go out there and create a marketing campaign, you know, in, in 10 minutes, doesn't mean it's gonna be any good. But then go out and say, hey, I've created a marketing campaign or I've created xyz. AI just made that so much easier, so much quicker to do. So how you need to start thinking about separating yourself is demonstrating how you think. And that is so hard to do on your own for all the reasons that we've talked about. You need something to help you articulate how you think about thinking. And when, when people can understand how you think about thinking, they understand you, there's a lot more trust built in. They feel like they know you better. And again, you're, you're different from everything else out there. When you can explain that.
- [34:09]
A
You know, it's intriguing because I think about anybody who is a psychologist or a psychiatrist or whatever they call them. They've got, had regular meetings with people and they understand people's mindset and their thinking because they get to know them really well. A coach, I have two coaches and one of them has been with me for over five years. He totally knows my mental models because I've spent an hour with him every week for like five years. So it's really easy for someone else who's trained to be able to see this in you, but it's really hard for us to see it in ourself. And so how does AI like? I'm very curious. I mean, obviously some people's alarms are going off in their head saying hey, AI shouldn't be our psychiatrist or, or a counselor. And that's not necessarily what we're, we're proposing here, but what we are proposing here is like, hey, AI might help you see the models that you employ. And when you use that, that might be valuable for the following reasons. Explain why that's so important for anyone who's listening when they actually are aware of their models, magnet, you know, thinking models.
- [35:10]
B
So from a personal level, the validation that you will feel when you see these things that you've been doing for years, you know, when you see this being read back to you or, you know, presented back to you, when you see that that level of validation of how awesome you are is just incredible. So just, number one, the level of validation and how good, like, how good that feels to see how awesome you actually are, you know, when you're operating in your zone of genius is one. And then the second part of the. That on the emotional and kind of personal side is the confidence that comes with that when you are able to then articulate what it is that you do. You know, every marketer will tell you this. Like, it's very hard, you know, just a different differentiation to properly explain what it is that makes you different, how it is that you do what you do. And so when you are able to then now articulate this, it creates just such a confidence of, oh, now that I could do this and now I could do this and talk about how different I am in this incredible way. It's gonna. I'm gonna want to create more content about it. I'm gonna want to go share it, I want to create offers about it. I want to go, go, you know, talk about this to help people with this. Right? You know, so that's the second part of this is when you can do that and you can train AI like that, and you can now articulate it and you could see it now you can start creating intellectual property about it, right? You could package it, you could sell it, you can create coaching programs out of it. You can create facilitation guides out of it. You can create sales programs out of it. There's, I mean, the, the. The amount of intellectual property unique to you that you can create out of this is just incredible. And there's so many fun ways you could do it and mix and match it and pull it out of live stuff. It's really cool.
- [36:47]
A
So how do we actually get AI to do this for us? So what are we telling AI to do in order to come up with this information, Right? I mean, is it, is it coming up with all four of these for us or are we just focusing on the metacognitive here?
- [37:00]
B
We'll have it pull out all four. So I have a series of prompts, almost starter prompts at the beginning to. To form your. To form your cognitive fingerprint. Right? And it is, it's just going very deep in looking and looking for all the places in the transcript where it can identify all the structural things going on, all the procedural things going on, all the conditional things, all the metacognitive things. I like to also have it look out for different blind spots that are happening, different things like that. And as it starts building that on transcripts and starts building these databases and your cognitive fingerprint file, or your fingerprint file, whatever you want to call it, it knows now, okay, now I'm pre trained, I know that these are here. It will now also actively look for other places that these are showing up in those transcripts. Right? So the more you give it, the better it's being trained to go look for these things.
- [37:48]
A
By the way, folks, he's going to go ahead and give us these prompts at the end of the interview. So y' all can do this yourself. But give us a couple of tips on like what we're going to be prompting here just so people in their brain can understand. Like maybe they could just go try something right now if with some of their data. Like what's a couple things were telling it, like, give me a little guidance
- [38:05]
B
on this, you know, just off the top of my head, AI is smart enough where if you, you know, just lay out these four layers of, you know, declarative, procedural, conditional, metacognitive, like we talked about earlier, just having the conversation with AI hitting that key and getting whisper flow fired up and just saying, hey, I'm trying to understand. You can call it your cognitive fingerprint, you can call it just, you know, how I'm thinking and I want to break it down into these four different layers. Based off these four different layers, can you go in and identify all the different places inside of this transcript where these are starting to show up? And so at a very basic level, that is how you just get it. Starting to be pre trained to start pulling these things out for you. And then as you go again, they build on each other and there's other ways you make them compound and combine together to find some really cool stuff.
- [38:54]
A
So once we've got AI doing this for us, what is the output you already mentioned, like a fingerprint? What is the output that we're going to get from this? Are we going to get some sort of a big descriptive prompt, for lack of better words, that kind of describes how we think and how we operate? And is this something we can use in our CLAUDE projects, for example, to ensure that whenever we prompt it, I mean, help people understand how this helps them in the future? Get AI to sound and think a little bit more like them.
- [39:23]
B
When I run this for people, especially the more transcripts they have, it will come through and put out a 20 to 30 page doc sometimes. While this is actually super fun to read for yourself, what becomes even more valuable is getting it in a little more concise form, trimming out some of the fat of that and using that as such a rich context file. So we call it a fingerprint file. You can call your cognitive fingerprint, personal file, whatever you want to call it, but it is such a rich file that you can almost think of something you could take anywhere you go, plug it in, and it's going to immediately be caught up to speed on how you think, how you show up, how you act at such a deep level. And that is so important these days because as everybody knows, these models change so fast, right? Fable came out a couple weeks ago, so everyone jumped on that. But before that it was Codex 55, before that it was Claw. So that they just continue to leapfrog each other and people are continuing to switch. There's also the different things like openclaw and Hermes Agent. So some of these more agentic things that you keep on your phone as kind of your sidekick. Think about anytime a new one of those comes out having this very rich context file that you know, if you want to test it, if you want to get started immediately, you could just plug that in. So it's like a. It's almost like your little micro brain USB file that you could take around and just plug in anywhere you go. And no matter what happens in the world, AI because you know something's going to happen, you know stuff's going to change that that can come with you. You can plug it in and hit the ground running right away.
- [40:49]
A
How has it helped you, Max? Do you have any examples of what it's enabled you to do?
- [40:53]
B
Oh man, yeah, it's. It's helped me so much. I mean, from a personal level, it's helped me one just understand how I think, right? Like just being able to see it and read it back and being able to. It's hard to even explain sometimes. But just like again the, the, the validation, the confidence. But then when you could see it, you could almost dig in deeper on it to understand yourself, but then also improve, right? So when I see these, this stuff written down and I can actually improve on it, work on it. I've developed my entire consulting practice around like five, you know, like how I actually show up in these situations. But then the how you do, especially with teams, too. So there's the ability to do this with teams and see how people on a team interact with each other in team meetings. So when I've been on teams, we've all used this and it's made it so much easier to see who should actually be doing what in certain, you know, taking on certain projects or certain tasks based off of how they're actually showing up in those projects, in those meetings, where are they engaged, where are they, you know, all those different things. You could do it in leadership teams, too, when you start thinking if, you know, if it's you and a partner being able to see both of your blind spots and seeing, like, how you guys complement each other from a blind spot, or is this person, you know, are they. Do they think much more analytical? And so, like, if I were to go to present something to them, I need to make sure I'm presenting it with a lot more data first. Whereas me, like, I. I'm a visual person. I want to, like, hear stories. I want that. So if someone were to present something to me, if they understood my thinking, how they presented new ideas to me would be completely different. Right. So there's so many different ways it could show up.
- [42:27]
A
What I love about this is there's so many of these tests out here, Right. That are personality tests, but they're self assessments.
- [42:34]
B
Yes.
- [42:34]
A
What I like about what we're doing here is this is actually a assessment based on actual things that you've said, not on you assessing yourself and answering questions. How would you respond to this or that? You know, there's. There's a bazillion of them out there. There's strength finders, there's Myers Briggs, there's Disc Ian. You know, there's a bazillion of them. So what I like about this is this is actually something that is an assessment of you based on the things that you actually say and do, which I think is super, super cool. Max. I know. We have just simply scratched the surface.
- [43:06]
B
Oh, yeah.
- [43:07]
A
Of how you can create something like this. So why don't you tell everybody where they can discover more about you, where they can connect with you online and where they can get that prompt. Sure.
- [43:16]
B
So the easiest place to find me is at CognitiveFingerprint AI and I put the prompt for you guys on the site. So if you do SME at the end of it, I'll put the prompt. I'll put some other goodies on there and I'll have all the other links for all my social platforms and newsletters. I write about this at least once a week on my sub stack. Just talking all about Invisible expertise and Expertise Paradox and some of the science behind this too. And some really fun stories of how this shows up. So some really, really cool stuff and a lot of. A lot of AI. You guys know I'm a big AI guy, so you'll get a lot of that too.
- [43:50]
A
Max Bernstein, thank you so much for coming on the show and sharing your wisdom with us.
- [43:53]
B
Yeah, thanks for having me. This is awesome.
- [43:54]
A
Hey, if you missed anything, we took all the notes for you over@social mediaexaminer.com A114. Be sure to follow this show on your favorite podcasting app. And if you've been a longtime listener, we would love a review on whatever platform you're listening on. And do check out my other show, the Social Media Marketing Podcast. This brings us to the end of the AI Explored Podcast. I'm your host, Michael Stelzner. I'll be back with you next week. I hope you make the best out of your day and may AI help you become more successful. The AI Explored Podcast is a production
- [44:27]
B
of Social Media Media examiner.
- [44:31]
A
Before you go if you want to stop grinding through the AD creative process manually, the AI Business Society can help. Caleb Cruz is teaching members how to automate the entire process on July 16th. Full recordings and all the resources are included with your membership. Join at social media examiner.com AI again social media examiner.com AI.