Connection Is the New Currency with Jason Yeh, Part 2

This week, Dan picks up right where he left off with Jason Yeh, founder of Scaling Freedom, for a rapid-fire dive into the real-world side of AI. Jason shares battle-tested frameworks for deciding when to use AI (and when a simple script is smarter), explains why human connection will matter even more as automation spreads, and gets personal about parenting, education, and staying nimble in a market that changes every week. If you’re wondering how to weave clarity, technology, and meaning together—without losing the human spark—this conversation is for you.

Show Highlights

• Treat AI like a smart “drunk intern”—speak out loud, let it interview you, and watch the prompts improve [00:00:31]
• Build a custom GPT that turns raw thoughts into perfectly engineered prompts [00:02:37]
• “AI on a bad system just scales the mess”—why process always comes before tooling [00:04:36]
• The hidden gold mine of manual data entry and how scraping invoices unlocks insights [00:06:51]
• Jason’s 4-point litmus test for any AI project: Painful · Existing process · High impact · Strong data [00:09:56]
• Planner vs. Solver: use one “conductor” agent plus narrow specialists for higher accuracy [00:17:13]
• Klarna’s chatbot fiasco & the lesson: some moments demand a human, not a bot [00:18:28]
• “Connection is the new currency”—brands that stay human will win customer loyalty [00:23:09]
• Staying nimble: why Jason refuses to ossify into a single product in a hype-driven market [00:24:45]
• The next frontier: AI-personalized education and raising kids who still think critically [00:28:12]
• Using Perplexity for curiosity walks and co-creating bedtime stories with your kids [00:32:37]

For more updates and my weekly newsletter, hop over to https://betterquestions.co/

To learn more about Jason Yeh, check out the websites below: 

https://scalingfreedom.ai/

https://www.linkedin.com/in/jasonjyeh/

Transcript:

[00:00:07] Hey everyone. Welcome back. You are listening the second part of last week's episode. Connection is the new currency. Let's jump back in.

[00:00:15] I'm curious, like obviously you do a lot more with AI than just the chat interface, but because that's kind of the consumer model of it. Mm-hmm. How do you think about like when in your own use of ai, just for your personal use or whatever, how do you think about like asking it the right thing?

[00:00:31] Or is it just a matter of like you said, like, what question should I ask you? That's kind of an easy one to always ask, having it prompt you. Uh, 

[00:00:39] yeah. Two, two tips. One is stop using the chat interface by typing in to chat. Oh, gotcha. And that's, that's a hot take. And, and what I mean by that is treat AI as a person.

[00:00:56] I know it sounds weird. I, I heard this other analogy the other day I thought it's quite funny, is, um. Treat AI like a drunk intern.

[00:01:06] So, you know, it's like, 

[00:01:09] you know, and you're slowing it down a little bit. You're like, all right. 

[00:01:13] Yeah. Yeah. And it's, it, and it's, um, and I think that analogy encapsulates it perfectly because what it tells me is that you still need to give it a little bit of guidance. 

[00:01:24] Yeah, 

[00:01:24] right. But, but it's, it's, it's not just a drunk intern.

[00:01:27] It's a really, really smart, drunk intern. Mm. And so, and what I meant by not typing is I think if you could just talk to it like a human being almost. Yeah. Right. And then put it on voice mode, uh, and just have a conversation. Just treat it like a conversation. Then I think you stop getting in the way of yourself.

[00:01:51] Now I know there's gonna be purists out there that says, whoa, whoa. Like, you know, I. Human language, you know, that may not be the most [00:02:00] optimal way to prompt to get the best results. Well, my argument is that the OMS were, you know, and the GPTs were trained on human language. 

[00:02:10] Right? 

[00:02:11] Right. So why not give it back what it's used to?

[00:02:13] That's one point. And then the second way is if you truly want to be purist is, um, it's a, it's kind of meta. You can actually design a custom GPT for prompting specifically. Yeah. And, and so every time you want to say, drop your stream of consciousness, Dan, maybe you could drop it into your custom GPT that focuses on prompt engineering.

[00:02:37] Mm. And let it 

[00:02:38] transform your words into an ideal prompt, and then copy and paste that into the chat interface. 

[00:02:43] Yeah. 

[00:02:44] Another hack. I, I 

[00:02:46] do love that. Right. It it, I, I, that's the other thing, right? It's like not thinking of it as one thing. But thinking of it almost as like a, some kind of underlying engine that you could put in a bunch of different vehicles and chain those vehicles together.

[00:03:00] Right? So a lot of AI automation is, you know, I'm, there's 17 different prompts and I'm chaining the output from one to the other and clearing the other. And it's like, you get these like really wild, interesting outputs. Yeah, I think that's, that is super great advice. Okay. So, all right, so today when you are, you know, you are working with businesses integrating AI into their workflows, right?

[00:03:24] And you, one of the reasons that I, you know, we were talking before we hit record and you were like, you know, why did you guys wanna re reach out to me? One of the things that really sort of jumped out to me was like, your systems focus, right? And like, one of my complaints about just AI, I don't wanna say AI media or AI hype or whatever is like.

[00:03:48] You can throw AI onto a terrible process, and then it's just a terrible process with AI attached. So it's like, it just makes it weirder, but not necessarily better. You know what I, yeah. And so, you know, and, and I, [00:04:00] I've always just been like, we, we'd never even really figured out systems before jump to AI.

[00:04:06] Right. And, um, so I'm curious, like what you see as the big opportunities for businesses in terms of maybe different ways of integrating AI than they're thinking about it right now. Right. Like the classic is like, we're gonna have a chat bot on our website, and I'm like, yes, I'm, you know, great. Uh, but like, what are, what are people missing or, well, maybe a better way to answer that question is what's interesting to you now in terms of how people integrate AI or can 

[00:04:36] Yeah.

[00:04:37] Uh, you know, Dan, it's, it's kind of funny to paint this out because, uh, I, in, in my, in my experience. Marketers, like, no, no job against marketers. I know you're a maker, 

[00:04:50] but we're all terrible. We, my friend, uh, different friend Patty, uh, de Patti Dio, really great friend of mine, SEO person, she'd always say this, I remember this saying forever, say, marketers ruin everything.

[00:05:04] And it is absolutely true. We ruin everything. That's okay. You're a estate. I have a lot 

[00:05:08] of good friends in, uh, that are marketers. So I, you know, I, I say this with a lot of love as well. Um, so, um, so, you know, when, when I fi, when I first started looking intently at the space, I just realized that a lot of the use cases were the same.

[00:05:23] Um, and, um, you know, another thing I like to say is that, you know, while Mark marketers are loud, you know, operators just kind of get their stuff done. 

[00:05:31] Mm-hmm. You know, 

[00:05:32] behind the scenes. And I, I can say this because that's actually what happened when I was at Nike, right? You know, Nike's known as this giant marketing machine.

[00:05:40] I. How do those shoes get actually shipped around the world? 

[00:05:44] Difficult. 

[00:05:44] Um, so when I started paying attention to all the AI hype and news that was out there, it was a lot of the same thing, right? It's like, um, hyper personalized ad copy, right? Or like personalized like emails, um, you know, automated, uh, [00:06:00] AI that would, um, write these customized email to Dan and say, Hey, like I read your last couple posts, you know, here's what stood out.

[00:06:08] And it was all AI generated. Now you just kind of delete those goes in spam. Um, you might also talk, you know, you know, there was also a lot of chat bot opportunities as well. And, and, and because I've been in the trenches as an operator the last two decades, I was like, I think there's, I think there's so many more opportunities than, than what I'm hearing.

[00:06:29] Um, and so how, how can I let leverage AI as a tool to solve the existing operational problems today? 

[00:06:36] Yeah. 

[00:06:36] Um, and to answer your question more directly, um. One of the themes that has been, well, two, but I'll share, I'll share the most prominent one is, do you know how much manual data entry work exists today?

[00:06:51] Like, think about manufacturing industrial operations, think about construction, think about healthcare, retail. Um, uh, so imagine you're a, uh, retail business, or, um, let's say you are a, uh, or, or, or a gift shop for instance. And you know, you're receiving a ton of invoices from your vendors. So these come in, um, you have to, uh, typically there's someone looking at the invoices.

[00:07:23] You manually entering it into A-P-O-S system, um, like toast, for instance. And think about how many of these jobs exist. Yeah. And I was like, I. I think the question I think every business owner and founder should ask themself is, how can AI help me solve this? How can AI help me solve my number one piece of constraint in the business today?

[00:07:48] Yeah. And I'm like, I, yeah. There has to be a way. And so what we're like, a lot of the work that we're doing right now is like taking all the, the manual work, um, of [00:08:00] reconciling documents. What we're doing is we're scraping that information, organizing it into clean data outputs. Once you have clean data output, then you can have AI synthesize and give you insights Yeah.

[00:08:14] Of what information you're looking at. Do you, 

[00:08:19] in what I'm curious about, particularly about a, how a AI works with data and the, this is, this is gonna be a slightly wonkier tangent on this podcast, but That's okay. Like, at what point are you, so here's a question that I, I often find myself asking, which is like, I can use AI to do a thing, right?

[00:08:46] I could also use like a JavaScript to do this thing. And it's like one of, one of my worries, and I can't tell if this is based, this is more a holdover from like older models, but it's like particularly asking things to say like, okay, for example, I, I run Google Ads. It's a lot of what I do, right? Let's say I'm generating a client report and I can put statistics, I could pull it from a spreadsheet, I could put in the report and I could say like, give me insights based on this.

[00:09:14] Yes. But sometimes it's doing like calculation and I'm like, okay, well I don't, I don't really know where that calculation came from. Like there's no, it's not necessarily running a script to do it. And so I go back and forth on like, well, how much. How much am I trying to shoehorn this into AI? Because I think AI is cool and I want to use the cool tool for the thing, versus how much should I just like send someone the spreadsheet?

[00:09:37] And I'm curious like how you think about like, when do you pass, well maybe ask this question in a way that's more useful for people. Like when do you decide to pass a task to AI versus more traditional automation tools? Like for example, just scripting it or running it through like a set workflow that's not necessarily gonna change.

[00:09:56] Does that make sense? 

[00:09:57] Yeah, it does. I love this question [00:10:00] because I, I think it's kind of like the, um, you know, fitting a, what is that called? Like, you know, if you're trying to like fit the same object into the same, you know, well you have squares the square. Yeah. Square pack into wrong. Yeah. Thank you. So, so what I typical, how I typically advise founders is, um, I think of these as these four check boxes.

[00:10:24] As like the most ideal and optimal case for AI. So number one, is it painful or manual? Okay, number one. Number two, is there an existing process for it today? And I think a lot of founders miss this. Number three, what is the impact of solving this? Does it either hit your top line, does it, is it a big impact to bottom line?

[00:10:54] Right? Otherwise this might not be worth solving. 

[00:10:56] Yeah. 

[00:10:57] And number four, this one isn't one that a lot of folks miss as well. And I think goes back to your question is like, do I have strong data? Like what is my data foundation here? Yeah. Because the more robust your data is, you know, it's, it's like the whole, um, garbage in, garbage out that a lot of data folks know about.

[00:11:18] And so. If it meets these four boxes, it's likely a good candidate for AI. Now, to answer your question directly, I would say if you can use it without AI, start there first. Yeah, always. Like I always ask, I always like to better understand what is the current state and what is a future state, and then trying to understand what that gap is and then trying to figure out, well, what's, what's the ideal match for AI to solve that gap?

[00:11:53] Because maybe it doesn't make sense for AI to solve the whole gap. 

[00:11:57] Yeah. It's like [00:12:00] slotted in where it makes sense to slot it in rather than trying to shoehorn the entire problem into one tool because. Yeah, theoretically you can. That's like kind of the temptation, right? It's like, well, I can just dump the whole problem in the chat bot or whatever, and like it'll do something.

[00:12:17] I dunno what it'll, I, so what percentage of people who come to you and they're like, okay, great. I I am, I'm on the I bandwagon, let's go. And you get started and, or maybe you're, you know, you're in discovery with them or something and you just realize like, oh, like your real problem is you have zero data, like legibility.

[00:12:35] Like there's, it's the boring thing that you need to do before the exciting thing. Like, is that, is that happening a bunch? Uh, when people come to work with you? 

[00:12:44] So the first thing I do is I temper their expectations as like, 

[00:12:50] this is class that you have to do this. 

[00:12:51] This 

[00:12:51] is, yeah. 

[00:12:52] Yeah. It is. Like, uh, I think a lot of folks come to me and, and they think the world of AI and, and um, and I think AI can do some really impressive stuff.

[00:13:03] I think it's important to scope out. Something that is meaningful but also achievable. 

[00:13:10] Yeah, 

[00:13:10] right on. If, if you were to graph a two by two, like, what's that intersection of lower effort, but high impact, ideally something that's high visibility to your organization. And so then you get buying and adoption as well.

[00:13:25] So people are really excited and they're like, holy crap, that, that's a wow moment for me, me. Yeah. Because I, I think it's important to figure out what is it that's going to create the wow moment for your stakeholders. And uh, that's, that's something that I'm, I'm always striving for. And, and to answer your question more directly, um, I like, we, we try to do as good of a job as possible upfront during discovery to ask those tough questions, right?

[00:13:54] Like, well, what is your existing tech stack? Right? How do you envision, [00:14:00] you know, the output of this. Plugging into your existing tech stack, what is the outcome that you're looking for? What are you gonna do with this? An AI output, like what is it going to solve? Like, what problem is it going to solve? Um, we, we typically scope out A-A-P-O-C that doesn't require complication.

[00:14:20] Like that's more for like the full delivery. Hey, once, once the POC is good, you sign off on it and it works, then we could worry about, you know, full integration with your existing tech stack. 

[00:14:32] Yeah. I was gonna, I mean, one of my kind of perpetual questions about AI and particularly AI as integrated into the real world economy and real world businesses, right?

[00:14:45] Mm-hmm. Is about complexity, because it essentially, you know, not, not so much the approach that you were talking about, but let's say it's like, okay, well I. We, we throw a bunch of stuff into an LLM and we get out a thing on the back end, and essentially what you've done is insert a black box into your work.

[00:15:08] Mm-hmm. Right? Mm-hmm. You're, I mean, I don't know about you, like in the, it's probably 'cause you're, you're doing this professionally. Like you, you probably have more sale fail safes in place than, let's say, I would, who's more of a tinkerer. But like, what I will often find is like some percentage, it would be anything from like 10 to 30% of runs of some kind of, let's say an automation that uses AI that's like more than 10 to 20 steps longer, whatever.

[00:15:34] Some percentage of that output is just gonna be unusable for what, for one reason or another, right? Yeah. Like it just, this was supposed to be 3000 words and it's a hundred words and like 

[00:15:44] Yeah. 

[00:15:45] And you can try to go back and figure out like where that happened, but ultimately, so much of the. Machinery is just out of reach.

[00:15:54] Right. Even if I worked at chat GPT, I couldn't tell you why you got this output in [00:16:00] particular. Right. And so I always worry about this sort of like additional, this addition of like, I don't wanna say like chaos, but like weird. I always say like weird every, it's like things are become gradually weird because like all of a sudden, like Phil from accounting sounds really weird.

[00:16:19] And it's because like his email like emailed you on its own and it was like Phil quits or whatever, right? And he is like, I didn't mean that. You know, like, it just, it is, it adds this layer of weirdness to things that it is. I don't know, it's like obscure in a lot of ways. Mm. I'm curious like how you think about that or like whether you, whether that's something you can solve for in the kind of systems design element of what you do.

[00:16:45] Um. In terms of the precision or in terms of the more human touch aspect of it? 

[00:16:50] I think I, I think I'm asking about both, but you could take that question wherever you want to take it. 

[00:16:54] Yeah. So in terms of precision, uh, another framework that might be helpful to solve that is planner versus solver. Mm. So we, we, we, like, we design and build hyper specialized agents that solve a very, very narrow problem.

[00:17:13] I think a lot of the challenge arises when folks try to build a super agent that this one agent that solves 10 different problems. Instead of doing that, we have a planner who's more the strategist or the CEO you might call, um, in AI terms. This is known as a conductor agent. So they're more of like, kind of think of the CEO hat.

[00:17:34] Yeah. 

[00:17:34] Um, it's more the planner and then you have these specialist agents. Let's say those 10 problems, you would have 10 different agents that will solve that neuro. Problem. And I, I I, I find that the more narrow, um, the agent is against that specific problem space, the, the higher the quality output is, is what we found.

[00:17:55] Yeah. And then in terms of the human thing, my, my hot take is [00:18:00] that as AI evolves and as AI starts to almost, um, encompass a larger and larger part of our interactions, whether in terms of business, work or life, that the human part of it is going to become even more and more important. Meaning like what makes Dan special and why I like having this conversation today is because this is a human connection.

[00:18:26] And, and, and from a business context, um, my advice to founders and business owners is what, what, what's an area of your business that, where that. Human interaction to your customers is actually more valuable than the AI interaction with your customers because business owners automatically think of what, like chatbots.

[00:18:50] Right. But I would, my challenge is that there's probably areas of your business where chatbots does not make sense. Where your customers would actually appreciate more human touch. Oh, a hundred 

[00:19:03] percent. A hundred percent. Right. Because I don't know, I've never met your wife. My wife, if she has to talk to a chat bot when she wants to talk to a person, she almost murders.

[00:19:16] They walking by, you know what I mean? Like Yeah, yeah. Right. You can see her like tensing up and it's like the moment you can tell, you feel rejected almost, or you know what I mean? Yeah. It's like I'm trying to talk to you and you like put up a wall between us. It just implies a sort of, uh, did, did you 

[00:19:36] hear about the Klarna debacle?

[00:19:38] No, no. So Klar is like a, it's like a credit 

[00:19:41] thing. Yeah. It's like, it's like a buy now pay later type of service. So if you show up at the checkout counter and, and you don't, you don't wanna pay for it, um, a hundred percent you can go on a payment plan. So they're based in overseas and um, they, I think they automated like, uh, [00:20:00] 700 or so of their customer service staff replaced 'em with chatbots and it went horribly.

[00:20:06] And one of the reasons was, um, they actually did not disclose that they were AI also. That was like, not only, so that was like one aspect of it. And the other key learning was it was making the chat bot do very complex things like, Hey, maybe there's a fraud alert on your account, and it would just give 'em a rigamarole, like, 

[00:20:28] and 

[00:20:28] they would, and it took 'em a long time to finally escalate to human agent.

[00:20:32] And when they. Escalated to a human agent, the human agent wouldn't have, or the real human, rather. Yeah. The, the human wouldn't have the context of everything that was discussed. Yeah. So there's a whole case study on it, and I think it's fascinating. 

[00:20:44] Wow. Okay. Cool. Dude. Do you know, was that published somewhere?

[00:20:47] Like a Harvard Business Review or something, or? Yeah, 

[00:20:49] no, I, I mean, I, I think it's, um, I mean, it's all over the news. I think it's just All right. I'll look for it. I'll look 

[00:20:55] probably to it. Yeah. I am like fascinated by that stuff because it's like, I almost, like, I know AI doesn't, I don't wanna say no, I don't think AI has feelings necessarily, but as someone who's like compulsively polite to my AI, like, just because I'm like, I don't know, someday maybe you'll come for me and I want you to remember the, the good ones or whatever.

[00:21:15] It's like, I feel, I feel just bad for everybody involved, right? Like I feel bad for the, the human that got past the call 'cause they just got, got a call that's on fire and now everyone's mad. I feel bad for the caller and I feel bad for the AI 'cause they're like, I'm doing my best. I just have zero context for like the real world or whatever.

[00:21:33] Um, I mean, I, I, I think you're so right about the human element, right? Because I think there's a lot of people out there who, like, their first instinct is to get rid of as many human elements as possible. Mm-hmm. Right? Like the ideal in their head is like this machine that operates with nobody involved but like.

[00:21:51] Every single time in history, we've increased the amount of dehumanized systems that we come in contact with. Whether that was like the rise of the industrial [00:22:00] Revolution or it's like post First World War, like whatever. We have these like increasingly depersonalized systems. All of us feel more alienated and we all feel worse and we all feel like life is terrible.

[00:22:12] And it's like, it's just like you said, it's that ability to just connect that really centers us and makes us feel whole, makes us feel appreciated. It makes you want to do business with a company 'cause you're like, oh, they made me feel that. Right. And like for so many companies today, it's like they get so confused by the fact that like it, it's what isn't here, that isn't what really matters.

[00:22:35] And they can't just like let go of that old paradigm and actually almost go back to an earlier paradigm of like being able to connect with people, right? Mm-hmm. So I think you're absolutely right is as it becomes more of a, um. As it becomes more available as an option to like just weed people out.

[00:22:53] Mm-hmm. There are gonna be companies that like, well, we deliberately don't do that. Yes. And so, you know, oh, like, you know, if I call you, you're gonna have a person, like, I'll pay $10 extra a month. You know what I mean? Like, just Yeah. You know? I think that's amazing. 

[00:23:09] Yeah. And that totally resonates. I, I, I think that in, um, in an increasingly AI driven world, connection is a new currency.

[00:23:18] Yeah. I love that. That is huge. Absolutely. Plus like, not to mention being able to generate AI, video of yourself and AI, audio of yourself. Just being able to be like, okay, well, do I, do I actually get the real you? Mm-hmm. Yeah. Um, I think that's absolutely huge and it just makes, you know, makes all the sense in the world, man.

[00:23:40] So that's cool. But I mean, so you're, you're doing so much stuff with your company and you're, you're in a space where it changes so fast. Which is the other thing that I always, I always wanna talk about. It's like when you, when you are in a field where the skillset changes so rapidly mm-hmm. How do you, how do you plan [00:24:00] your future?

[00:24:00] Like, and what do you think about your next three to five years? Like, what do you kind of see that journey as being for you? Um, I, 

[00:24:09] I, I treat it as a form of discovery and I, I think it's, there's a lot of founders out there in the AI space that are saing their product. Yes. Yeah. And so there's this whole debate, right?

[00:24:25] Is it better to stay nimble and try, um, you know, do more custom bespoke AI work? Or should I narrow down on a, um, solution that solves a, you know, deep problem. And I've chosen the former path and I think, I think right now, um, I. Silicon Valley is like littered with AI startups that are, you know, either, either going outta business, um, I think like a builder AI is, is a, is is a good story on that.

[00:25:00] Um, where there's, it, it, there's so much AI hype around what they're doing. They raise a ton of money and for whatever reason, um, it didn't work. And the other side of the story might be, you know, like these big incumbents like the Googles and the Microsoft of the world, they have distribution, they have users.

[00:25:20] So you're one feature away from, from being killed, right? And, and so, you know, a lot of these startups that got a lot of funding, um, you know, their whole business model is disrupted by one AI enhancement. Like, that's as a feature stack on these incumbents, right? So I, I made a conscious decision and choice to say, I am, I wanna stay nimble.

[00:25:45] I'm still, um, I, I, because the space is evolving so quickly that I think it's important to stay adaptable and nimble. And this gives me the ability to see trends to, and I, uh, insights for, to [00:26:00] surface. And once I start seeing and pattern recognizing, I think that's the time to double down. Um, so that's, that's why I'm excited.

[00:26:07] I don't know what is to come over the next three to five years, but I do know that I, you know, might staying nimble and adaptable, I can adapt to that. 

[00:26:16] I, I think that's so smart. And, and I think it's ultimately, I think the most successful people in AI are these people who, like, they're able to just, like you said, sort of like dance on the wave as it moves, right?

[00:26:32] 'cause we're so early, right? It's, it's actually impossible to be like, well you, you have people who are like professionally are like, well, here's what's gonna happen in the next 10 years. We're gonna do that, da. And I'm like. Dog. I literally, none of this makes any sense. Every day I'm like, oh, okay. Now this, you know, like, it, it's such, and I do feel like we're also gonna bump up against limitations that we don't know yet.

[00:26:55] Right. So it makes so much sense to be nimble and agile and to be okay with that uncertainty, right? Yeah. Because ultimately that's what you're saying. It's like, I'm okay with the uncertainty. You have confidence in your skillset, and ultimately it's like, it's not a big deal to just say, I'm gonna put off, you know, ossifying everything, setting everything in stone to me.

[00:27:16] Like, this is exactly what I do. This, you know, this, I'm gonna put that offer now so I can take advantage of these opportunities as they arise. I think that makes so much sense, dude. I think that's the approach that, uh, smart people will take in this space. Um, so I think that's cool. So like, what do, what do you see as your biggest challenge?

[00:27:37] Like, what is the thing that you're like. That's the thing I'm focused on, you know, for your own business, for you, that's the thing that you are focused on addressing. 

[00:27:46] Yeah. Without, 

[00:27:47] without. 

[00:27:48] Yeah. I'm, I'm gonna give you a non-work answer 'cause I think it has deeper meaning and I think the big challenge, um, [00:28:00] for our educators and our parents is actually how AI is gonna be used to transform education for the next decade.

[00:28:12] Yeah. And so I'm, I'm thinking about like, hey, I'm, I'm working in ai, I'm have, be building solutions in AI for my clients, but what's the world that my, my kids are growing up into? Right? We, we talked about, you know, as each generation, like they're being more and more used to technology in their hands from the wo right?

[00:28:32] We talk about the tablet generation and now it's like the AI generation, right? Like they're born into AI. What is that world gonna look like? Uh, and so that, that, I think that is a big challenge. I don't know how to solve that. And I, I, I think a, I think most parents and educators automatically go and think like, oh, it's, it's like, AI's gonna do your homework.

[00:28:53] But, but I, but I think the implications are far beyond that. You know, it's, it's, you know, you know, if the classroom, if we believe, if our thesis is that the classroom is, you know, in addition to parenting is where we shape the next generation, then how our classrooms going to evolve to take the best of AI and, and amplify those efforts so that it, it facilitates learning and not, you know, so I'm, I'm most excited by like, if, if we can figure out how to leverage AI to build custom customized education to meet each kid.

[00:29:35] Based on where they're at. Right? If you think about the traditional public school system, you have one teacher to 30 to 40 kids going through the same default curriculum. Like, does that make sense? Yeah. Now, if, you know, if, if you compare the top performer or someone that truly understands the subject to someone at the bottom, there's a wide gap.

[00:29:59] And [00:30:00] I, and I, and I believe there's a way to fine tune the classroom experience with the help of AI to create a truly customized experience for every kid. So I don't know what it's gonna look like, Dan, but I'm, I'm, I'm excited about, um, the opportunities there. 

[00:30:18] Uh, I mean, I, dude, I love that you talked about that.

[00:30:21] Like, so I was a teacher before I started doing my agency. My mom was a teacher, my aunt was a teacher. My grandmother was a teacher. Yeah. I was, I have two young boys, like they're both in school now and I think all the time about, um. How do I want them to use this tool? Right? Because right now, I mean this, I'm transitioning to just personal talk at this point, but it's like they're very limited in terms of their device time.

[00:30:45] Like they're very, very, yeah. It's, it's gated behind chores. It's only a certain amount per week. Like it's on specific days, right? Mm-hmm. And they mostly use that time to play video games. 'cause it's like, that's what's more fun. Yeah. Yeah. My oldest son particularly is much more interested in, in computers and you know, and we've done some time where we're like, okay, well let's see if we can use AI to make a video game.

[00:31:07] You know, we, you know, cool. Yeah. Like, yeah. And I, and I, I think about all the time, like he's the kind of kid where whenever he has trouble in school it's because he's bored. He's bored. Yeah. And I'm like, and he's bored because the work sucks and everyone knows it sucks. We all know this is pointless. You know what I mean?

[00:31:26] But it's like you're just gonna make him do a worksheet because that's what you do. And like, it's just this whole thing. At the same time, I worry, like, I don't want him to get out of the habit of critical thinking on your own. And it's really easy to be like, well, I could think critically of this problem, or I could have a super intelligent alien life form.

[00:31:45] Tell me what it thinks. Yeah. You know, like, gimme a thing. And so I, that balance is, it's so critical. And I think you're right. It's gonna be the, I mean, who knows, but it feels like it's gonna be the big question of the [00:32:00] next 10, 20 years. So it's like, how do we use this to help kids rather than like limiting them?

[00:32:06] Um, so it's an incredible thing. I'm curious, so do you have a plan for your kids? And your kids are, are pretty young, um, but do you have a plan for them or do you think about like, you're like, here's how we use this here. Do you set limitations on them or how do you guys interact with that tool? 

[00:32:24] Yeah, we, uh, so my kids aren't using AI yet, but, um, two, two things that I've integrated into, um, uh, you know, my interaction with them that's been really fun is number one, perplexity.

[00:32:37] Yeah. Um, so I, I, I, perplexity has replaced school for me. Mm-hmm. We'll stop. And one of the coolest things is, let's say you're, uh, you're at a park, you know, you're at a, you're at a, um, or you're anywhere outside and, uh, they're playing in the, the jungle gym, et cetera, and you see an animal that you don't recognize.

[00:32:59] And so the other day I just took my phone out and I just snapped a picture of of, yeah. And it was like, oh, Egyptian geese, right? Oh, cool. Yeah. That's cool. Right? And it's like, and it's like factual, right? It's like, oh, you know, so, and I, you can almost play this like tour guide or zoo guide for your kids.

[00:33:20] And I did this, I did the same thing with plants. So we, we noticed that there's, there's these, um, I never played super close attention to the flowers in my backyard, but now we're able to go like, you know, species by species and be like, oh, what is, what type of flower is this? You know? And, and so you kind of engage 'em in curiosity as well.

[00:33:41] It's like, oh, dad doesn't know what it is, you know, let's figure this out together. Right. Um, so that's, that's one thing I'm, I'm doing. And the other thing that is fun is, is storytelling. So, you know, thinking about, uh, so my kids love stories and he's always say, Hey, dad, can I [00:34:00] hear a story before I go to bed?

[00:34:01] So, um, and I'm always making stories up on the fly. Yeah. And what if you can, and, you know, leverage AI as your storytelling thought partner? Mm. So you take like, what are your, you know, like what are your kids' favorite characters? What are your kids' favorite food or your kids' favorite things to do? Now give those ingredients to AI and help the, you know, have it help you draft three versions of a bedtime story that you can kids.

[00:34:32] Yeah. Right. And then you say, Hey, like, you know, make this more suspenseful. You know, add a hook there, you know, add a cliffhanger there and, and don't make sure that you leave it there for this story to be told tomorrow night. Yeah. Right. So there's some really creative things you can do. I love 

[00:34:50] that. Dude.

[00:34:50] I, uh, and it, again, there's so many ways to use this tool. Like we were talking about, it's the anything box, right? Which is, which makes it tough. But I think those are like such great sort of real world examples. And you're building in, you're modeling for them a sense of agency, right? It's like, I had a problem and I used the tool to solve the problem.

[00:35:12] Like I wanted to know about the plant. Here's what are my resources, here's a resource available I could get. Right. Just that, that process of being curious. Seeking out information about the world, I think is so critical for them. Um, dude, I, you know, we're, we're past time. I don't, I don't want to, we will talk to you literally for like five hours if you let me.

[00:35:33] So I'm, I'm gonna make you get off the call, but I just wanna say like, thank you. Like, literally thank you so much. This was like, such an amazing like, sort of interview and you, you opened up so much about yourself and how you view the world and your journey. And I think you're doing just incredible things in the space.

[00:35:52] And you're like, you're so thoughtful about it where there's so many people who are like, just, they're like, you know, AI is like [00:36:00] peanut butter. It tastes good on everything. It is smear it all over, like whatever thing they're doing. Right. And I think I love your thoughtful approach to really thinking about where it benefits and where it benefits people specifically.

[00:36:11] So yeah, I just wanna say thank you so much for being here. It was a, it was an absolute pleasure and I really appreciate it. 

[00:36:17] Thank you, Dan. Uh, really appreciate your, uh. Insightful and introspective questions as well that drew that out of me. So a lot, a lot of it is, uh, you as well. And by the way, I forgot to give you a shout out.

[00:36:29] Um, I have on my bookcase this book, um, uh, I don't very, yeah. Uh, I think, I think you interviewed, uh, it was a setter Chin. Yeah. Your first episodes. And so, uh, I am almost done with the book, and now I'm like, oh, working at XMR, you know, chart. Oh, he 

[00:36:50] might connect you and Cedric over email, Bethany, because he's the man, he MR charts.

[00:36:55] And I might, my kids are like, please stop talking about XMR chart. Be good on March chart. That's amazing dude. Alright, well for people who wanna learn more about you, we talked about, uh, 

[00:37:07] scaling freedom.ai. Mm-hmm. You're also on LinkedIn, uh, which is linkedin.com, you know, whatever back slash and your name on there is Jason j yay.

[00:37:17] Which is YEH. So Jason, J-Y-E-H-I can go look up on there. Anywhere else that you're like particularly active online that you want people to know about or look you up on? 

[00:37:28] Yeah. LinkedIn's the best place I am. Uh, I, I write, uh, about two to three times a week there. Um, so I'm just sharing lessons learned and case studies and, um, I I, I'm, I'm taking more of a leading learner approach Yeah.

[00:37:45] To, to all this because the space is evolving so quickly. So if you like to follow along, that's the best place 

[00:37:50] I, and I love that that's the best type of content to read. Because it's like, oh, you know, people are just like, I'm an AI expert. And it's like I'm 19 and a half years old, I've [00:38:00] been doing this for three months.

[00:38:01] And it's like, okay, alright. We're all learning this together. Like, uh, and I, I read some of your posts, you share really fun stuff. You shared a cool thing on actually about perplexity and like taking pictures of, uh, parking sucks, which really, and I wrote down 'cause I was like, that's sick. So if you want to go check out that post of many more Jason j Yay on LinkedIn or scaling freedom.ai.

[00:38:23] Jason yay. Welcome or not welcome. Thank you for being here, my friend. I, I cannot tell you how much I appreciate it. This was great. Thank you so much, Stan. That was fun.