The Koerner Office - Business Ideas and Deep Dives with Chris Koerner - How to Piggyback Your Way to Success: AI, Med Spas, Golf Courses and More.⏐Holdco Bros Discuss! Ep. #134

Episode Date: February 21, 2025

HoldCo Bros are back! In this episode, Nik and I kick things off with a funny story about going to Raisin Canes while wearing a "Think You Should Leave" t-shirt. Then we talk about the booming Med Spa market, which is really driven by millennials who are both vain and obsessed with their health. Nik and I then discuss how there's potential in integrating Med Spas with golf courses and other membership-based businesses. I share my idea of starting a sauna/cold plunge business inside a gym or country club. Then we shift gears to AI agents, and discuss how to leverage them for market research and identifying new opportunities, and also how to extract insights from interviews. We wrap things up by talking about OpenAI's new tools, which are GPT Operator and Deep Research, weighing both the pros and the cons for different business applications.Learn more about Nik here: http://linktr.ee/cofoundersnikTimestamps below. Enjoy!---Watch this on YouTube instead here: tkopod.co/p-ytAsk me a question on or off the show here: http://tkopod.co/p-askLearn more about me: http://tkopod.co/p-cjkLearn about my company: http://tkopod.co/p-cofFollow me on Twitter here: http://tkopod.co/p-xFree weekly business ideas newsletter: http://tkopod.co/p-nlShare this podcast: http://tkopod.co/p-allScrape small business data: http://tkopod.co/p-os---00:00 Highlights00:33 Family Outing at Raisin Cane's01:07 The Iconic '55 Burgers' Shirt03:11 Med Spas and Their Growing Popularity04:50 Med Spas in Golf Courses: A New Trend?06:25 Revenue Models and Business Strategies10:24 Cold Plunge and Sauna Business Ideas16:17 Leveraging AI for Business Success23:40 Practical Tips for Entrepreneurs26:54 Unveiling Deep Research Capabilities28:16 Exploring Chris Crone's Profile with Deep Research29:26 Deep Research: Asking the Right Questions30:46 Maximizing Deep Research on a Budget31:58 Introducing GPT Operator and Deep Research33:36 Impact on Industry Reports and Market Analysis36:07 Challenges and Limitations of GPT Operator39:20 The Future of AI and Self-Driving Technology43:57 Deep Research: Pros and Cons47:17 Practical Applications and Early Adoption

So med spas are growing at five times the average rate of growth in the United States economy. Those pools cost 130 grand. But of those families that have pools, how many of them have coal plunges or saunas? They're growing like crazy. They have great margins, like 15 to 20 percent. Some of them can even go higher than that. So there's always going to be demand for these things that seem like a trend. By the time agents become usable, people like Chris and I are going to be way ahead of you.

So just get educated now. Bottom line is we're early. Start playing around with it. You're going to make a million bucks by being detailed and understanding every part of it. So we go to Raisin Cains as a family. Okay. Shot out Todd Graves. Raising Cain's founder billionaire.

And I'm wearing my shirt that says 55 burgers, 55 fries, 55, right, from I think you should leave. Yes. The drive-through skit. And I was just, I just happened to be wearing it that day. We weren't planning on going inside Raisin Cains, but I had my mom there. We had a bunch of people there. And I'm looking at the menu and I'm like, man, this place is expensive.

It's going to be like a hundred bucks. So I just ordered a catering order. I ordered a big tray of fries and a big. tray of chicken and it was cheaper and it was more food for feeding everyone. But then I didn't even realize I was wearing the shirt, which is like ordering way too much fast food. And everyone in the kitchen was like, oh, I love your shirt. I love your shirt. When I say everyone in the kitchen, I mean one guy. But still, it made me feel good. What are the chances I would order all this fast food. So, but I got a shout out. If they don't know what they were talking about, go just search

on Google 55 burgers, 55 prize. It's a, it's a sketch from I think you should leave. That's so good. I got another compliment from a stewardess on an airplane for that shirt. Go on. She was like 60-something years old and she liked my shirt. And she got the reference? I assume. Otherwise, she just liked a really random looking shirt.

Oh, I really like your shirt. That's what I get every time. 55 burgers, 55 fries. For those who don't know, the sketch is this. Tim Robinson goes through a drive-through. He orders. He goes to the window and he goes, hey, I want to pay for the guy behind me. I don't know

Paying it forward Maybe it'll catch on Maybe it'll catch on I'll pay it'll Just thought I'd do something nice And so And so the lady The lady takes his payment

She's like oh that's sweet And as soon as he does it Because he knows he's starting A pay it forward chain Right Like the idea behind that is the next person It's like oh they paid for me Cool I'll pay for the person behind me

Tim screeches out Pulls around Goes back to the order And he goes I want to order 55 burgers 55 fries 55 Coke

He's just like ordering a bunch of food because he knows the person in front of him is going to be paying for his order now. Let's just say hilarity ensues. Then the guy gets mad at him and then he finally starts to apologize after Tim says, I just wanted to do something nice before alcohol class. And then the guy feels bad and he starts apologizing. And then right when Tim has him, he goes, oh, I can just run. It's so good. It's so good. All right, dude, it's been a couple weeks since we talk.

I'm freaking stoked. I have two main ideas I want to talk to you about. Which one do you want to talk about first? Med spas and golf courses or the good and bad of AI agents. Let's talk about med spas because I got a joke to go along with that. So we're getting older. We're millennials. Millennials are vain, but they're also worried about their health.

And so we're seeing massive explosion in med spas. For those who don't know what med spas are, they are essentially cosmetic places for people to go and relax, but they give things like Botox, injectables. Some of them do more health-related things. So they'll do like the semi-glutide treatments, right? Everyone's heard of Ozempic. That's what that's what those are. Hormone replacement therapies. Some do testosterone replacement. But essentially it's aesthetics. It's coming in. You're aging. You want to get some fillers. You want to get some Botox. Whatever. They're taking off like crazy because our generation in particular, We're not only obsessed with health, but we're obsessed with aging.

So here's some stats. The current medspa market in the United States is about $4 billion, but it's growing at over 10% a year annually. So just for context, the U.S. total economy is probably growing at 2% to 3% a year. So med spas are growing at five times the average rate of growth in the United States economy. So they're growing like crazy. They have great margins, like 15 to 20%.

Some of them can even go higher than that. And you're doing something that's like one of my favorite frameworks, which is taking advantage of a current trend. Like you have the tailwinds pushing behind you. There's a quote that I like. A quote that I like, it's like, my favorite boat is a boat that has a tail wind. So it's like, cool. You know how to do a row boat really good?

That's great. I'm just going to go to the place where I know the wind's going to be pushing at my back. So to me, it's taking advantage of like a couple of these things that are coming into place. And so I love meds spots. A friend of mine, it has a membership at a golf course and some golf club. And a new buyer out of California bought the golf club. They're shutting it down. They're renovating it.

And they're going to be building some new condos on. They're making it a lot nicer. They're putting some amenities in. And he approached the owner and it was like, hey, what if we added a golf or a meds spa? And we gave a discount to the members. And the new owner is like, that's really interesting because it allows to give another benefit to its members. It's sticky. It's high margin.

The people who are able to afford the meds spa. or excuse me, the golf membership are probably also going to be able to afford the Mets spa. Hopefully we don't get a lot of hate mail, but probably 90% of golf memberships are men golfing. So this gives an opportunity for like men to bring their spouses to the club and give them an added benefit. So like it reduces some tension there. Like they're in talks right now of developing a med spot that would like go on the campus of the golf course with this new bill that's happening. So as I saw that, I was like, dude, that's super interesting. He's piggybacking.

Like it's what we talk about all the time. And it just happens to be on the local level. I think it would be really interesting if you were, if you were interested in doing the meds spa in looking at country club memberships, golf membership, tennis club member, wherever, and scoping out spaces near those establishments for space at a med spa. And you could test this idea a million different ways. But that's sort of the general premise that I've set up. What do you think?

Do you know what the structure is? Like, is he just going to pay rent? Is it going to be a revenue share agreement? They haven't worked it out yet. According to my friend, the owner of the golf club is like not interested in having equity. This is according to him. I can't imagine a scenario where I would do that. I would think there'd be some type of a rev share or affiliate agreement or something.

But according to him, he's like, no, he just likes the idea of there being an added benefit for his members. And so I would just come up with the money. I would do the remodeling. I would provide the physicians and the clinical. oversight. I would do all the business and give a discounted rate to the club. He would have no equity ownership in it. So according to him, that's the current model. I don't think that's normal, though. I think that these golf clubs would want a share. And frankly, that's how I would pitch it.

It's like, hey, you would get a piece of this because I'd really want them bought in to the men's spa. Yeah, I look at it as kind of like a Walmart model or an indoor shopping model where the parent leaseholder like Walmart or a shopping mall takes not a percentage of equity but a percentage of revenue percentage of your sales like when I had an iPhone repair store inside a mall we had to pay $2,200 bucks a month plus like 10% of all of our revenue and that was standard. That's normal. So to me that would make sense because if it's like a if it's a big med spa company, they probably don't have a vehicle for giving away equity in a location to some random golf course owner, but they probably do have a vehicle to giving way a percentage of revenue, right?

Yeah, that might work. The other example of this are like eye clinics, I don't know what the word to call them is, but like every Walmart has an eye center. Great clips, subway inside of Walmart and all that. Right. Costco has these as well. So it's like, it's not unprecedented. Oh, my members are this demographic.

They also like these things. Let's put that offering in there. The golf course med spa is a good pairing, but I'm sure there's tons of these types of pairings anywhere that you have memberships. So that's why I mentioned like the tennis club, country club. I'm trying to think of some other whatever membership types of companies. If you look at that, there is some other adjacent offering that that current membership isn't giving that you can take advantage of. I think working your way backwards, again, from the distribution are some cool ways to identify opportunities. Fun fact, did you know that there's a

regulation that requires optometrists to have an exterior door. And so if you find a Walmart that has an optometrist inside, you will see a random door on the front of the Walmart leading only into that optometrist. My friend Dan owns an optometry clinic inside a Walmart in Washington. Tell me that. I mean, that doesn't surprise me. Like home health and hospice, we had to have separate entrance and exit doors. So I'd like from a healthcare standpoint, I guess I could see that being the case. But no, I did not, I did not know that. So every Walmart, eye center has a separate entrance. Looks really random too when you look for it.

How do we make money off of that? I got a couple thoughts. So I like the idea of piggybacking on golf courses in general. I wonder what else you could do there. On one hand, this med spa idea is like, oh, this is really interesting because now husbands can bring their wives to the golf course

and the wives can just go to the med spa and the husband's going to golf. And on the other hand, that's not a behavior. You know what I'm saying? That's not a current behavior. So that makes me nervous. Oh, I see what you're saying.

If it's 90% men that go to a golf course and 90% women that go to a med spa, is that doomed for failure? Men are using Botox more and more as a percentage of the population. This particular golf course has other amenities for families. So there are a lot of families that come. Yes. Yeah. That's the best way to say. It's not just a golf club.

It's a country club. I don't disagree with you, right? because you're making some assumptions on behavior. But that's probably why the golf course is like, yeah, cool. If you want to open it, it'll be good for our members. But they're not necessarily taking skin in the game. But that's, yeah, it's a good. Another idea.

I had a guest on my podcast a few months ago that opened a sauna slash cold plunge business inside a gym. And it's going great. And the build out is not that much. He bought two cold plunges for like five grand each and then two sonnas for like 10 grand age and just plopped them inside this empty square footage that the gym had right when you walk in so like you walk in there's the desk you go to the right to go to use the gym and then like straight in front of you there was this empty room that they weren't using and he just rents it from the gym

and that was his first location why not like you like yeah that we heard for golf course yeah golf course is country clubs right that's the again that's the perfect population those are going to be the guys who are like yeah i got my country club membership and we have now a cold plunge in a sauna Well, just as like a framework, it could be find a facility that already has a locker room, showers, and a bathroom, and then go piggyback your cold plunge and sauna on top of that. Maybe it's a gym. Maybe it's a country club. Maybe it's a pool, like a public pool, right? Yeah.

What do we think? Dude. Okay. I'm going to take your framework. Bring another framework. Seemingly unrelated, but I think it is related. I think all of it, well, at least you and I have definitely had these experiences. experiences life where we're like, oh, I want to buy some Amazon stock. Nah, it's overvalued.

It's probably the high is ever going to be. I've been saying that for 20 years, right? Tesla stock. You could have said the same thing. Apple stock. Like, whatever the stock is, you're like, I think it's kind of hit its peak and it's only ever continued to grow. But at some point, you are right. You could be right. Sure. Sure. So there's two sides of that. I do think that there is a point where you reach market saturation and you're kind of like capped from that standpoint, but 20 years ago when we were looking at this, if we had been smart and been like, hmm, these are the user adoption trends, wow, 50% of the market's still not on the internet. Oh my gosh, what percentage of people actually order things from Amazon?

Oh, how many people have an iPhone? Like if we would have done more critical analysis, we probably would have said, yeah, this is the way that things are going and now I'm better at doing that. That's the same way I feel about some of these trends. So you might say, oh, cold plunge and saunas, those are, it's over. saturated at this point. But the truth and the reality is we're not. Yeah. If you want to go and open a cold juice smoothie place, yeah, we're probably oversaturated. Jamba Juice did that 20 years ago. But cold plunges and sanas, no. So like just finding those things that are trends that you see

becoming more and more popular and then putting them into places to give people access is an interesting framework because most of the time I would think like coal plunges and sanas are They're oversaturated, but they're not. Like, they're just not yet. Like if you look, what's that graph that's like the distribution of early adopters? Oh, yeah, yeah. But it's like early adopters, adopters. And like by the time it goes mainstream, it's still not even done growing, right?

So it's like, I don't even think coal plunges and sonas have gone mainstream. So there's an opportunity to place it. On Twitter they have, right, which is not representative. But I'm thinking of the midwit meme, bell curve, right? this is more related to like market size in general on one end it's like anecdotal data like my friend doesn't have a cold plunge so we're early right and then you go to the middle and it's like yeah but that's not an accurate sample size and then on the other end I think that's still true like you have to be experiencing something that's happening in that industry for it to be relevant or not for

instance, I go to church with a bunch of people, my similar age, literally, and this is not an exaggeration of like 50, 60 families I go to church with, 80% of them have pools, right? And maybe of those 80% more than half were since COVID. And those pools cost 130 grand. But of those families that have pools, how many of them have cold plunges or saunas? I don't think any of them. Right. So if you can afford a pool, you have a family and you're of the demographic that values the benefits that cold plunges and sanas offer, then statistically speaking, all those people I go to church with

are probably going to own a cold plunger or sauna at some point, but they don't yet. Adding on to that, I know I just crapped on Jamba Juice a second ago, but Jamba Juice has proved that there's demand for smoothies. There is probably somewhere in this world, many places in this world in the United States, where you could plop a Jamba Juice down

and it would do really well because there's some percentage of demand, right? So there's always going to be demand for these things that seem like a trend. Crumble cookies is an example, right? Crumble may plateau at some point, but there's always going to be a place that would support a Crumbull cookies franchise. So just going back to it, it's about thinking through the distribution. I love the gym idea.

I freaking love that idea of placing these saunas. I love the idea of placing these onas at tennis clubs, at fitness clubs, at country clubs, like golf courses, wherever. You just need to think through, okay, where are the people who would, use this going and how can I get them access? Like, how can I get them distribution to this product? It's just a really cool way to think about it. I agree. We're so smart. I had some other thoughts on that. So when you were talking about med spas and golf courses, I started thinking about other piggybacking opportunities, right? And how like, this med spa is going to go one of two ways.

It hasn't opened yet, right? Yeah. No, no, no, no. In my opinion, it's going to crush or it's going to completely tank, right? That's you, and nothing in between. That's usually how new things go. Like a med spot of golf course is a new thing. He's either going to find perfect product audience fit or it's just not going to go well because golfers don't go to med spots, okay? How far away are we from, remember the game, the Sims, right?

How far away, and this is way too niche. Like not niche for the people listening, but it's niche for the general public, where we can have an AI enabled Sims game where we can like play out synops. scenarios like that on a computer on a simulation. For instance, with the perfume vending machines. One thing I said in that first video they went viral was, I wish I could take a hundred of these things and place them in a hundred different types of venues, gas stations, medspa, all different types, and then just look at the data, say, wow, for whatever reason, the perfume vending machines right next to the ice cream store, just crush, just crush every single

one of them, without exception. We need to go find all the ice cream stores, piggyback and put a perfect like you can't do that affordably you can't do you got to spend millions of dollars and you got to call you got to place them but like what about an AI enabled something where I'm thinking like a deep research right or a chat GPT pro version where you can just type in a prompt and it's going to run simulations based on everything it knows about traffic counts about human psychology human behavior without spending a dime outside of your $200 month subscription fee you can get that question answered you. I love the idea. There's two things that you're talking about. The first thing is you need to, like, figuring out what the demographic is of your ideal customer. And by demographics, I don't

just mean like white, black, male, female, rich, poor, but like, oh, no, 35 year old living in this type of city, they drive this type of car, they go to this kind of a store, like just very granular stuff. So figuring out your ICP, that's one of them. And then the other one is like testing whether or not this would actually work. Because you can have your ICP and you can think, oh, in this area, it's going to work. But then you go to test it and it's like, oh, this didn't, this didn't work. This isn't where you were going with it. But one thing I thought would be really interesting is this would be for someone who actually had some money. Let's say, how much did these coal plunges cost? I think five grand or so. Okay. So let's say you and I are like, let's, let's,

Let's drop 20 grand and buy four of these. So like one or two coal plunges to sanas, cool. And then we just go to different places and for a month, test them at each place. It's like, hey, Jim, can we just rent this from you for a month? Cool. And then see what demand is like. And then we go to the mall. Hey, mall, you have this open spot.

Can we rent this from you for a month? So there would definitely be some upfront costs. You'd have to buy the sannas. You'd have to pay some rent. But you'd like really quickly get to be able to test demand. that's one way to do it. Or you could just do the growth hack things that you've done in the past, which is like, go to the gym and say, hey, Jim, can I just put a sign up here that says in two months,

we're going to have a sauna and cold plunge? Sign up here if you're interested. Just to test demand. Get your first month free by giving us an email with a QR code. Yeah. So like there's a bunch of different ways I think you could go about testing demand. The question is, because I'm with you on this, how do you then leverage AI to make sure that you have a higher probability of success.

That's the piece I'm like, how would you, how would you use AI for that? Okay, I got an idea. You ready? Okay. Let's use the cold plunge example. So let's just speak hypothetically. We want to know what type of location could we piggyback a, let's just say cold plunge. No sauna, just cold plunge.

Cool. All right. So what data is out there? You've got Ivis World Reports, okay, on industries. so we can have operator go pull. And now I know that costs money, right? Hear me out. It goes and pulls an Ivis World report on the cold plunge industry,

and it can tell us in what states they're trending the most. Okay? Okay, so we can whittle it down there. Then we can take it a step further. You can go to Google Trends, and you can find out which states, proportionally speaking to its population, are searching for cold plunges the most.

So that's a second data point for the same type of research. And then you can say, go to the Wisconsin franchise search because Wisconsin is one of the only states that forces franchisors to publish their FDD, their franchise disclosure document freely for the world. So then you could say, you probably won't even need to tell it. It would just do this in this scenario. It would go to the Wisconsin franchise search web page. It would search for any franchise that offers any type of coal plan services.

It would read the entire franchise disclosure document. It would find the locations that are performing the. best. Let's say Colorado is a state where everyone's searching for it. It finds all the franchise locations in Colorado and it finds the four best performing stores and it learns that all four of them are next to an LA Fitness. And like it could stop there and automatically you know, okay, LA Fitness, they might have a contract with this franchise so I can't go there, but I can go to a planet fitness or I could go to a mom and pop gym in Colorado and I know that my chance of success will be significantly higher. And it only looked at like two or three different things.

And it's it's already capable of doing all that today. What are you thinking? I freaking love that. As you, the minute you started talking, I had the idea, do you remember a few months ago or weeks? I don't know what it was where we were like, you need to be thinking now how you're going to use AI. This is one of those perfect examples of in the past, man, sonas are really taken off. Where could I place them? Hmm, I don't know. Google. Where do I typically place saunas, right? You're using Google as like this question and answer. Now your frame of you can go to, I'm going to use AI. They're my employee. They're going to help me find this. Even before you go to the steps that you said, you're just like this prompt, you will have PhD

level experience in the sauna industry. I would like to not only learn about the industry, but find the best locations for this. Help me walk through how I would go about gathering this information. How should I think about this? How should I think about this? And all of a two-way sort of communication with AI and it's like helping you think through it. And then eventually you'll get to like, okay, cool, cool, cool. Yeah, that's good idea. I should look at whatever X, Y, and Z criteria.

Great. Now take that criteria and you can do that with chat TPT with 20 bucks a month. Then you go to deep research and operator and you can say, hey, look, this is what I want you to do. I want you to go to these websites. Then I want you to pull this data. Then I want you to aggregate it here and put it in one place. you can look at the data and then you ask deep research to say okay this is the data take the data

i've collected take all the data on the internet merge them together tell me the top five opportunities give me the justifications and how much it would cost me to operate in those markets like that right there you could do probably in a day or two that would have normally taken us i don't know two weeks right like it's like you find like one subreddit and then all of a sudden you talk to a guy and you have a conversation and it takes us a long time to kind of like aggregate all this information but all of a sudden you've got two employees working for you with AI and you can very quickly get to the point where you're like cool now I know this is my target demographic these are the places I need to look in my city this is where I want to place them and all of a sudden you've increased the probability of your

success by 50% just by having the right research I like I love freaking I love that framework so video's cool, but you know what's better? Long form audio via podcast and my newsletter. TKOPod.com. Go there to subscribe for free to my newsletter. It's one email a week, very tactical. And then go to my audio podcast, three episodes a week, stuff like this. You're going to love it. All free. No sleazy sales, bitch. TKOPod.com. Well, dude, and not just that, but how many times have you heard a random podcast or watched a random YouTube video that had like something insanely insightful that you've never heard anywhere else, right? I interviewed this guy that has this business inside a gym, son and coal plunge and he got super tactical in my interview and he this guy doesn't go on podcast

I think mine was the only one he's ever been on and he was very like this is what we charge this is what our churn is this is how many sales we did the first year this is the address of the gym the gym has two locations our second location is not going to be in a gym like no one's going to find that you're just not but AI could find that I have a tip for people that I think will make somebody listening to this if they implement it a ton of money a ton of money I'm not And like, this is not hyperbole. Go to YouTube. Muhammad, make that the hook.

Is this called the, is this called the just the tip segment? This is just the tip, my friend. Let's say that you find that interview, right? And it's like, go to Chris's interview, copy the URL, then go sign up for this software called Descript. I think the free version offers this, even if the free version doesn't, it's like maybe five bucks. I don't know.

Copy the URL, paste it into Descript. Guess what's going to happen? Descript is going to transcribe that whole entire conversation for you. If you've ever tried to transcribe anything from YouTube, it's a pain. It doesn't give attribution to who's talking. The breaks and the lines are like really difficult. The time markers make it weird. But you copy and paste into Descript, you'll have the entire transcript.

And now you can take that and put it into chat, TPT, and say, this person talked about sonas on Chris's podcast. I would like you to summarize their business model. Help me understand distribution, startup costs. margins, who I need to hire, like any questions that you want. And it will spit it out for you because someone gave a super insightful interview. And now you have the transcript that you can then put into chat TPT or Gemini or like whatever else you use, right? But I've used this so many times like just prepping for podcasts. There's a ton of value doing this.

I promise you if you're listening to this, do it. Like you like and thank me later. It will make you a lot of money. That's a great tip. Just the tip. It could do all this research on these public. publicly available databases, Ibus World, Wisconsin Franchise Search. It could look at S1 documents, publicly traded companies. It could go find interviews with the founder of the cold plunge.com or whatever.

They're selling tens of millions of dollars. Yeah, but I would say, I would say everything that you just named the S1, the FDDs, all the publicly disclosed documents. Everybody else is doing that. It's that next step that you just said, the lab very last thing, find an interview with the founder. Like, if you do that next step, people aren't doing that yet. Maybe if they're getting paid full time and they work at a private equity fund or a venture fund or something like that, a family office. They're doing that. But most people are not doing that. Go find the interviews.

Go find the conversations that they've had because they're going to say way more in those conversations. You could take it a step further and say, okay, the three LA Fitnesses in Colorado where this does the best, I want you to tell me the crime data for those three zip codes, the unemployment rate for those three zip codes, the economic growth rate, the percentage of retail square footage. Like people would not think to do that, but it's all just cross-referencing. And it could do that.

You haven't messed much with deep research. Deep research does this. So I've done this a couple times with deep research where I've said two things. The first thing is, before I even get to that, if you were looking at this data, what are some non-obvious correlated data points that I should be extracting? And then it will tell me, oh, you should look at, it's crime statistics, you should look at income, you should look at disposable income spent at Whole Foods. You know what I mean? Like things that I didn't even think of. Cool.

and then you take those and you say, great, do me an analysis, I want you to see how correlated these two data points are and whether or not the correlation equals causation, because it will run a regression analysis, which is freaking nuts. And it will spit out as like, this is the R squared. This is like, this is how not only correlated they are, but how much causal effect this variable has on this variable. So it's pretty insane. Because like there's a website. I can't remember what it is, But it will show you Iowa security guards named Jim and the price of cacao in Africa. And they're like, they're like perfectly grapped over time. Right.

It's like, okay, just because they're correlated doesn't mean there's causation there. But you can get that information by using deep research. I've used it, but not for like graphing stuff. That's pretty nuts. Yeah, it was nuts when I asked it to do it. I was like, holy crap. Because math just hasn't been as good for these models. Yeah.

So I was going to do a segment on Chris Crone. our favorite ridiculous guru. And can I'm not going to do it because it's just not good content. But can I tell you what the prompt was and what I love about deep research? Please. So my prompt was do a deep dive on Chris Crone, his controversies, upbringing, personal life, family, crazy stories, quotes are sayings, impressive feats. And what I love about deep research is before it gets going, it asks you follow up questions,

which is so nice, right? Because it's taking, like we've been saying for months now, like ask chat GPT, what to ask chat, GPT. Ask it how to use it. But now, like deep research, it takes it a step further. And it just, it asks you, right? It skips that. So then it said, would you like a balanced overview, including both as achievements and

controversies? Or are you looking for a more specific angle? Do you want information, public records, interviews, or social media insights? And I said, balanced. And yes to the second question. So it took five minutes, exactly. it cited 29 different sources and it gave me 3,700 words worth of like really in-depth content. Really in-depth.

Well, like what you're saying on the deep research side is what I'm excited about because typically you ask chaty-pitya question that gives you an answer and you have to like almost think of a way to ask it to then prompt you to ask the right questions. I don't know if that makes sense. Yeah. But like I did this with these research as well. I'll read it to you. Hey, I want to research.

I want you to make me this list of these. providers in this area. And it was like, okay, cool. To provide you a comprehensive list, I want to clarify the following. Are you looking for all providers, including small independent agencies or just major providers? Do you prefer a specific revenue range? Should I prioritize data from official sources like Medicare, state health departments or business directories? How would you like the data formatted in a table? It's like, yeah, I didn't even think of those things. That's a good question. How do I want the data prioritized? Do I want you to like go to Reddit or do I want you to go to the Bureau of Labor Statistics.

Hmm. Like that's what's great about the deep research is it. It's all of a sudden a PhD level assistant to you to then follow up with you. Ask you the right questions because it knows you're going to ask them later anyways. If I have a really intelligent assistant or employee in real life and I ask them to do something, like the best of the best ask me six follow up questions because I'm super high level, super broad. I don't give nearly enough detail ever.

So they're okay, do you want this, this, this? Like I know that they're on it. And they have that attention to detail that I don't have. That's what deep research is. Another cool framework, if you're not spending $200 a month, it's just to think like from open AI's perspective, they don't want to give super long in-depth answers because that costs them more money. But you can still get that from it.

And so if I'm using the $20 a month plan or the free plan and I want like 10,000 words worth of content, I will first ask it for an outline, like a table of contents or bullet points. And then I'll say, all right. Yes. Number one, like I want to, I want you to be as detailed as possible on number one. Okay, number two, be as detailed. Like, it's a little tedious, but you will get tens of thousands of words. That's so funny.

We've never talked about. I do this. I do the same thing, which is nice now on deep research. I don't have to do that. But when you didn't have, you know, 200 bucks a month before you couldn't do that. So it's like, all right. First, how should I do the outline? Spits out the outline.

Great. Let's just do one section at a time. And if you want, you can even get more granular with like each bullet point. I do, dude. I do the same thing because the devil is in the details. You're not going to make a million bucks with some high level overview or scaffolding of an idea. You're going to make a million bucks by being detailed and understanding every part of it. So, dude, I love that idea.

So in the last two weeks, Open AIA has come out with two, what we feel like are game-changing additions to their product. One of them is called GPT operator. What this is is like the first iteration of an AI agent that will, go out and execute a command on your behalf. In the past, I'd ask ChatTBTT a question like, how would I do something? Boom, it returns an answer. Now I can say, operator, I want you to pull this data from this website

and put it into this table, and it will go out and do it. So that's Operator. And then there's ChatT-DPT deep research, which they came out with about a week after Chat-T-PT operator. Deep research will go, you ask it to go do a bunch of things. hey, research all of these topics. It then takes that information, spits it into a report, essentially, that answers your question. Many companies do this. So you mentioned Ibus World. I worked at Ibus World for a while. It's like an industry report company where you can go and

purchase industry reports from them. And there's an analyst who on the back and did all the research and put it together. There are investment banks, large banks, they'll put together research reports for publicly traded companies, consulting companies do this as well. like there's a broad market for these type of consulting research reports. Essentially what this is doing is that are those research reports. This isn't the first time this has been done. Google did this as well. They have a deep research project product as well. That's really good. But anyways, chat TPT introduced both of these. And if you pay $200 a month, you can have access to both. And Chris and I have been diving in to both of them. So Chris, I'm going to start with you.

Operator, pros and cons. Tell me about it. I got to get this thought out of my head first. Are we worried about Ivis World? Like they got it. They're going to have to have to drop their prices, right? Because if I'm Ibis World, I'm just going to use this to generate my reports, right? I'm not going to like just go sit in a corner and cry and shut down shop. But these reports are significantly less valuable now. Are they not? Like it's going to take a while. It's not going to happen overnight. But how do you charge three grand for a PDF? I worked at these companies. So I know this for people who are listening. Ibus World does this. hey, I want to know about aftermarket parts manufacturers in the car space.

Cool. I want to learn about the industry. And so companies like Ivis World will produce a report. They'll say this is how big the total adjustable market is. This is what the typical profitability of an operator is. This is the growth trend. These are the major problems we see. These are the major tailwinds that we see.

This is what the next five, 10, 15 years look like. It gives you a holistic overview for somebody who's interested in investing into the space. So what Chris is saying is, isn't I going to put them out of business? I actually don't think so. I think what it's going to... I don't know. I don't think it's going to put us out of business. I think...

Go ahead. It's not what I said. It's not what I said. Muhammad, edit that out. I don't think they'll have to... They could. They could lower their prices. But I think what it's going to do is their reports are going to become more granular.

And again, this is my personal opinion. I think that the reports are actually going to become longer. because the people that are asking for the reports are going to have their own AI infrastructure to say like, cool, the report was generated. Please digest it. And now I'm going to have a conversation with you to ask the things that I normally would have asked or for you to educate me on these things. So the short answer is like, I don't know.

But I don't think that the pricing will come down. I think it will probably stay the same. I just think AI is going to cause the price of almost everything to go down. Why would they be exempt? So let me say it a different way. I think we're in agreement. What I'm saying is they will probably charge the same price, but they will give much more in their product.

One of those two things is going to have to give. Either they're going to offer the same product and their price comes down or price is going to stay the same and they're going to offer way more. It could also be price comes down and they offer way more. That's what I would think, the last scenario. I just think there's going to be less demand for their product. They could sell it all they want. It's three times longer now.

It's like, yeah, I just don't need it anymore. So if demand comes down, price has to come down. But anyway, so it's very hit and miss. Like I posted my video about my piano experiment about Facebook marketplace. And tons of people were like, it doesn't work on Facebook for me. How did you get it to work? I just, I went to Facebook marketplace and started messaging people. It did for me.

And I had to log into my Facebook account within the operator browser and all this. But a lot of people, it just will not work for them. And I can't explain it. I didn't do a growth hack or a trick. It just worked. right so it's just buggy that's one downside it's glitchy agreed it's clunky yeah like slow it times out it doesn't exactly follow your instructions like there are there's quality control pieces that you have to implement like it makes it does make mistakes it's slow like you said

it's literally going at the pace if not slower than slower a human yeah it reminds me of the lewis CK skit where he talks about being on a flight the first time that like internet on airplanes was invented and they announced they're like here on jet blue we now have internet on airplanes and people logged in and they were like oh this is amazing and then it it broke like 10 minutes later and the guy next to him on a flight he's like this is bull crap this is bull crap and he's like what are we talking he's like you're sitting in a seat at 35,000 feet you're sitting in a chair in the sky yeah and like there were like good a half a dozen comments on that tweet that was like, it's freaking sucks, dude. It's not worth it. It's so

slow. It sucks. I'm like, what? It has to go to space. Like, this is artificial intelligence. I'm like, week two. Like, what are we talking about? Can you be grateful? I think the, I think the other problem is they, there's just a bunch of guardrails that if they had to put on it, right? I think it'd probably really say it'd be way faster, but it'd probably be doing some weird stuff at some point. You know what I mean? So, yeah, I agree with you. This is, this is version one. This is version one. If you're listening to this, that's when you should be playing with stuff like this. You can have an edge.

Like, that's the whole point. Do you have any other cons? Certain websites like Facebook, like the trillion dollar companies out there, you have to like babysit it. You can't even navigate away from the tab. There are third party tools you can use that will like move your mouse for you to circumvent that. But most websites, it will run in the background.

You can run six at a time, I found out. But with some of the more valuable websites, it, I don't. It doesn't work at all, like YouTube. I couldn't get it to work on YouTube. Or it makes you babysit it. So this is the best analogy that I can make for operator. And I've told several people this. Anybody who has a Tesla on the freeway, they can do self-driving.

It's amazing. Oh, it's so great. But you can't. You got to keep your hand on the wheel. Like, you still need to have somebody behind the wheel. You're there. If your hands off the wheel, the wheel will shake. Like it will, it will alert, alert, danger.

while Robinson. Chris freaking bought this like weight thing from Amazon and put on his steer was it an ankle weight to put on a steering wheel so that it was like oh it'll trick the system into thinking that my hand still there. So it's like that's what operator reminds me of a lot right. It's like you still got to sit behind it. You still got to make sure that it's doing its thing. It's going to ask you. It's going to make sure that you're paying attention. Blah, blah, blah. However, so I'm just switching from the cons to the pros. However, anybody who's had self-driving on in their task. Tesla, you've seen the future. You know what's coming.

It's like, oh, this is going to happen. Like, yeah, this is annoying right now. Yeah, I got to put me behind him the steering wheel. But full self-driving is going to happen. When I got my Tesla two and a half years ago, I had to do all those things. I bought it for full self-driving and it didn't work well at all. Like I used it 2% of the time. And my app tells me that.

So it's not an exaggeration. I bought an ankle weight. My friend Keith had like this fancy little weight that he put on it. And I went to buy it. It was 150 bucks. I'm like, I'm not doing that. Ankle weights were 10 bucks.

But it's still, so that got around like the having to put my hand on it, but it still didn't work. It was unreliable. Today, today, you don't need your hand on it at all, ever.

You just need to be looking at the road every now. Is that right? Yeah. That's gone. And it works amazingly. Like it works. My car drives itself now. The one thing I hate about it is like,

I can't look at my phone. It won't let you because the cameras, watching you. That sounds silly, but like, I want to look at my phone when my car is driving me around and it won't let me. So whatever. Do you do self-driving not on the freeway? Because there was a time where it was only self-driving on the freeway. You can do self-driving off of the freeway now, too? Yeah, I've been able to do it anywhere since I've had the car for it. It was just glitchy both on and off the freeway. But now it's not glitchy anywhere. Really? Yeah. It's legit. How does Waymo get around it? Because like, who posted about the James Camp maybe?

He took a Waymo ride the other day and there's no driver. It's all autonomous. Why is it Waymo able to get around that? But you're not. Is it Texas? Is it like local? It comes down to LiDAR. Waymo has LiDar and that's like a one to $200,000 investment per car in addition to the cost of the car.

Elon has said from day one, LiDAR is unnecessary. It's too much. It's overkill. It's redundant. When really he's biased, like he can't sell $200,000 cars. So LiDar has to be irrelevant. right yeah i was going to ask is it a pride thing is you like oh no dude it's it's just a freaking we don't need it bro it's just a bias thing but they have like these benchmark tests now like now that full

self-driving is a thing and waymo is by far the most accurate like the least amount of collisions per million miles traveled Tesla is not great compared to it so like lidar gets you the safest self-driving car or ride possible but waymo is owned by google right yeah Okay. I was going to say, Elon should just buy Waymo, integrated into his company. Yeah. But Google's not going to sell that crap. Have you seen Waymo's out in the wild? Yeah. No, personally, no. I live in Idaho, dude. I know. I don't think they're in Dallas, but they get shot. They get people in Idaho see a Waymo. They're like, oh, what the heck is? That thing's driving itself. Oh, turn to the aliens. Karen, get your gun. When we went to California, they're everywhere. And I saw a stat. I need to look this up.

I knew that San Francisco allowed it, and I thought that San Francisco's municipality had granted it access. So I still think that there are local ordinances that are going to allow or disallow these self-driving vehicles to operate. So like it'll still be a patchwork approach for a period of time because there's going to be cities that are like San Francisco that are like, cool, you can be self-driving. And then cities like Dallas that may be like, no, screw this. You can't have this inner city. But right. I want to say there's like 15 or 20,000 Waymoes. on the road today, it's, it's a lot more than you think. No way, Mo. It's a ton. It's a ton.

All right. So like, that was my analogy with operator. The Tesla thing is I've seen the future. Like, I can't unsee what I've seen. As clunky as it is, it can click into things. It can pull data. It will go to different websites. And you can like you can see that, oh, okay, this is going to happen. Like this is, it's not just pie in the sky. And it's going to happen faster than you think. Unfortunately, we're all biased because Elon's been saying FSD was coming for 15 years. And it's like, still not here, bro. What are we doing?

Oh, yeah, AI is probably going to be the same way. It's not going to be the same way. It's iterating at a much faster rate than we have ever been used to. So like this 15 year time horizon, like shrink that down to 18 months until this is probably fully operational. So that's how I feel about operator. All right. So that's operator. Deep research.

I play with deep research more than you've played with deep research. You play with operator more than I play with operator. I like it. Here are the cons. On some things it takes forever a week ago, let me just explain what it is. You can tell it, hey, I want you to do research in these areas and bring back some market analysis for me. And you can be as specific as you want.

So I gave it some parameters. I'm looking for these types of health facilities in this state. This is the revenue size. Go to this website. to validate it, go to this website, to pull this information, put it into a table, identify who's on public, who's private. This was over a week ago, I still haven't gotten the report. So, like, I could have put that together in probably a day or two.

I still haven't gotten a report. So I'm like, what the, like, how is that not put together? And I've every day been going and like, hey, any update on this? It's like, this information is important. And we want to make sure that it is put in the right format. That you remember my, but never, anyways, that's a con. Like, it can take a long time because it's still pulling a lot of information. information. The other con is less about the technology and more about just the user, in my opinion,

unless you've gone through like a PhD program and you understand how to conduct research and how to put together these reports. And so you're pulling correlated and non-correlated data and you're looking at causation, et cetera. It's kind of hard to tell deep research to do something in a way that's going to give you the most value out of it. So I would like to see some type of an airface on deep research, it's doing this, but not enough where it's like, what do you want to learn? Okay, I want to learn about opening up saunas near gyms. Cool. And then it just asks you a million questions, not a few, but like a lot of questions to get really granular on what you're looking for. So I would say that's the con from using it. One thing I wonder about deep research is, is the length of time

it's taking to find your report going to be correlated to the quality of what's in there. Do you know that yet? Don't know, because I haven't seen it. Yeah. And I specifically asked it to do a report that I've done in the past because that I know takes a lot of disparate data sources and sort of puts them together to build what I'm trying to build. Like I'm trying to build basically all of the providers in a specific state of health, so healthcare providers in a specific state with their revenue from different revenue sources.

And I've done that before pulling from different websites. And I want to know if they're public or private. Because if they're private, they're more likely. to sell than if they're public. So I've done this before. So I asked chatty B to operate or deep research to do it because I just want to know how good it is. And once it gives it to me, then I'll be able to like, look, this is what I did. This is what it did. Let's compare the two. And if it's if it's good or like 90% of the way there, my mind will be freaking blown, right? But I just haven't gotten that yet to be able to validate it.

I'm curious to see how it is because I mean, either it's going to be amazing or you're going to be disappointed for the time it spent. And it might have just like hit a glitch and just stayed there for a week, you know, it's kind of what I think so. It's kind of what I think is up on it. Right. The pros on it are you have an analyst. Like most of the time you'd hire an analyst, $50,000, $60,000 a year out of college and they're just going to do a bunch of research for you. You now have an analyst that can go and do the research as well as a fresh out of college analyst. So, and it puts it in a format that is digestible that you can read. You can ask it, follow-up questions and it'll answer it really quickly. So it's it's a really cool product.

If you're just trying to get a quick approximation or quick up to speed on how something works. So for that reason, I really like it. Like I could see you using it for a ton of your content, right? Like, hey, I'm researching above ground pools in Wichita, Kansas, help me understand the market. Like I could, I could see you using that and then going and making a video of it. You know what I mean? because you know the questions to ask it, and it's going to give you a lot more analysis than typical JAT TPT would. I asked it to find the list of the top 10 episodes from Greg Eisenberg's podcast and my first million by views and who their guests were so I can go snag them as guests as well.

Like that's something I asked Kamal to do yesterday. Yeah. I don't want to see if it can do it for me. But it has problems with YouTube, you know, so I don't know. All right. Why don't you give your pros on chat TBT operator? real quick. So it's super good for doing like very low skill base level research. For instance, I had to go, I had to travel last week. And so I said, find me a hotel in city name within five miles

of this exact address with four stars or more under $150 or night. And it did. And I watched it. I watched it on price line. And it, it felt filters. It thought of a couple things that I didn't even prompt it to think of and boom like that's something that an executive assistant would do and then I just went in and booked it and it could have done that for me so that plus like find me a Toyota tundra 2015 within 100 miles of my zip code under 100,000 miles under 30,000 dollars and so it will save you like 10 minutes there and then you can like and it will go a step further and start messaging people but the more you ask it to do the more it kind of glitches out but it's really really good for doing stuff like that. And then for notating stuff like that in a Google sheet, it's great at.

Bottom line is, we're early. I think if you can afford it, start playing around with it, agents are going to be a thing. Because by the time agents become usable, if you're trying to get up to speed then, people like Chris and I are going to be way ahead of you. So I just get educated now. Well, and don't take its answers at face value. Like push back on it as much as you can. Like it solved the capture for me, but then it wouldn't submit it. And I, It still wouldn't, but I kept asking like, no, submit it, submit it. It's like to capture. What do you mean?

It solved a captcha, like the ones that have the weird words and the function shapes. Yeah, yeah, yeah. Yeah, I thought that was kind of basic. Like I'm like, at first I was amazed. I'm like, oh my gosh, it solved the captcha. And then I went and asked it. And I was like, is capture solving a thing? Like, because you can do it manually.

They're like capture farms in Indonesia. Most people don't know this. And if you're running a bot, the bots can't get around captias. So it will literally pay people in Indonesia like three cents a capia. like three cents a capture to manually solve them remotely. Oh my gosh. That's a thing. But I asked ChatGPT for, oh, if like solving captures has been around for a while,

like via automation and bots. And it's like, no, this is brand new. So it solved the capture for me. And then I'm like, okay, now enter it. And it wouldn't. You know, it didn't. It had a guardrail. All right.

What did you think? Please share it with a friend. And we'll see you next time on the Kerner office.