The AI Trap: Why Buying More Tech Won’t Fix Your Broken Business

It turns out ChatGPT hallucinated the entire thing. If you don't actively use AI, you're gonna fall behind. It's not gonna replace humans. You're dealing with a very young but very powerful capability. If you can't articulate what you're doing in a way that makes sense, nothing is gonna save you.

The theory that AI is gonna replace us all, I don't see that happening. David, I really appreciate you taking the time, man. Um, this is a topic that is near and dear to my heart. I'm an armchair kind of mindset, peak performance psychology, um, and I love talking to guys that actually know what they're talking about and are, and are trained and think about this stuff every day. So I appreciate your time and looking forward to getting into it, man.

Thank you. Thank you. I, I'm, I'm excited. Yeah. So, so one of the things, um, when, when your, your people reached out and I was digging in and, uh, I, I got really excited to chat with you was, you know, you kinda live right now and, and in your career in general, but, but right now you live in this space between, you know, AI, um, and, and, and peak performance and kinda the, the, the mindset and, and neuroscience that, that sits in the middle.

And, and as we talked beforehand before we went live, like, you have, you have the platform that you're building and the technology you're building, as well as this is what you've done for a very long time, uh, for a living, as a career. You've been a consultant, a coach, and you've done this. So what I love is i- is, is kinda this isn't just you did a, a, a 201 regression analysis in an MBA class and said, "Oh, there's an opportunity in neuroscience and AI," like this is, this is your life. And what, what I kinda wanna start the equation or-- start the equation start the conversation with is a conversation around, um, like, like what's, what's real in this space, being that you have your hands in AI, but you've also done the work offline. Like, if someone's-- if someone is, is sitting here going, "You know what?

I'm a little nervous to hire a coach and go all the way to hire someone, but I'm not really sure how much AI can actually help me or, or how real it is or, or what kind of real insights can I get?" Like, maybe we just level set the playing field for, like, how much... Like, like what is real today and what isn't in terms of what someone can extract from, say, an AI tool in general? Like, can they trust these things? Can they dig in?

Or is there still human in the loop necessary? Just I'll leave that kinda there and let you roll. As you ask that question, I'm hearing two very different questions, and let me just start and k- tell you what I think I heard because- Yes ... those questions lead to different paths. Yes. Uh, what we do at Optivs is we are building, um, the capability to make AI aware of your physiology.

And so our, our technology allows the AI to know are you paying attention? What's your cognitive workload? Kinda are you in the zone or are you in kind of a tilt state? With the theory that allowing the AI to become physiologically aware will make the AI more effective at working with you. And so one way of interpreting your question is how real is the stuff we're doing, you know, in terms of take a bunch of brain data, analyze it, and can you actually read what's in the brain in a meaningful way?

Can you use it to make the AI better? And I heard a completely separate question, which is kinda how real is AI generally? Which is like forget the brain data, but, like, if I go onto ChatGPT or Claude, kinda, you know, can I use it? Is it, is it useful? And that's, like, independent of us.

And I'm happy to talk to either of those things. Kinda my expertise is in the, the first one, but you know, I, I live right in the middle of the AI world, so you know, I'm happy to take either one or both. Why don't we start at the higher level and work our way down to your specific, uh, product and what you're doing? You know, my company, we're always in fundraising mode, right? And so, um, one of my existing investors had a friend, called him up and said, "Hey, you should look at this company, Optivs, and think about doing an investment," right?

And the guy went into ChatGPT, and he was like, "You know, look, here's what Optivs is doing. Tell me if hyperscalers are doing the same thing." Like, you know, "Give me a sense, and, you know, tell me what Google is doing," right? And it gave him a PowerPoint presentation and said, "Okay, this is what Google is doing," you know, which Optivs is doing. And the PowerPoint presentation said Google is building kind of the same thing.

They're further behind, but they're kinda doing the same thing as Optivs. He-- So he's like, "You know what? I'm not interested 'cause, you know, Google's gonna eat these guys," right? And so, you know, and he sent us the PowerPoint presentation. And you know, we freaked out a little bit 'cause we were like, we didn't know about this.

And so I, I spent, like, last night doing a bunch of research. It turns out ChatGPT hallucinated the entire thing. It created a PowerPoint deck for this guy saying this is what Google is doing. And on very deep research, he prompted it in such a way so it was just, like, trying to please him, and it created a picture which is not at all what Google is doing. And in fact, you know, for a variety of reasons, our belief is they're gonna be very interested in what we're doing in a year or two.

But it was a real... You know, and it was an amazing thing, and it sort of speaks to You know, I mean, this guy made an investment decision based on ChatGPT making shit up completely, right? And so it is... You know, the problem that we're all having with AI is like when it works, it's insane, and it sounds really confident. But every once in a while, it's either completely idiotic or it just makes stuff up completely.

And so, I mean, at a high level, I and everybody I know spends a lot of their time using Claude, using OpenAI, and everybody I know gets value out of it. Like, you know, if you're a coder at this point, either Claude is doing all of your coding or it's doing eighty percent of your coding. And if-- and with that, like you go to bed at night, you tell it what to do, you wake up in the next morning, and then like it just did three months of coding. But at the same time, you gotta know it makes a lot of mistakes, and it's not, you know... So, you know, so it's like a...

It's like when you hire an intern, and they don't know what they're doing. Like, they'll do a lot of work, and they'll-- but you gotta check everything, and you gotta just sort of, you know, have some judgment about it. So what I would say is, yeah, AI is insanely real. If you don't actively use AI, you're gonna fall behind. Like, and it, and the problem is every month it's getting better.

And, you know, y- it's essentially a force multiplier. But at the same time, at least today, and I think, you know, for the foreseeable future, it's not gonna replace humans, and you need to give it a really short leash 'cause it'll make stuff up. It'll lie to you. It'll, you know, it, you know, it'll, it'll think it's doing the right thing, and it does the wrong thing. And so, I mean, that's kinda my answer, but I, I, I don't think anyone could survive without it because where it is good, it is so staggeringly powerful.

I completely agree. I, I... And that story is scary. And unfortunately, I deal with a decent number of early-stage companies, um, both helping them get investments, and I do my own angel investing. And I have seen multiple times that scenario that you just described.

A company's looking for investment. They send over their, their thesis or their SIM or whatever they're using over to someone. They throw it into ChatGPT, run a quote-unquote, "competitive analysis," and now all of a sudden, these investment decisions are being made. And, and not just investment decisions. This is, you know, all kinds of things, whether to use the tool or not.

I, I see this all the time. We were talking before we went live. Um, I tend to work in the, uh, property casualty insurance industry a lot. That's kinda what I would say, like my home industry is. And a lot of in...

They call them insurtechs, a lot of insurtech startups, um, they get... They're, they're getting killed right now on u- by users who think they can either quickly code something or just hack a solution that they have created in an enterprise tool in inside their Claude or their ChatGPT. And one, what I think is really interesting is the lack of awareness of cost, right? Like they're using the twenty dollar basic min version, you know, the, the first step up from free. Um, and there's no guardrails.

They're not using skills. They don't have plugin. You know, so there's no sophistication to the tool. Um, they're not even using custom projects, uh, which may have a little more context, uh, and a little more guardrail. And they're, and they're using this and, and making decisions.

And I, I find it to be... I, I understand why they're doing it, especially if they're maybe tend to be more of a Luddite or just, just not as technologically advanced or haven't spent as much time with it. But the scary part is they're taking it as truth, and that is crazy to me. And even simple things like not understanding how a context window works can take your entire conversation with a chat prompt way off the rails if you're not, you know, opening up new sessions or compacting the context and that kind of stuff. And, and I don't mean to get too technical, guys, if, if you don't know what all these terms mean, but you know, what David just described is a very real scenario.

And, uh, I'm interested in your take on maybe how you've thought about fixing that issue. I built myself just a little application where, um, it's actually a skill where, let's say Claude gives me an answer to something. Um, if I want, I can go to my committee, and the committee skill that I built will then, will then reach out to Groq, uh, ChatGPT, and I think I added DeepSeek, um, just for like a open source model. And it essentially creates a committee. And like, and I'll-- the-- it'll s- it'll say, it, you know, act as a critic of this opinion and poke holes in the argument, you know, and it'll hit this committee of models, and then that tends to, to, to get me closer to truth and, and certainly gets rid of, uh, some of the sycophantism.

But I'm wondering, like, have you thought about a solution to this, or how do you make sure you're not just, uh, uh, riffing off of data that that's been hallucinated? The industry writ large, you know, the Googles and OpenAIs and, and the like, are very focused on this. And there are, you know, there's, I mean, really tens of billions of dollars going into trying to figure out how to solve this set of issues. And it's, um, you know, it's a combination problem where hallucinations happen still today, like the story I told you. Um, but the other thing is You know, there's just this huge issue that AIs make assumptions that are just wrong.

You know, they're, um... You know, and that's not to talk about the sycophantic kind of nature of it, although, you know, I think you can actually work with that with the prompts, the, the sycophantic issue. Um, I mean, I guess what I would say, Ryan, is as a, is a general principle, there's two things. First of all, you can't be lazy. Like, y- you know, the bottom line is you're dealing with a very young but very powerful capability, and if you take it at face value, you shouldn't be in business, is the bottom line.

But if you also become a Luddite and you're like, "You know what? It's too complicated. I don't trust them. I'm gonna just do it the old way," there are some industries where that is gonna work, but more and more and, like, month by month, you're gonna get rolled over by anyone who's using AI because it is so intensely powerful. So, like, the bottom line is it's just kinda you've gotta, you've gotta deal with it where you're aware of its limitations and you manage these things.

And there's not, there's not, like, a silver bullet. There's not like, oh, open this window or do this prompt. Y- you, you have to be a power user of, of AI. I mean, as the bottom line, it depends on your whatever, but you... You know, where the competitive advantage is, is if you know what you're doing with AI.

You know, it, it used... Like, in coding, they used to say whatever, you know, a good, a really good coder's worth ten or twenty times what a mediocre coder is. That's now, like, a hundred times or three hundred times or something, 'cause a really good person can train an army of agents. Now, all of a sudden they've got this whole army and, you know, a mediocre one screws it up. So it, you know, it, it's just kinda, kind of the good news and the bad news is working hard, being smart is, has always been a competitive advantage, and it's equally, if not more true today.

I mean, that's my answer. No, I, no, I, and I, I agree with you. Would you, would you say that... So I was talking to somebody the other day, and so this isn't my original idea, but I just wanted to put this by you. So essentially what this guy said was he felt like his business had missed the digital era, that he, um, he's in his late fifties and, you know, he just was skeptical of, for whatever reason, he had a very successful business.

You know, his father had done it a certain way for a long time, and when he took over, he was fully invested in that model, and he kept it rolling, and he didn't wanna be disruptive, and they were making money, and he's like, he's like, "Look, like we made money through the kind of twenty ten to twenty twenty digital era trend." He's like, "We were making money, but we didn't really invest in it. We were always kind of behind. We were always the last one to take on a tool," you know, uh, that kind of stuff. And he's like, "I'm looking at AI as a chance to say, 'Okay, I was wrong.'

You know, a- as much as we still had a business, we still did it fine." He's like, "I could've doubled or tripled what I did if I had kind of gone all in." And AI, his take was he saw AI as an opportunity to not make that mistake again, and by going kind of, quote unquote, "all in" on AI, learning it both himself and figuring out how to, how to integrate deeply into his company, he felt like, "I can now say yes, I may have missed the digital era, but I can leapfrog a lot of that by going all in in AI." Do you see that as an opportunity to almost, if someone's sitting here and feels like they may have missed the digital wave for whatever reason, that this is an opportunity for them to, to get back in and start pushing again, that this can leapfrog them back into a great position? There was a study done, uh, this is now more than a year old, so, you know, in the AI world, that's ancient news.

But MIT took a look at a bunch of companies who made an investment in AI, and basically what they concluded is ninety-five percent of them lost money on the investment. That basically they put a bunch of money in, they built AI, and it didn't actually add value. All it did is added overhead to the companies. But five percent of them, it was incredibly valuable for. I don't know what that number would be today if you did the study, but I will tell you, you know, it's, there's a really...

Now, and look, I deal with much larger companies than, than kind of the people who you've... My client- Yep ... base is more people with whatever, twenty employees. Yeah. Like twenty employees. You're dealing with big enterprises.

Yep. I gotcha. Big e- big enterprises, you know, but the theme is the same. Right now, if you are not Microsoft or Google or, and, you know, Anthropic, but you know you're, I won't name a company, but just, like, a big company that's not a kind of an AI native company. Right now, your board has been screaming at you for six months about what are you gonna do in AI, 'cause everybody's freaked out about it, and there is this sense like, "If we don't get on this bandwagon, we're gonna lose out."

And It's kind of a weird phenomenon bec- and, uh, you know, this was true in the days when the internet was starting as well, where it's incredibly easy if you wanna pacify your board to go hire whatever, a big consulting firm or go hire a bunch of AI people and do a big project in AI. And as a general rule, kind of the way, the way stuff works is you're gonna wind up just wasting all your money, and it's not gonna do anything. So you're gonna sort of discover we got a way of working, it works, this is just gonna mess us up. And, you know, but at the same time, the boards are right, and it's true, and if you don't sort of get on the bandwagon, you're screwed. And, you know, I guess what I would say is, um, y- you really, you really have to be thoughtful about what you're doing.

So you, you got a company, your dad had your company, you have the company, it's working, it's making money, you got your processes. Um, you've really gotta go in very rigorously and try to figure out what allows us to kinda move to the next level. And there is, there's an interesting phenomenon in Silicon Valley where they talk about whether a company is AI native or not. Now, my company, by virtue of when we were born, is AI native, right? So w- we, we live and breathe AI.

You know, we have some of the better AI scientists in the world. We have this, like, insane competitive advantage over a company that's, like, five years older, where they did it one way, and now they've gotta kind of retool themselves. And it's... I mean, it's just, like, really hard. I mean, even doing a technological redo ten years ago was hard, but AI is, like, super hard.

Like, um, 'cause the, the field is evolving really fast, and it's, it's technically difficult. So what I would say is... I agree with your friend. It is my prediction that kind of when the dust settles three years from now and we look back, uh, you know, a lot of companies are gonna go out of business 'cause they're not doing AI, but there's gonna even, you know, there's gonna be even more companies who spend the next three years investing in AI and get nothing of value out of it. So it's, like, really...

It's kinda like the, the world has shifted so profoundly because of AI. You don't wanna be hasty, and you really wanna go in with a theory of the case about, am I an AI company? Am I not an AI company? If I'm gonna be an AI company, what does that actually mean? And you rethink the business from the ground up as an AI native company.

But I think if you just, like, plop AI on top of it and, like, okay, I'm gonna replace customer service with an AI agent, right? Or, you know... I, I mean, remember, the other thing is, you know, for people listening to your podcast, they don't have twenty thousand employees, right? And they don't have, you know, okay, I can throw a couple billion dollars at this. You know, and it is an unfortunate-- It is a game where, you know, the, the stakes are pretty high.

And so I think you just gotta be thoughtful. But at the same time, like, there's these insane stories of, like, two guys in a garage build a billion-dollar business in a year. So, you know, doable if you're smart and if you're thoughtful and if you do it right. So again, it's just like, you don't wanna be impulsive. You really wanna have a theory of the case and say, "This is how I'm gonna use AI.

This is why I believe it's gonna work. This is why it's doable. Here's a plan." And you don't spend a dollar until that plan is super clear. Yeah.

So, okay, two, two things I wanna say there. One, um, and I, I'm gonna fix the way that I ask them 'cause it leads into the next question. So those cases that you just broke down where large enterprises are implementing AI and putting millions, if not tens of millions of dollars behind these projects and then, you know, six months later popping their head up and going, "Hey, we, we're not really seeing any improvement," that to me is like the old adage of, like, the middle manager who says, "Let's buy Salesforce." Like, that person's never been fired, right? Because if you just recommend...

You know, if you just take the broad thing and say, "Hey, let's do... You know what's gonna fix our sales process? It's not our scripts or our leads or our flow or even my leadership, it's 'cause we don't have Salesforce, and all we have to do is get Salesforce, and everything will be fine." It seems like, like that mentality is being applied to AI, and that just, to your point, doesn't work, right? 'Cause you, 'cause you have to be thoughtful. We need to have a plan.

So what I have learned, um, through, through my own dabbling, playing, and, and I've built a few things as well, uh, most of them are just personal tools that I use myself, is the planning process. And the, and many of the, um, like, the s- whether it's the s- the CLI function or whatever you're using, like, there's a planning mode to these tools now, and almost no one that I come in contact with uses it. Yet when you use this, and I'm, I'm-- You don't have to use the planning mode. I wanna talk about planning in general. But, like, when you use that planning mode, what you get as an output is exponentially better.

It's not even close. You can't even compare the results to a general prompt and planning first. So, so with that said, how do you, you know, kind of maybe taking it into your own work, your own business, like, and this idea of how important planning is Like, how are you thinking about planning? What does planning actually mean before you deliver it to an AI to execute on a task? Like, do you have thoughts, ideas, or a system around creating a plan that then you can give or just some high-level ideas? 'Cause David, I don't think anybody is doing this, or at least most of the users are not planning before they use these tools.

I'm kind of a bad person to ask for this because, again, we're an AI native company. Yeah. Yeah. And so, like, I don't think we would know how to do it without planning. Like, it's just like it's so in our blood, you know, 'cause we're an AI company, right?

So... And everybody's, is there and so... A-and, you know, look, as a result, we've, you know, we've got a competitive edge, uh, 'cause it's just like, right, we live, breathe, and speak the language. I think... You know, so again, like I'm not a great person 'cause I haven't even dealt with-- I, I've dealt with big companies who are idiots in terms of what they're doing around AI, and I've seen what they've done wrong.

Um, you know, but I haven't kinda seen the specific use case. What I, what I would say is this: There's nothing that replaces common sense and kinda strategic thinking. And so I've had the great privilege in my life of knowing a bunch of people who've made a bunch of money building businesses, right? So I'm, I'm an entrepreneur, you know. I've got a ton of friends, and I've got lots of friends who've built, you know, multi-billion dollar businesses, like from the ground.

You know, and I think one thing which characterizes successful entrepreneurs is they think long and hard and, you know, business ends up being a pretty common sense kind of thing. So, you know, independent of the tools or how do you write the prompt or what agents do you use, kinda if you can't articulate what you're doing in a way that makes sense, kinda using a pencil and a piece of paper, um, nothing is gonna save you, right? And, you know, so if you've got a business and you're like, "Okay, I've got this business, it's a hundred years old, I'm gonna retool it," you gotta have a theory. Like, and, you know, it's like, okay, so we have these business processes. Am I gonna use AI to fix...

You know, let's say I've got eighteen steps in my business process. Am I gonna use AI to say, these three steps are gonna get more efficient, or am I gonna say, I'm throwing out all eighteen steps and we're starting from scratch, right? And either way, it's like you gotta be able to tell a story. So I, you know... It's very interesting, um...

You know, again, I'm, I'm just sort of random, but it's a kind of a probably a useful metaphor. Um, so we've done a lot of stuff in my company working with financial traders, and so I've worked with a bunch of big trading firms, right? And, you know, nowadays with AI and machine learning, you know, a lot of trading comes from you build a computer system, and whether it's an AI or a machine learning model, you build a really complicated black box model, and then you throw it into the market and have it trade, right? And as a result of that, I also know a bunch of people who invest in these trading firms, right? And you know, there's some very famous firms.

Everyone I know who's made money investing in trading firms has told me the same thing, which is like, if they can't explain their strategy in a way I understand it in like less than ninety seconds, I'm going ghost. Like they, they're just like, you know... I mean, Warren Buffett actually had this great line where he said, "Beware of geeks bearing formulas," right? I mean, it, it just kinda... And the reason I say that is it all comes down to just like fucking common sense.

It's like, if you wanna use AI, you better explain what you're doing and it's gotta make sense. And if you can't explain it and if it doesn't make sense, you're, you're gonna fail, right? You, you have to have a theory of the case, you know. And, and the minute you go into AI and you have it planning, you're effectively doing that geeks with formulas, right? Now you're trusting the AI to understand your business and how to make your business more effective, and it's like zero chance that'll work.

Zero, right? But if you're like, "Look, I'm gonna replace customer service with AI and this is why, and here's how I'm gonna do it, and these are the tools I'm gonna use, and it currently costs me this, and I'm gonna drop my cost to this," or, "I'm gonna do sales this way and I'm gonna enable it and now, you know, my return on investment is gonna go from here to here, and this is why," that's the kind of like, okay, makes sense. And so I mean, I think, you know, again, kinda so much of business to be successful is just kinda stupid simple. And the more AI comes along, the more important that becomes, 'cause it's just so easy to get caught in, you know, oh my God, this is incredible. And it just...

It is, but it's so easy to get, like, misuse. I'm, I'm gonna tell you just one more thing. I was, I was just at an AI conference a month ago, and there was a slide that was put up, um, by a guy from Anthropic, um- And it was really interesting. It showed the number of new apps that have been released year by year, and then it showed, like, app sales year by year and kinda the number of successful apps. And what, what the slide showed, and it was really staggering, is the number of new apps released, like, basically per month has been growing exponentially. 'Cause now you can go to Claude and you say, "Okay, I wanna build an app."

You go to bed, the next morning it's built an app for you. And so everybody and their brother is releasing apps. Dollar sales for apps, the number of apps that have become successful hasn't changed. So what, what you're seeing is you're seeing a world where it's just harder to build a successful app than it was a year ago. But kinda, you know, the message is if you just go to Claude and say, "Build me an app," you're not gonna succeed.

You gotta still have something which is a good user, user experience, which solves a real problem. And if you do that, yeah, sure, your costs come down, you can succeed. But it's, it's just kinda... Y- I mean, that's the theme is AI doesn't replace business judgment or common sense. So what I heard you say basically, y- you know, is the core tenets, the core ideas, the core structural id- uh, um, um, drivers of what makes a business successful or not have not changed.

And, and essentially what you're saying, and correct me if I'm wrong here, is that we can't outsource the ideation and decision-making to the AI. That, that judgment, taste, these things are still incredibly important are, and may become the most important concepts that a human, human actually brings to the AI is, is the judgment and taste behind it, what, what the customer experience is, what our hook is gonna be, who we're serving, how we're serving them, et cetera. And if we're just vaguely throwing these things into an AI and hoping somehow it's gonna, like, make these decisions for us, that is just an absolute recipe for failure. I would concur with that. And it does kinda speak to, um, a much larger question about whether and when you're gonna see the rise of autonomous AI.

And I mean, of course you're gonna see it. Um, but there's a huge debate in the industry, um, you know, there's a lot of debates. Like, people fight about everything there. It's like, you know, is AI conscious and are we gonna reach AGI if we've already reached it, right? But one of the things, you know, people are really trying to figure out is, are we building an AI that's autonomous or, you know, is the model for the future more AI teaming with humans where, you know, there's always gonna be a person in the loop and kinda you need that.

And it's actually... I mean, that particular question is worth hundreds of billions of dollars to kinda figure out the answer to. And, you know, there's a... You know, there are people with very strong views on both sides of that equation. And of course, obviously you're already seeing autonomous AI, right?

So, you know, you're seeing it with missiles, and you're seeing it with cars. Mm-hmm. And, like, there's, there's plenty of situations where, you know, the, the AI is effectively just operating on its own. Um, you know, I think I don't have any special wisdom or knowledge on this, so I'm just speaking as, as, like, a guy. But based on my experience, based on what I know about neuroscience, based on what I've seen, I think that the, the theory that AI is gonna replace us all is pr- Yeah.

I'm very skeptical about that. You know, the whole- Yeah ... five years from now we're gonna wipe out 70% of all the jobs and everyone's gonna be unemployed and working for the AI overlord. I mean, I just, um... I don't see that happening. Um...

Yeah. That doesn't mean I'm right. But, um, you know, as of today, AI has added jobs to the economy. And sure, it's replaced a few things here and there, but as of today... And you know, the truth is there have been many, many technological advances over the last 200 years where every time it happens, people are like, "It's gonna eliminate people."

And so far, you know, it'd be very hard to point to a technological advance which has led to lower employment rates. I'm with you. I, I, I'm huge, uh, techno and AI optimist in general, while, uh, optimist for the human, the, the, for the future of humans and their relationship to AI, huge, huge optimist. I guess I have to say that. You could be, I guess, a, an anti-human optimist and just think the AI was gonna take over, and that's how you're optimistic.

But I guess very optimistic for human in the loop and humans in general. And I think a lot of the conversation around, you know, the, the, the doomer conversation around AI, uh, all of those arguments, when you boil them down, uh, they tend to misrepresent one primary piece of information, in my opinion. And the in- the, uh, kind of, I guess, analogy that I would make is it would be like saying the Industrial Revolution wiped out all the farmers without telling you that all of those farmers became factory workers. They didn't lose jobs. They still had jobs.

The job just wasn't the same. So yes, we didn't need as many farmers because we had tools and machinery to help them do things that used to take humans. But whether it was running those tools or moving into the factories that came along with that technological innovation, those people were not dying in the farm fields because they didn't have jobs and couldn't pay for things. They just, it just transitioned to where they work and how they work. And I-- in every one of these technological advancements in which someone, you know, the, the, the, uh, i-in this, most of the time it's, it's the incumbents with territory to lose who make these arguments.

They never reference where the jobs went. They just reference the jobs that were lost as if that happened, uh, on an island. And, and that's the part that I find maybe not purposely disingenuous, but certainly something that we always have to consider when we read those arguments. So I'm with you. I wanna transition to Optios, and you said something in the green room that I have been dying to get to, which is y-you kinda sit in this world of hard science, AI, and actual practical performance.

And the example you said was you can use y-you know, your tool to analyze someone shooting foul shots, but at the end of the day, if they don't make more foul shots than the, than the, the AI doesn't do its job. So taking now, let's, let's spec-specifically talking about Optios and the work you're doing there, how do you marry and how do you think about and build towards, you know, all of this, you know, like you said, every day there's a new model coming, a new idea, a new way to get data out, a new way to structure speed, okay, with, uh, like what your clients want at the end is a practical outcome and improvement. Like, how do you marry those two things, and how do you make sure you're getting them? Because I think to something you said in the very beginning, there are a lot of people today that are implementing AI, using AI, building AI without any real idea of what the practical outcome should be or, or even tracking what the practical outcome is. I think a, a good starting point for that question is to tell you about a research experiment that was done fifteen years ago out of DARPA.

Uh, you know, for those of you listening, DARPA is an agency in the government started in the '50s, uh, to basically fund frontier research, which is kind of too early for corporations. And, you know, it was built out of the defense industry, and a lot of the most important inventions that have ever been made in the US came out of DARPA funding. So fifteen years ago, DARPA had a theory that you could use neuroscience to an, to improve a war fighter. And to test the theory, um, they asked just a really simple question, which is, "Can I use neuroscience to measure when someone's in an optimal brain state?" And the task they started on was marksmanship 'cause, you know, it's the military, easy to get a bunch of data.

And everybody wanted to shoot a gun back then, right? Incidentally, that is much less relevant to the military today, but back then that was, and hugely interesting. So they took several hundred marksmen, and they scanned their brains when they were shooting. And, you know, sure enough, they discovered that there is an optimal brain state associated with shooting a rifle. The expert marksmen were pretty good at getting into the state.

Novices didn't know how to get there. And it was-- And then n-now you had the ability to measure what people call, you know, the flow state or the zone state with regard to marksmanship. So that was actually a big deal for the study. But then they went on and did something really interesting. They said, "And now that we can measure the zone, can we use technology to train it to accelerate learning?"

So they invented, you know, arguably the first neurofeedback device ever. Now these things are everywhere. But it, it was basically a sweatband. It had sensors in it. It's just measuring if your brain's in the zone or not, attached to a haptic motor that clips onto the collar.

So the way it worked was if you're not in the zone, it's vibrating on your neck, and then as you get into the zone, the vibration goes away. And that's the term neurofeedback. And they had novices train with this for like total of two hours over the course of a month, and the results were just stunning. What happened is, because the brain is plastic, they rewired their brain over the month. They learned to access that expert state very quickly, and with that, these novices moved eighty percent of the way up the learning curve toward being experts.

So it saved like six months or a year of training time. And then they went out and they showed that it worked with like intermediates, and they showed it worked with experts. And so, you know, kind of the, the core thing that came out of that was this idea, if you can measure things in the brain that are relevant to performance, you can exploit that information to help accelerate learning or to help improve the performance, right? So you talk about a foul shot, and you can imagine the same thing. If you're trying to learn to, to shoot a free throw, and if you can get metrics about, "Is my body moving in the right way?

Is my brain in the right state?" You can teach yourself how to go through a pre-shot routine to get into that optimal state, and that'll improve your free sh- your free throw accuracy more quickly than just normal shooting a basket. And there have, you know, since that time, the government spent like seven and a half billion dollars doing research all around this notion of, can you measure the brain? Can you use it to improve performance? And there's, you know, hundreds of studies that have come out where, you know, you can process information faster, and you can improve your memory, and you can learn at like two hundred and fifty times the speed, and the like, right?

And so that's kind of the underlying science that's informing what my company does. Okay. You with me so far? So, you know, what, what we are doing is we're focusing on two problems. Problem one is to kinda get really good metrics physiologically that you can use in the real world, right? 'Cause all this DARPA stuff was done in a laboratory with, like, graduate students or, like, snipers or, or whatever, but, you know, it, it hasn't been deployed.

And so we've got this massive database of brain data and eye data and stuff in association with tasks. So y-you can basically say, you know, was a trade profitable or not? Did the sniper make the shot? You know, we've got stuff with basketball players and football players and traders and pilots and the like. And it's all about kind of building an AI kind of model to say, "Now I can do real-time measurement in the brain."

Right? And, and to do it in a way where, like, millions of people could use it. And then the second thing is to actually build a closed-loop system where you actually use the data to have an impact on making someone learn faster, you know, or improve performance or, or even to just tell an AI, "Here's the state of the human, so here's how to interact with them better." But some kinda human-in-the-loop system where we're kinda really providing that human state later, that physiology data to make the AI better. Okay?

Makes se-- So you're with me so far? Right. And, you know, so I guess what I would say is, um, I think the evidence is beyond dispositive that this can make people better. I mean, it's just kinda like if you give someone information about their brain state or you give an AI information about the brain state, it will improve performance twenty percent, thirty percent, three hundred percent, you know, depending on the use case. But, like, every time we've tried this, every time somebody else has tried it, it kinda works, right?

So... And it... You know, if you think about it, it makes sense, right? You manage what you measure. So if, you know, if you're trying to learn to speak Spanish and you've got an AI agent and now that agent knows, are you paying attention?

What's your cognitive workload? You know, is the information getting in there? And it modifies what it's doing, it'll double your learning speed, right? Like, it's, it's just gonna happen. And, you know, we've done stuff in golf where we showed you could improve putting accuracy by, like, thirty percent and stuff with, you know, pilots where you can improve performance at a simulated flight task by, you know, somewhere between thirty and seventy percent, and stuff with the National Geospatial-Intelligence Agency, where you actually got a tripling of productivity.

So kinda the, the science I think is very real. Um, to say something is real and works in a lab is super different than you got ten million users and they're using it, and now you'll never learn Spanish unless you're kinda wearing a headset or monitoring your eyes 'cause it makes it better. But I, I think kinda the science is compelling enough and the problem set is compelling enough. It's very hard for me to imagine that three years from now or five years from now, you're not gonna see this kind of toolkit built into all these AI agents, or at least, you know, certainly the ones for education and gaming and, you know, sports and stuff. And, and we're already seeing kind of a big movement in that, and there's a lot of Silicon Valley m-money kind of supporting the thesis I just gave you.

Yeah, it makes sense that I think intrinsically we all understand that if you train your body, you do it in a deliberate way, you watch and listen for feedback and iterate off of that feedback towards positive performance, that your body, you know, physically starts to respond. And what I hear you saying is now with, with AI and the ability to scan the brain, we can do the same exact thing with our brain. We can understand the mechanisms, the states, the processes that need to happen in order to improve our functionality from not just a physical perspective, but how our, our mental state, our mindset, our focus, et cetera, also improves our performance on a task. And not just in-- so in the case of like trading, right? Not just hitting a baseball or taking a foul shot or, or a putt, right?

It's actually, are you saying like even the decision-making that we're making on, say, like a trading floor if we're, you know, trading stocks or something? We've done three studies where we showed if you take a day trader and you put a headset on him, you can predict in advance if the trade's gonna make money based on whether he's in a good state of mind. We did a project with a professional baseball team. You look at someone's brain before they step into the batter's box, like seventy-five percent accuracy in predicting the outcome of a swing just based on their brain state in advance. Like it's, it's real and...

Oh, yeah, no, this is, this is real, published, like, not, not the big... The, the trading thing and the baseball thing aren't published, but no, they're, they're solid. Um, you know, and if you, if you think about it, kind of the world of AI is transformational for this problem set. 'Cause what, what's AI about? AI is about training really large models where you're given a massive amount of data that's too big for a human to comprehend, and you say, "Build a model," right? So, you know, now you've got self-driving cars.

Those were trained with just gobs of information, where now it's like, oh, okay, that's a stop sign. That's a puddle. That's another car. You know, that's someone crossing the street. You can do the same thing with brain data, right?

So we've got this, you know, there's like-- we, we get like a hundred million data points per hour out of the brain when we put sensors on them. We've got data from the heart and from the eyes. You know, you take that and you build a large enough dataset, and you're like, "This is what it looks like when you made a free throw or you didn't, or you sank a golf ball, or you did a training, uh, decision well." Over time, the model is like, "Oh, okay, now I know what the brain looks like when you do well or poorly." And what's interesting is, because it's AI, when you give it information about basketball, that informs how it looks at a trader, and when you give it information from a trader, that informs how it works with golf.

It's all about large datasets, right, and diverse datasets. Um, so yeah, I mean, it, it's real. I mean, there's, the, it is... The, the results are staggering and, you know, I, I think it, it feels utterly inevitable that it's just gonna be part of the AI ecosystem. That, you know, if you're a football player and you're watching film, the AI is gonna track your eyes and track your brain movement, and it's gonna be like, "Hey, are you paying attention?

Hey, you know, you were supposed to look at this coverage. Did your eyes look at it? Did your brain register that?" And that'll be part of the film-watching experience. Like, y-you know, it's like, it's just like one out of eight million examples.

You know, if you're playing a video game, the developers are gonna wanna know what your brain is so that they can create the game to make it meet what you're... I mean, it's just gonna be part of the AI tech stack, I think, inevitably. Do you think there will be a wearable, um, that maybe isn't like a, a whole brain scan hat and, you know, all the devices on you? W-will there come a day where, say, I'm just a s- I, I don't wanna say just. I'm a salesman, and I have ten sales calls today, and I pop on my necklace, contact lens, whatever, right, earpiece, and it's able to help me make sure I have my mind in the right state that it needs to be in in order to be successful on that sales call or at least position mys- you know, like, like that kind of practical everyday use.

Do you see that as like a wearable that we have and getting real feedback from? Sort of. Like I, I mean, here's how, here's what I think is gonna happen. If you're a, you know, l-let's say you're a telesalesperson, right? So you're sitting there in front of a, a monitor and you're making whatever, a hundred calls a day, five hundred calls a day, whatever.

Um, there's gonna be a system that tracks your physiology in association with those calls, and it's gonna be built on top of data from tens of thousands of salespeople. So it's like, this is good, this is bad. Now, that is gonna need to be multimodal. So it's gonna need to look at your eyes. What's the size of your pupils?

Where are you looking? It's gonna want your brain. It's definitely gonna want sound from the audio. You can get a lot of information. So you're gonna use AI to decode the audio, not only what were the words, but what's the intonation.

And y-you gotta realize, once you get a large enough dataset, the AI will be insane. Like, the AI will know in advance of a call, are you likely to close this or not? And that'll be useful information that you can use to manage people and screen people and train them, right? But if you think about the nature of it being multimodal, you can't really do it through a wearable. You're not gonna have...

Like, it's not gonna be like an Oura Ring or a Fitbit where now you wear it. I think it's gonna be integrated into the overall system through a number of sensors which are interchangeable. So I don't think you're looking at a hardware solution. I think you're looking at a software solution coupled with, you know, a commoditized set of hardware devices, right? And so, you know, you already have data from microphones, so you're gonna be able to pull that.

I think there's gonna be much better cameras, 'cause right now you can't pick that much up, so you're gonna... I mean, there'll be much better cameras where you can track the eye movement and look at the eye. I do think there's probably gonna be sensors in the headphones. Remember, these guys are already wearing headphones, right? So what you'll do is we'll throw a couple sensors into it.

Now you can read the brain data, and it's gonna be part of the headset. My guess is everyone who makes these headphones five years from now is gonna be putting sensors in, 'cause it'll just be kinda... And then there'll be some sort of software platform, which I hope comes from Oprios, which integrates it and then feeds that information as a model context protocol into the sensors. Like, I think that's almost certainly where the future is headed. Um, you know, what's gonna be interesting, which is the bigger issue, is how many of those salespeople are gonna be replaced by AI, right?

Um, that's a bigger question. But for the people who are still on the phone, I don't see any way that's not happening, and it's gonna be in the next, whatever, two to five years. You know, and this is, this is one of the places where I think being a Luddite is gonna hurt you. Embracing that type of technology, embracing understanding like, hey, let's say Johnny's your number one salesperson and he shows up in the morning and every sensor attached to him is, is signaling that he's stressed out, overworked, something's on his brain, he's not in a great place. You know, you can cut him off from making maybe his first 25 sales calls and maybe sit him down or just tell him to take a break or have a meet, you know, something to help him recalibrate before he wastes the first three hours of his day banging on calls that are never gonna be successful because your sensor data is telling you he's gonna be short on the phone because he's just, you know, didn't get enough sleep or whatever's going on.

You know, you're, you're pulling that, um, and the ability to manage in that way. You know, one of the things I think is really interesting, and this is where I'd like to, to close out our conversation today, is just we've talked a lot about planning, talked about feedback, you know, the massive amount of data that's gonna be both at our fingertips as a whole, as well as, um, uh, uh, leveraged by AI tools that can kinda synthesize it and produce outcomes. My position is, and this is what I'd love your, your take on, is I think AI moves more of the burden of success off of the mainline production, uh, producer of the value, say a salesperson, a customer service person, and puts more of the burden on leadership today. And the reason I say that, and I'll finish up this idea and then I'm very interested in your take, is that because I now have insights into, let's say it's two years from now, Johnny's mindset when he first shows up at work, and I can kinda jump in and sit him down or maybe just grab a cup of coffee with him for ten minutes and try to help him reset his brain and get him into that right state or, you know, uh, uh, uh, grab a customer service person who's maybe had two or three really tough calls in a row and, you know, whatever, whatever.

I, I believe this, you can't check out as a leader anymore, right? There's, there's no excuse in either terms of delay of data, lack of data, lack of insights, et cetera, lack of ability to train. Like, it feels like more and more responsibility and more and more of the burden of success is moving to the, to the leadership layer and, and it, it's now paramount that we have high-quality, thoughtful leaders versus the mainline, right? Where, where a great mainline person could make up for a poor leader. I think today you, you have to have that high-quality leader or your, your boots-on-the-ground people are gonna really struggle to be successful.

Does that, does that make sense? Is it-- Does that argument make sense? I agree completely. I mean, you, you... The problem I'm having is I don't know what to say other than I agree.

Like, I agree. I think, I think your analysis is correct. So if I say anything beyond this, all I'm gonna do is just r- repeat what you just said. Um, I mean- Well, that's okay. That's okay ...

I, I, I mean, it goes back to what I said before about the difference between a really good coder and a good coder has been magnified ba- by AI. That's true everywhere, right? I mean, you can now do so much more with so much less because of AI, and what that means is leadership and strategy become more of a differentiator to businesses than they were five years ago. And it's always been the key differentiator, but I just... I agree.

It's more so, it's more so today. And, you know, look, I just think you're gonna see more and more division between winners and losers than ever before, 'cause if you use AI correctly, you're gonna be able to eat the lunch of others. And the speed of development, like, you know, things that used to take ten years in business now take six months, and so you have to be smart and all the rules just keep changing. I mean, it's just kinda, um, if you're good and you're hardworking and you're smart, you can make much more money much more quickly. And if you're not, you're more likely to get wiped out than ever before.

So I, I kind of, I agree with you. I think that's a great place to finish our conversation. I, I couldn't agree more. I do wanna, I do, I do wanna say one more thing. Yeah, please.

Please, keep going. I wanna go back to the thing I said at the beginning. So let's talk about this guy who runs the sales center, and I just wanna kind of temper what I just said. I think if you're putting your head in the sand, you're not investing in technology, you're gonna die. But I think if you approach it, you know, like with the energy of a five-year-old kid who just walked into a candy store, you're gonna probably kill your business as well.

And so today I, you know... Th- there's a lot of limitations to AI, right? There, there just are, and you... In many ways, the old ways are still better than the new ways, and it's like you, what you were saying about Salesforce, right? And so it's not a band-aid, it's not a silver bullet, and, and so I think it's important to be very aggressive with technology but also very conservative and skeptical.

And so, you know, it's interesting, um- I do a lot of stuff in sports, right? And sports is, you know, it's really famous for having kinda older coaches who believe they know what they're doing and resist change, right? And if you remember, like, the movie Moneyball, again, right, it's all about this like we have this new technology that's gonna change everything, and you remember it's all about, like, this old guard, you know? W- but the, the people who are successful, like the great coaches today embody what I'm talking about. They look really hard at technology.

You know, I mean, I, I know a university, um, basketball team that did incredibly well recently, and their coach is very tech-forward and does a lot of things that are just not standard in basketball. You know, using w- force plates and how high someone jumps as a screening tool to decide who you're gonna put into the game kinda thing. And he's got 30 things like that. You know, cameras everywhere using AI and so on. But, like, what characterizes this guy is he's very tech-forward, but he's also very skeptical, very commonsensical, and is really willing to call bullshit if he's not convinced that it's giving him an edge.

So the reason I want... You know, the reason I wanted to close with that is I, I, I think you have to keep both things in your mind. You do have to be very aggressive. You know, it's idiotic to pretend AI's not gonna change the world, but you also don't wanna be impulsive, and you never wanna think the technology is gonna solve the problem that you have to deal with in just like... You gotta, you gotta think hard, and you gotta be commonsensical.

And I think if you can manage that balance, you know, it's the best time to be an entrepreneur ever in the history of the universe. Ah, I love that. And I'm, I'm glad-- I'm actually glad you came back around. I think that really... I think that is one of the more important ideas that we have discussed, and I'm very glad that you did that.

I know my audience is gonna wanna go deeper in your world. Where can they learn more, and is there any socials that they can follow and, and kind of hear your voice and what you're trying to do? It's so funny. I, I'm now doing podcasts, and, like, everybody asks me for a plug-in. Like, we just suck at social media.

So I've got... Like, every once in a while, we post things on LinkedIn, but I like- Sure ... you know, basically w- it's not easy to follow us. We don't have, like, a newsletter or an Instagram or anything like that. So, um, you know, maybe, maybe you guys- Well, we'll send them to the website, and the LinkedIn is a good place to start ... we have a website, yeah. But there's not even any- Yeah ... social links on the website, so, um- Awesome ... but, um, thank you so much.

It was a great pleasure talking with you. Yeah, no. And I, and I love your perspective. I really do. I think, I think there is a nuance to how you are talking about this stuff that, that is, that is the vein.

And, um, I think it's the exact right way to be thinking about it, and I'm so glad that you shared it. Appreciate you. Um, you know, anytime you wanna come back on, I would love to have you on 'cause I love this topic, and I feel like we just started scratching the surface. So I appreciate you, David. Thank you so much.

Thank you so much.