- Speaker #0
They ask investment managers how often do you use ChatGPT and how much do you like it. Different investment manager says like well I use it every day, I use it three hours a day, whatever and I really like it, it's so stunning or like yeah you know it makes mistakes, it's kind of annoying, it's too friendly, whatever. So they rated those answers and then they compared it to the investment success of the investment managers and lo and behold the more you trust ChatGPT the worse your investment decision was. Why? Because people don't understand the shortcomings.
- Speaker #1
What do you think is the biggest shakeup e-com that's coming that most people that are in e-com, that maybe are the sellers, are just not paying attention to that they need to wake up and smell the coffee?
- Speaker #0
If you are an aggregator via Amazon or via Google, that business model is going to break down.
- Speaker #2
Your watch on marketing misfits. Ron Farrar and Kevin King.
- Speaker #1
What's up, Norm? How you doing, man?
- Speaker #3
I'm doing great. In fact, I'm having a cup of coffee.
- Speaker #1
I think you and our guest today, I'm maybe the only one not drinking the coffee, because I hear that coffee and AI go together really, really well.
- Speaker #3
That's what I just heard, so I can't wait to get into this podcast, by the way.
- Speaker #1
Or is it AE? I was listening to a podcast the other day, and this guy kept calling it AE. And when the AE does this, I'm like, what? No, dude, it's AI. And then he's like, no, the A-E, it's like, it's like, it's like Wi-Fi, you know, like my ex from Latin America, because the I is pronounced like an E. So it's not Wi-Fi, it's Wi-Fi and it's A-E. And I'm like, no, it's A-I. It's just like, is it A-E-O? Is it, is it G-E-O? Is it A-I-O? What, what, what is it? Nobody knows.
- Speaker #4
E-I-I-O. E-I-E-I-O. I think that's more than maybe. I mean, yeah. But,
- Speaker #1
but nobody knows. And. You know, there's a lot of people out there jumping on this AI bandwagon right now and agencies. We have an agency, you know, that's doing AOS for some people, for e-commerce sellers. And they're jumping on this and there are no experts right now. Nobody is truly an expert. And but there are people that are guests. Right. But there's people that are way ahead of the pack and on the cutting edge because it's a rapidly changing industry. And you and I both keep up with it pretty, pretty thoroughly. And our guest today is one of those guys that's... on the cutting edge. And I think he's going to blow some minds and including possibly even ours, he promised to. So we're going to hold him to that. And this is going to be, I think, a really cool discussion that everybody needs to listen to. I think this is going to be one of those episodes. We haven't even recorded it yet. So we'll cut this out if it's not true. But I think it's going to be one of those episodes that's going to be game changing for people that actually listen to it and actually believe what they're hearing and understand what they need to do because the world is The world is about to radically change. This is going to be cool.
- Speaker #3
Can I call you Kev-ee?
- Speaker #1
You can call me Kev-ee. No, it's K-E-V-I-N, so it'd be Kev-in. All right, I'll be in.
- Speaker #3
All right, let's introduce our guest, Kev.
- Speaker #1
All right, let's do it. We've got Lux, Lux Finger, coming on as soon as Norm figures out how to hit that button. I hit the button. There you go, there you go, Norm.
- Speaker #0
Fresh coffee, ready for you both.
- Speaker #1
those are awesome how you doing man good to have you on life is amazing it is it is uh so uh not a misfit amazing life is amazing well they have this people people always ask me like they say uh how are you and i say it's another day in paradise yep yeah i always say it's another day in snow but it's because you're up in canada that's it it's another white day right yeah um welcome uh let's uh so uh you're i mean i think i'd heard about you uh the reason i tracked you down uh is i think it might have been the jason scott show or it might have been some something like that where i heard you or you were mentioned as and i started looking at some of the stuff you're doing i was like wait this guy is someone we got to get on the podcast so i'd reached out to you um i think back in the summer of uh of 2025 and actually uh finally we got you on and uh We were chatting a little bit before, and I think this is going to be cool. So for those that don't know who you are, just give us a little bit of a background of where you come from.
- Speaker #0
Yes, absolutely. So I'm a faculty member of Cornell University. I teach about AI and data and products. And for 15 years or 13 years, I have been in that data space. We call it AI today. We called it machine learning before. We call the... data for that so i made every of those hypes i built a couple of companies out there sold them my last company recently sold to xgen ai and i'm now the head of ai go figure right i'm the head of ai for xgen and we are helping e-commerce to um adapt to the new world awesome and you you did some work at google and linkedin a few other of the big boys now yeah totally i sold my last company like the the company before i sold and then i went to linkedin i built up the analytics team then i went to snapchat uh like was there as we went public so i I cleaned up the data and tried to set some data sanity there. Then I went to Google. I was one of the founding guys at Google Health. So we went into health care and used stuff like you can use data to predict that somebody goes to a hospital six months before that person actually enters the emergency room. And if you have that knowledge, then you can save people's health care costs. I then went to a NASDAQ-on-L.I. company in New York, where I became the president of AI and products, also in healthcare. And then I started my own company, the second one, where I built essentially generative AI models and AI recommender systems for e-commerce. As well as I dabbled into, and we should talk about this, the new space of GEO, generative engine. optimization also called LLMO or AO. Like there is somewhere an O. Everybody should have an O.
- Speaker #1
So a lot of people don't understand that AI is not new. A lot of people think that AI, when ChatGPT came out to the public in 2020, was that 2022, late 2022, that was their first exposure. But it's been around for, like you just said, for ages.
- Speaker #0
Let's do a pop quiz. Let's do a pop quiz. I won't hold it.
- Speaker #1
47 or something like that.
- Speaker #0
When did AI start?
- Speaker #1
When did AI start?
- Speaker #0
Almost the first discussion of people thinking about this AI thing to be honest.
- Speaker #1
It was in the 40s, 1940s or something, right?
- Speaker #0
Yeah, a little bit like 56. So the first like mathematical models. Now, honestly, if we think about AI, about data. helping us to predict the future, you know, then like a linear regression could be considered AI. And I think because you have data for a linear regression and then you use a data point to inference on that linear regression. And so you have an artificial intelligence way telling you something. And fun fact, neural networks, chat GPT and all of that is nothing else as billions of billions of linear regression stuck in a kind of a function. But it's... It's essentially that. So if you did your K-12, you know what is AI.
- Speaker #1
So those that don't know, what's a linear regression for the audience?
- Speaker #0
Okay. Everybody knows how babies grow. Let's start there. Like a baby comes to the world, has a certain size, and then you feed it and it is growing a freaking linear, meaning every day the baby becomes taller and heavier, right? So if... I know what the start rate is of a baby and I know the growth rate of a baby. I can tell you how heavy and tall the baby will in one month, two months, three months. At some point in time, this stops to be linear. That is a linear growth, right? So now how do I know what a growth rate is of a baby? Well, I train a modeler. Sounds totally funky. It's nothing else. I go to a hospital, grab a couple of baby data, month and size, and chart them on a like. month over weight and then i have a linear line through it and that's a linear regression and that is ai because now i can ask my super fancy model oh let me ask how heavy is a baby with eight months so i look at my data i have a bunch of points there and that gets gives me the amount of the weight for baby in a month so this is like guys and i didn't want to drive this down to a technique around it. I promise you. That knowledge about how linear regressions work will help you to sell better on chat GPT and perplexity. Trust me, it's really true. Understanding a little bit of AI goes a long way.
- Speaker #3
How old are we as, are we an eight-month baby yet?
- Speaker #0
In terms of AI? Well, I mean. AI is a tool, guys. It's kind of like, yeah, like we, like people get so excited about ChatGPD, but essentially it's just a new interface, which... Could be pretty exciting, right? I mean, like as we had the last interface changed, we called this the World Wide Web. And everybody, every shop had an interface called a website. And every marketing guru was meant to understand that website. And now we have a new interface. We can chat to it. We can ask a question. It finds the right things for us. It's an interface.
- Speaker #1
Yeah, it's a very... powerful interface that's growing super power more powerful by by the day it seems like but a lot of people i think i agree with you when you say it's a tool i agree norman i agree that it's a very powerful tool if you know how to use it but most people what's a chat gpt has like 800 million that they users they say i don't think that's really 800 million people i think there's businesses and other accounts i think that's accounts or something but anyway whatever it is most of those people have dabbled they've written an email if they've done something and maybe a little bit of research to ask you a couple questions. They don't... A lot of people don't understand the true power of what it can do. They're playing with Sora or with VO3 or something and making some graphics. What are we looking at here on true game-changing, earth-changing, humanity-changing power of what is happening right now and what the future holds when it comes to AI, in your opinion?
- Speaker #0
Well, that's like, how many hours do we have? so I I do a whole course on this about like, I think, Yeah, the quick summary is ask ChatGPT about it. No, so... So if AI is a new interface, then this means how we interact with the World Wide Web currently is about to change. Meaning all the shops who are selling via... a homepage or like a web page need to reconsider.
- Speaker #1
It's going to be like they have like notion pages or something.
- Speaker #0
Yes. Not quite. This will be a different interaction and we can go into where the horizon is. But see, it's not that we haven't, like I tell you a dirty secret about academics. We always look backwards, right? So it's not that the world hasn't seen this before. Like the internet came around and suddenly news organizations around the world suddenly said, it's not this paper thing anymore. We create the homepage. Well, Facebook came around and killed that homepage and now it was a feed, right? So here we have a new interface. It's not the feed anymore. It is our chat interaction and our conversation and that will change. But it does way more than this. And therefore, Kevin, if you ask me, what does a future hold? loads of things. But definitely the way we shop will change because it goes from the web page, Amazon's web page is essentially not as important anymore, over to your trusted chat conversations. OpenAI just recently launched their agentic commerce protocol in order to facilitate exactly that. But you can go further. You can say, hold on, I can copy everybody in this universe. I, for my own course, I copy myself and create a Lutz Finger Copy so that my students can 24 hours a day, seven days a week, talk to me if they want, even with a German accent. Right. So that is helpful because I support them throughout their journey with my knowledge way better than I could do in person. Now, if you can copy me, then you can copy anybody. That has a completely change on service workflows, custom interaction workflows. But it has as well, obviously, a complete change in the notion how to represent yourself. What is true? You might recall the song from Drake and the Weekend, Something Up My Sleeves, which wasn't really Drake and the Weekend. It sounded like it, but it's not him. It was fake. Drake got very upset about it. But hey, he shouldn't. Somebody took his image, his vision, his space and made a copy of it. So what's the new world look like? We have a completely different question on IP rights. We have a completely different question on brand rights. We have different questions on marketing. And we have a different question on selling and sales approach. All of this is changing.
- Speaker #1
Hey, Norm, you'll love this, man. I talked to a seller the other day doing 50k a month, but when I asked them what their actual profit was, they just kind of stared at me.
- Speaker #3
Are you serious? That's kind of like driving blindfolded.
- Speaker #1
Exactly, man. I told them, you got to check out Sellerboard, this cool profit tool that's built just for Amazon sellers. It tracks everything like fees, PPC, refunds, promos, even changing COGS using FIFO.
- Speaker #3
Aha. But does it do FBM shipping costs too?
- Speaker #1
Sure does. That way you can keep your quarter four chaos totally under control and know your numbers because not only does it do that, but it makes your PPC bids, it forecasts inventory, it sends review requests, and even helps you get reimbursements from Amazon.
- Speaker #3
Now that's like having a CFO in your back pocket.
- Speaker #1
You know what? It's just $15 a month. But you got to go to sellerboard.com forward slash misfits. Sellerboard.com forward slash misfits. And if you do that, they'll even throw in a free two-month trial.
- Speaker #3
So you want me to say go to sellerboard.com misfits and get your number straight before your accountant loses it?
- Speaker #1
Exactly.
- Speaker #3
All right. You know, I was watching and I don't know, let's see, you've frozen. No, you're not. Good.
- Speaker #0
I'm so cold. Join my up here in Canada, man.
- Speaker #3
I was watching the news and they were talking about this actress, AI actress.
- Speaker #4
Yeah.
- Speaker #3
And she was being represented by an agent. And what they were saying during that episode was that this is going to happen more and more, that people are going to develop the character in AI. And these agents are going to represent them. That's where the money is. And I was just, I was sitting there saying, I didn't even think about it. Then you bring up Drake. And, you know, what's going to happen to people like that, that you duplicate? Like, who owns the rights to Drake's image?
- Speaker #0
Yeah, well, Drake should own the rights. Now you, like, and like, think about Drake. Look, Drake is somebody who has not reinvented himself many times. He has a certain style and he is copying the style over and over again. I would even claim that he's most likely not even writing his own songs. He has a team and he is directing this team to write his songs. Now, I can create an AI that represents his songs and creates in his image, in his idea. and um If you think about Princess Layla from Star Wars, in her latest appearance, where she is at the latest Star Wars film, she has a comeback despite the fact that she is dead. We don't need that human person anymore. We can superimpose images. And Hollywood does it all the time. When you start a Hollywood movie, they normally copy the body of the person just in order to make sure for the case you die, that not the whole investment goes waste. Therefore, they will keep you alive for the movie. That's what you signed up for.
- Speaker #1
They did that in Fast and Furious. The guy that died in the car wreck, when they're filming that, Paul, whatever. Yeah, they did that exact thing.
- Speaker #0
Case in point.
- Speaker #1
Yeah.
- Speaker #0
It was expensive. Now with Zara, I can, you know, if like, if your audience wants to follow me on Zara, I'm at Lutz at Zara too. Follow me, I follow you back and then you get access to my liking and you can make any cameo you want with me in it. Essentially, I rent out my space. So we definitely, if you talk about marketing misfits, we definitely should talk about that there is a whole space in geo llmo ao whatever meaning how do i impact the new aggregator in town how do i impact perplexity so when i'm as a consumer talk to perplexity that my brand is vision there right that's level number one level number two is how do i represent myself how do i make marketing imagery like image from something look at black forest labs one of the best tools out there in the market in terms of image creation it's not like you remember the times where you kind of created an image and suddenly you had six fingers or five fingers and it was always a giveaway that's quality like and the quality issue is fixed you need a control issue how do i if i create an image for cartier the puma jumping how do i make sure that this Puma is really the size of the Puma. The brand image from Cartier wants that. That's the control question. So then I work on that. Now, the third level is obviously now I have not only found a new sales channel. I've not only found a way to create, lower the cost of marketing creation. Now I give now my image out, my liking, my brand out, and I let others create my brand style. And I make every of my customers a creator. So here you have it. And this is only, like Kevin, I'm only, only touching here on e-commerce marketing space. Like I can go in way more depth into what it means in healthcare, what it means in finance, what it means in social networking. There's more to cover. I do this actually in my, like it's an open to the public course from Cornell, building AI solutions. But... If we spend only the moment looking at what does it mean for e-commerce, it means a different sales channel. It means a different interaction with your brand. It means a different way to create marketing. And it means a different opportunity how you protect and recreate your brand.
- Speaker #1
So is that where Web3 or the old Web3, you know, NFTs were a hot thing for a moment. And all the little cartoons and characters and stuff. But the underlying technology under that is sound. Is that something that you think is going to see mixed in with AI to protect these IP rights where it's going to be blockchain type of stuff that's helping? So when you share your likeness that there's something on the blockchain or an NFT that goes with that and it comes back and you're getting a little Scooby snack off of that. Do you think something like that is coming?
- Speaker #0
You're trying to get me on the hype route here. I'm like, does he follow the hype? Actually, so I never got onto Web3.0. I kind of thought it was BS and I still stick to that. um blockchain is definitely the right technology for us to have distributed um knowledge or distributed uh quality it's totally not clear how this works look i mean one of the mcp is a concept which was hyped six months ago i always said i'm not sure How we do this with security, how we do this with payment, how we do this with the privacy, still not solved. And yes, it happens as we would expect it to. Like the biggest player in the market tries to determine what is the best protocol for their needs. And the biggest player in that space is no longer Amazon. No, it is JetGPT and OpenAI.
- Speaker #1
And they're using AGP. instead of MCP, right?
- Speaker #0
Exactly. And how did they do this? They solved exactly those problems, which we described. I mean, look, this is like MCP, it's a P is for protocol. I can have any protocol I want. If I see you, I can fist bump you, I can shake your hand. You wouldn't care less as long as you know who I am, right? So like protocols are, anybody can create a protocol. The question is, do you get this implemented? And the move was OpenAI just did. brilliantly go to the number two in the market which is walmart and tell walmart look you're number two we can get you to number one if you just play nicely with us we have seen this verizon against at&t as apple came around well here we go next chapter and how do you dominate a market well done
- Speaker #1
Yeah, well, they started with Shopify and Etsy and then shortly came out in Walmart. And then Walmart, originally people were like, is it just the 1P products? And no, it's all the third party products too. And then a month ago, at the time of this recording, Amazon says, we're going to block it off. We have Rufus. We're going to block all the GPT, the service. And I think that was a massive mistake. I do too.
- Speaker #3
I think that's a huge mistake.
- Speaker #1
Huge mistake. And they may have to rethink that. I've been saying for like a year that Amazon is not going out of business, but the way you shop and you don't go to the mall anymore to shop. But Amazon has such a strong fulfillment system that if they adapt properly, they'll be okay. But Walmart has that too. And so it's interesting to see what's going to happen here. I think a lot of people are not on this bandwagon yet in the e-com world. The sellers, the people that are making their living. Not that... consumers I'm saying, but the sellers, I don't. I keep saying that guys doing $5 million a month right now on Amazon are going to be doing $500,000 a month or 500,000. I'm sorry, 5 million a year are going to be doing 500,000 a year in a year or two from now because they're not paying attention. They're still doing things the old way with keyword based and stuff and tech based. They're not doing any AEO and LLMs and getting their brand mentions and getting into the social media and getting the structured data and doing all this stuff. And you don't have control if you're just selling on Amazon on most of that stuff. And so it's what are you seeing when it comes to. e-commerce, you know, you have the browsers that just came out, several of them that have the agentic capabilities and they're kind of clunky right now and little, you know, have a few issues, but that stuff will get dialed in and it'll get better. What are you seeing? What do you think is the biggest shakeup in e-com that's coming that most people that are in e-com that maybe are the sellers are just not paying attention to that they need to wake up and smell the coffee?
- Speaker #0
Always a good one. He knows how to wake me up. I think for e-commerce, you can break it down in those three different areas. Area number one is how do you sell your sales channel? Area number two, how do you market? Area number three is what do you work with your brand? And very clear, if you are... an aggregator via Amazon or via Google. I don't care. That business model is going to break down. So if you now look into the whole world of our sales shops, most of them are aggregators. Most of them kind of have figured out a way.
- Speaker #1
You're talking about the marketplaces.
- Speaker #0
Yes, exactly. Well, our aggregator of intent, because they're going and grab keywords.
- Speaker #1
Oh, yeah.
- Speaker #0
Amazon is an aggregator. You're right. Amazon has a distribution network. I actually wrote. I'm publishing at Forbes. The research I do. And the stuff I talk about. If you look up Lutz Finger. And Forbes. You find me. In June this year. I wrote about. What will that mean for Amazon. Walmart. And Best Buy. Because these are very three distinct abilities category leaders can play for. Amazon, the homepage will be more or less less and less important because you don't go anymore to the Amazon page. You talk to your trusted LLM and order from them, as we see happening now. Amazon tries Rufus. Honestly, I did a public show comparing Rufus. It's so bad. And it is bad because it doesn't have the long tail of knowledge as Oak Mayoy has or Perplexity has. And I can give you a technical explanation to it as well as a product explanation. But let's not go on down that rabbit hole. The second part is Walmart. Walmart has been a price leader and they will continue to be a price leader. And they will position themselves within that. perplexity ecosystem as a price leader and then best buy doesn't have a distribution system doesn't have a price leader advantage however they have complex products and they will try to be the most knowledgeable about their product specifications in order to help the customer so and that is the business model which is most critical to where open ai positions themselves is them self so you have those three buckets in the sales part. So if companies don't wake up at the moment to understand that their trusted distribution strategy is completely going down under, then they're going to miss out. That's on part one. Part two, we can talk about how do you make marketing? How do you brand? How do you place imagery? Because there's a complete change in the market. Part three is... Do you have a brand? Do you have a brand story? Then you're probably about to survive. And now we can talk about what that means.
- Speaker #3
So getting into part two, I'm really interested to hear what you have to say about marketing. How do we go out there and compete? How do we market?
- Speaker #0
Yes. So essentially, and we... Here, it's a little bit important to understand how AI works and what AI can and what AI cannot do. Essentially. I can go to ChatGPT and ask me, write me a nice marketing pitch for Marketing Misfits session results finger. And it probably knows Lutz Finger, it knows you guys, so it will come up with something nice. Will it be exactly the style you guys like? No, because it has all the weirdness of the worldwide internet and it just takes averages and it doesn't fit, do exactly the style as Norm has it. So you need to tweak it, you need to train it, you need to massage it so that it's Norm's style, right? But it reduces your cost once you have it. And you can do this for text and you can do it for imagery. So very high level, we will see focused models for each brand containing norm style. And the two things you need are actually the three things. Two things. The two things to look out for. Anybody who tells you about AI. And look, I mean, I'm the head of AI for Xgen and we are. right in the middle helping e-commerce companies doing this. For us, it's two big questions, quality and control. Make sure that the text about this show is the right quality. Don't use swear words. Don't be too complicated. Don't have too old-fashioned wording only because you ingested a bunch of old books, right? Quality. And then the second one, control. Because Norm might say, no, no, no, this is not my style. I want it a little bit less, a little bit more flashy. And no, not too much flashy, a little bit more flashy. So here it's a control problem. So quality and control are the two things we need to bring in. And by the way, it's the same thing, not only for text, same thing for images. I mentioned early on, one of the companies I personally admire is Black Forest Labs. Why? Because they see pixels as an infrastructure. They allow you to create the right imagery of, let's take a brand, and then control that image according to your brand guidelines. And that's the important part, right? If you are, let's say you are Cartier. I mean, let's look at this Cartier. They have this Puma made out of diamonds running through the night. Obviously, this is all AI generated. Now you want to make sure that that That Puma, which stands for the brand, is really the Puma you want to have. You don't want to have a cheap copy of a Puma, even if it's an AI-generated Puma, right? So that quality and control is what becomes important for the second part. If you have that, you can drive down costs dramatically and create way more appearance in that space, which is an agentic commerce space. And this is important. I see so many people creating web pages now using ChatGPT. That's so stupid. Let me say this again. If you are in the business of creating web pages with ChatGPT, it's stupid. Why? Because you put the page out and then you are totally happy that ChatGPT crawls that page. Look at this. They're crawling the page. Yes, but they're not using it for training. Okay, let's come back. back to the babies. I talked about early on, how do you know how heavy an eight-month-old baby is? You select a bunch of baby data. So probably in your data set on eight months, your baby is somewhere between, I don't know, 15 pounds and 20 pounds. And there's a distribution of data, different data points. Okay. If I now use my model to give me the answer about an eight-month baby it will take me exactly to the line which i drew which means it's 16.5 thoughts if i create now create data using my model i get only 16.5 16.5 16.5 16.5 i lose diversity and my future model looks way worse this is called model decline sorry small technical detour however if you use chat gpt data to train chat gpt the model quality of ChatGPT goes down. And what will ChatGPT not allow? That the quality goes down. If Elon Musk reportedly spent billions of dollars a month on training the engine, the team will make damn sure that the quality doesn't go down. Meaning all of your ChatGPT created web pages, yes, they get crawled, but they get never used for training.
- Speaker #1
Are you looking to quickly boost new Amazon product launches or scale up existing listings to reach first-page positioning? The influencer platform Stack Influence can help.
- Speaker #2
That's right. Stack Influence pushes high volume external traffic sales straight to Amazon listings using micro influencers that you only have to pay with your products.
- Speaker #1
They've helped up and coming brands like Magic Spoon compete with Cheerios for top category positioning, while also helping Fortune 500 brands like Unilever launch their new products.
- Speaker #2
Right now is one of the best times to get started with Stack Influence. You can sign up at stack. influence.com or click the link in this video down the description or notes below and mention misfits that's m-i-s-f-i-t-s to get 10 off your first campaign stack influence.com so what you're saying then is basically any ai generated content whether it's your web page or maybe you're creating ugc video content with ai or whatever that will Maybe accomplish one purpose, but it's not going to give you that brand authority and that recognition and that's the training. So that AI then uses you as a human, like what are the humans? That's why Reddit and LinkedIn and some of those where it's human interactive are way heavily in Wikipedia, way heavily in the training data and authority. So you're saying that all these people are creating a thousand UGC videos and putting them out there. Maybe they're going to make a few sales, but that's doing them zero good in the future on getting any kind of brand authority or brand recognition. Zero. So you've got to be the human doing it. And AI can tell. I know there's some watermarks and there's some tools. And some people say, well, nobody would ever know this is AI. But there's actually ways to figure out that most of this is AI. So that's, I think, a myth. But so that's a very interesting point, because I think a lot of people don't realize all this AI content, all this, you know, whether they're writing a newsletter or they're writing a blog post or they're. creating websites or they're creating video content and they think they're putting out massive amounts of content that's going to help them and it's actually could backfire it may be a short-term win so let's clarify that like if you create content you might have two users one
- Speaker #0
is a human user right and creating content for human users might be helpful okay right well yeah you're right right i if like i would have written this by myself i need somebody to to read it. I want to help people, therefore I create content. Awesome. Human user gets it, says, thank you for the summarization. This was awesome. This was helpful. Human users is good. If you think about, however, and this is bucket number one, the sales channel, how do I make sure that I get positioned correctly in ChatGPT perplexity and other models? I need to influence those models and I can influence those models with content. If I use... If I have that as a use case in mind, that won't work. Now, Kevin, it's not easy for a computer to differentiate. Is this a human generated content or a computer generated content? There are tools out there. I believe this is all snake oil. This doesn't really work. Technically, it doesn't make sense. There is no way to distinguish. That's how we build it, right? Maybe you find a stupid mistake, like five fingers on a hand. Maybe something is a nutshell. Then it's a quality issue. We solve this over the time, but technically there shouldn't be any differentiation. The same way you cannot technically decide whether the 19.5 pounds, which I just told you, are because I really had a baby with 19.5 pounds, or whether this was just the answer from my line. Now you understand why I brought this very simple use case up front. So we cannot differentiate. The thing is somewhere else. If I train a model I use to work for Google. So we get data. And then we use the data to train. But then obviously we look, does the model know more? Is it better now after I trained? And if it's not, I just delete the model. Meaning if I take BS data, my model is not going to learn. I don't know what works, but the model knows.
- Speaker #2
It's like a loop. Yeah. Okay, I see. Okay.
- Speaker #0
It's like I might scrape the data. I sampled the data. I tested. I realized the model doesn't really care. Now, let me tell you something. If I tell you, Kevin, I today saw something amazing. I saw, hold it, hold it. I saw a cat. You're kind of like, you saw a cat? Yeah. So what? Because you, Kevin, saw so many cats in your life. Meaning I'm showing you yet another cat. is really not amazing right if i would tell you that i saw in something like this in anteater You haven't seen an Anteater before, I guess. I'm like, okay. Now he's like, oh, that's an Anteater? That's cool. Oh, interesting. Whoops, you saw an Anteater. Like, where are you living? Right? Now, totally. And that registers with you because it's special. It's novel. It's new. Your neural network gets trained. So would be any AI network. Meaning, if I cannot impress with novel, unique content, chat GPT. or perplexity or any other training set, they're just not registering it. It's just like me showing you yet another cat and you couldn't give a damn about it.
- Speaker #2
And recency matters too, right? On that.
- Speaker #0
Yes.
- Speaker #1
So just going back to the content, let's say you have something in and you have a whole bunch of cats in your content. Human intervention, like if I took that information, and I decided that I was going to bring in my editor to add some information, to change it around. Does that help or does that still, it's just still a whole bunch of cats?
- Speaker #0
Like depending on how the editor is adding the cats. But yes, like essentially it might help because a human has put, and like the problem, if you want to impress, like if you, like we are back to bucket one, how do you? get your brand into the sales channel, church of video perplexity? The answer is we don't really know because the people who built the models don't really know. What they do is they test. If it's something novel and unique, then it's okay. The model is impressed and learns it. Meaning for anybody out there, they need the human interactions. to actually create content good enough to be bored and taken up. Meaning if you are a brand or a community that has a lot of good discussions about your brand, about how your brand is used, then you are more likely to be recognized because humans come up with new ideas.
- Speaker #1
Very good.
- Speaker #0
By the way... And we know this from marketing. So you hire this marketing agency and they don't know anything about your brand. And then they come up with those concepts and you look at them as like, dude, that's not us. No matter how much drugs you take in order to come up with smart ideas, it's still not us. And then they say, oh, you know what? We do user interviews because they want the user input. So they talk to your users and they come back with, wow, you know what your user like about you. And then they make the story based on human input. AI is exactly the same way. We need human input.
- Speaker #2
So why is it then like Norma had this case where there's some backdoor, quote unquote, hacks or whatever right now to get ranked in the LLM. So he can issue. He has a press release company. He can issue a press release company on the right press release, you know, a tier one press release. And within hours, sometimes if you get certain intent based things, if they're included in that press release, starting to show up and be recommended on an LLM for those questions. So that's the recency. It's the authority. Excuse me. But those things are probably short term things and back doors, just like with SEO, there are back doors that will get closed at some point. Yes. So if you seem to be a lot of those right now that people were exploiting for the short term.
- Speaker #0
Yes. So if you go to QueryEdge.com, that's a company I built, QueryEdge.com. This is essentially a geo monitoring company. And fun fact, I started building this. It's focused on e-commerce. I can see which shops get recommended. I knew way earlier, like very early on that. OpenAI is working together with Etsy because for some miracle, they always recommended to buy something from Etsy, right? It's kind of like it comes up in my monitoring. Now, if you now want to impact it, meaning you put your stuff on Etsy, you create millions of pages on Etsy with always the same product and the description in order to impress OpenAI, maybe that works. Up till the moment OpenAI realized this is degrading my effort. Then we are down to an SEO type of arm length race, right? It's kind of... Smart minds like us try to figure out how we can impact perplexity and open AI. We measure, we build it in up to the moment that those tools recognize and then they reduce us. Now, that said, I very strongly believe that there are opportunities to do this. Especially end of this year, open AI will come out and offer you... probably pay to play, right? They offer you advertisements. And now you can actually evaluate what effort pays me more. And you kind of like can start moving between, should I try to place a message with OpenAI or should I rather pay for the message with OpenAI? So you can have, you think about this as a potential triaging. Totally A total opportunity. In the long run, companies who have a brand story, who have a real brand, will succeed. Companies who are just an aggregator on top of Amazon will die.
- Speaker #1
And that's the importance of just being wide, like having a wide brand rather than just doing a press release and hoping that that's going to happen. You know, you have to build that community around the brand and you can do that. And that's another question. The community that we can build with AI right now, where are you seeing that going? Where can we build that brand, that brand community?
- Speaker #0
So this is super interesting. So there is, you have a product and obviously people love your product. As you both know, I'm a coffee nerd. I love coffee and the machine behind me it's a decent espresso machine it's an open source project the machine is amazing it allows me to measure 10 different things with each coffee I shot like coffee shot I create I can buy extremely novel beans or like with taste like melon or strawberry whatsoever, and then work on... the setup. I can download profiles. I can do loads of stuff. It's complicated. Coffee is complicated. And now what you have here is a community that supports each other on usage of this machine. So now you have a community around a product. That's one that has always been there, that still continues to be a big driver for good marketing. Well, there's another way, which is upcoming, and that's the third bucket. Once you have a brand, you can actually create a lean in experience for your brand right um if i buy armani armani as a brand or cartier as a brand or coach as a brand or like you name all the brands they give me a brand identity which i put on how about i i'm allowed to participate in the brand identity because i can now go and create an Armani-like suit on Sora. As much as I could give you guys, if I say, join, like, follow me on Sora at Lutz, and I follow you back, and then you can create my image as a cameo. You're totally fine off doing this. So I give you my brand identity, meaning if I enable a brand identity to be copied, then there is a new business model. There is a new opportunity for me to lean in. as a brand. And let's take Drake and The Weeknd. If Drake would have said, like, by the way, Drake went ballistic against this fake music and says, that's not me. And he bullied everybody to not give the music out. So the only way that you can listen to that AI song from Drake and The Weeknd is actually YouTube. Everybody else wasn't allowed to carry this. Now, if Drake Drake could have said, look, this is my style. I endorse certain styles and you can make money with it. And I, because I endorse it, I get a cut from anybody who creates Drake-like music, right? You can think about a complete new brand experience, which is going to be the future.
- Speaker #2
And you could also create in-person experiences where all the people that have the brand name clothing or whatever can come and we're having a Louis Vuitton event in Los Angeles and all the Louis Vuitton people so you have you have You have that. So that's going to become even more important in an AI world. People are going to crave human interaction more than they may have in the past. And so the world is going to go more experiential in some ways. Is that what you're saying?
- Speaker #0
Yes, totally. So essentially, if you are thinking about a marketing approach and you go like, you can actually take the five Ps. but let's at the moment do... Not the price because it's not helpful, but like the place. We go from Amazon and go into an agentic experience, meaning chat GPT perplexity. You take the promotion. We go from stock photography of photos and text and music we had to create with a lot of cost to a low cost budget thing where quality and control is important. And then... The third one is that you are able to open up the brand to a larger creator audience and you create more of a community around your brand in order to carry you forward.
- Speaker #2
What about on the user side on this? A lot of people are saying the bottleneck is the users getting used to a different way of shopping when it comes to e-commerce, getting used to this aegyptic thing. And it's a different way. It's like going from listening to... to music on, uh, you know, it took a while for some things where I don't know, maybe music's not the best example, but from the eight track tape, uh, to the streaming, um, you know,
- Speaker #1
we're buying CD now.
- Speaker #2
Yeah. There's a progression there where people had to get used to consuming or interacting in a different way. How long do you think it's going to take for the average consumer? When you're a chat, GPT is talking about putting out a browser and, you know, agents built in where they're going to get comfortable enough to where they'll say, go, go, uh, I need some new shoes. Uh, whatever size I bought last time, go find them, find the best price and go ahead and buy it for me and make sure that they're here because I'm leaving town on Friday. Make sure they're here by Thursday. How long do you think it's going to get to where over half of the consumers are going to be comfortable doing that versus the old way of typing in keywords into a search engine?
- Speaker #0
So this is interesting that you asked this, Kevin. I would see this the other way around.
- Speaker #2
Okay.
- Speaker #0
We can talk about how fast the adoption curve was from chativity i mean amazingly fast men first women followed but amazingly fast now i have no question that this is um like that people will shop in a heartbeat they already do i do user interviews like i had one user interview i talked in my linkedin channel about it. One person I interviewed says, well, I uploaded all my blood values from my doctor to ChatGPT. And then I went and it recommended me to buy all these vitamins at this website. Where I've never been shopping before. I've never shopped for vitamins. I've never been at this website. So I just go to the website and buy for $200. I mean, that's reality already today. So your question about how long will it take is way faster than you think. The question is a different one. how long will the ecosystem take to react?
- Speaker #2
Yeah, that's, yeah.
- Speaker #0
And it is so easy for you to go on ShadGBD and have a conversation. And if they offer you buy link, you buy. And like, and then you're amazed that it's there and next time you do it again. That's easy. It's way harder for an e-commerce company to say, I sold through Amazon. Now I need to figure out how to do the agentic. Commerce protocol, I need to figure out what's the privacy thing. I need to build it out. There is where XgenAI is helping, right? So I need to be enabled to go out and to place myself. That's harder. But then a whole industry has to change because the whole industry of Shopify parts needs suddenly are as disrupted. So users will change fast. Companies are changed painfully. Industries are very hard to change. Again, dirty little secret. Let's look backwards. It wasn't the media industry. How difficult, Kevin, was it for you to pick up an iPhone? Not at all. You actually thought it's pretty slick. You got your iPhone and suddenly you could swipe and scroll. Man, were you happy. How hard was it for New York Times to change their news approach? Way harder. And the whole news or media industry almost went dolly up. Same approach here.
- Speaker #1
I've got a question about just... personal habits and routines. I am getting, I've got shiny object syndrome. I don't know which way to look. I don't know which platform. There's a thousand tools coming out at all times and they're all hitting me at different angles and everyone is, look at this, but I don't ever get a chance to become an expert in them. There's certain ones that, okay, I'll do this, this, and this, but there's so many great tools out there. How do you even or what should you be focusing on?
- Speaker #0
Value creation. How about that? Don't do shiny objects. Do stuff which makes your life better. Your business run faster. Your bottom line looks better. It's a dumb MBA answer. I'm sorry, but it is true. People run behind shiny objects. And then the MIT makes a publication saying, all the guys who ran behind shiny objects now saying shiny objects are not as cool. Yeah, because they run behind shiny objects. We have one of the big problems we have using AI is not everything which says has AI is good. Not everybody who's like, I can do you AI X, well, whatever. Does it help me? Does it help my workflow? Does it help me sell better? Does it fit my bottom line? That's very often forgotten. There's one study, which I love. They ask investment managers about how often to use ChatGPT and how much do you like it? So, you know, different investment manager says like, well. I use it every day. I use it three hours a day, whatever. And I really like it. It's so stunning. Or like, yeah, you know, it makes mistakes. It's kind of annoying. It's too friendly, whatever. So they rated those answers. And then they compared it to the investment success of the investment managers. And lo and behold, the more you trust Chachibiti, the worse your investment decision was. Very cool study. Why? Because people don't understand the shortcomings. They don't understand that it is not shiny which counts. It is, does it impact your bottom line?
- Speaker #2
Hey, Kevin King and Norm Farrar here. If you've been enjoying this episode of Marketing Misfits, thanks for listening this far. Continue listening. We've got some more valuable stuff coming up. Be sure to hit that subscribe button if you're listening to this on your favorite. podcast player, or if you're watching this on YouTube or Spotify, make sure you subscribe to our channel because you don't want to miss a single episode of the marketing misfits. Have you subscribed yet, Norm?
- Speaker #1
Well, this is an old guy alert. Should I subscribe to my own podcast?
- Speaker #2
Yeah, but what if you forget to show up one time? It's just me on here. You're not going to know what I say.
- Speaker #1
I'll buy you a beard and you can sit in my chair too. You can go back and forth with one another. Yikes! But... That being said, don't forget to subscribe, share it. Oh, and if you really like this content, somewhere up there, there's a banner. Click on it and you'll go to another episode of the Marketing Misfits.
- Speaker #3
Make sure you don't miss a single episode because you don't want to be like Norm.
- Speaker #2
So is that why you're seeing a lot of companies? There's a lot of stories been coming out like, yeah, we invest in AI, but we're not getting ROI. A lot of these big companies, they jumped on the bandwagon, the shiny object, and they're like, oh, yeah, we've done this, but we're not seeing really any difference.
- Speaker #0
By the way, again, look backwards. We had this before. There was the thing where everybody got excited about it. It was called, help me again, Internet, I think. And then there was this company called Amazon or something like that. And everybody's like, it's amazing. We will shop everything digital and we will get stuff delivered overnight. and We will buy only online and Amazon skyrocket and then people turn around. So it's like, but hold on, I still have all the brick and mortar and it's like, I'm actually not all shopping. And Amazon crashed brutally. It called the bubble. And if you look at the Amazon share price today, that bursting of the bubble is not even visible anymore. If you look at the full scale of where Amazon is today, we are extremely bad. in understanding impact in the long run. And the impact which we are seeing is huge. And I hopefully helped with those three buckets to describe it. But that means still somebody who is selling on Amazon as an aggregator is not going out of business tomorrow. But they will if they don't take the right steps today.
- Speaker #1
So what are those steps? What are like three steps that every e-commerce brand should be taking right now? What in your opinion? Because if you go to them and say, hey, you need to do AEO or GEO or whatever you want to call it, sometimes there's not immediate results off of that. I mean, sometimes there is. Sometimes it's more of a long-term plan. They're like, yeah, but we need to go send some emails or we need to go run some Facebook ads or something because I need to put some money in the bank. So what are like three things that… they should really be doing right now if you're an e-commerce brand?
- Speaker #0
So definitely you invest in a tool to monitor. Query Edge would be one. There are many others. You go and start monitoring. That's number one. Number two is you need to prepare for how do you get the human element into your brand, meaning you build up a community. you um um understand how people search. So Xgen is extremely good in search and discovery. So we help customers and like help brands to distill the knowledge of user interactions into something you can build off. Meaning you invest in community, you invest in customer care, you invest in a close relationship because that is the human element you need in order to impress ChaiGPT and not just showing. more cats to it right and then the third one is you need to experience about how do you lower your cost in brand and marketing and you like depending on your brand you can't even experience with the like like experiment with the lean forward approach i
- Speaker #2
gave you now four steps there we go and we are at the top of the hour so let's thank you so much for coming on i do have a question for you. And we ask this of all our misfits. Do you happen to know another misfit?
- Speaker #0
The biggest misfit at the moment is that people believe that being an aggregator off demand is sufficient. That's the biggest misfit. It's not.
- Speaker #1
Awesome. If people want to reach out and learn more about you or follow you or get in touch with Xgen or... Query Edge or follow you on LinkedIn or whatever?
- Speaker #0
They can follow me on LinkedIn. They can ping me on LinkedIn. They can follow me on Forbes for my Forbes articles. They can reach out to me. They can take my course at Cornell. If you take the course on Cornell, you are not only getting in-depth knowledge about AI. All of my students are afterwards in a big community on WhatsApp where we discuss and build on stuff. And we are over 700 people there like chatting and talking about AI and how to make it useful for businesses for themselves. Awesome.
- Speaker #1
Awesome.
- Speaker #2
All right.
- Speaker #1
Well, this has been great. Appreciate you coming on and sharing. This has been really cool. Really fun.
- Speaker #2
All right. We will see you later.
- Speaker #1
See you later. Have a good day, guys. All right.
- Speaker #2
Thank you. Enjoy your coffee.
- Speaker #1
Oh, I will. uh every time we talk about ai norm i don't know about you but it's like like you said you know your head just explodes like all right what is there's so much going on it's moving so fast and it's like you just got a core like you said what's the shiny objects that you just got a core just drill down to the essentials what is the core things and only pay attention to these little shiny objects if it actually is something very useful in your business otherwise uh you know what what's the point of doing it know about them know about you know if soar 2 or vo 3.1 comes out and it has this new feature that doesn't mean i gotta go run and play with it it just means that okay that's there as a tool if i need to use it uh for somebody or need someone on my team or something to use it but yeah i think the like he said the world is about to change and i think a lot of people out there not just an econ but across across all genres of life don't realize the drastic changes that are about to happen or are going to happen you know we'll look back and these may be some of them will be very quick some will be a little bit more gradual but it's going to be interesting 10 years from now to look back just like it is when you see an old steve jobs uh interview or uh bezos interview from 30 years ago you know it's going to be interesting to see and we're you know it's not going to change kev it's going to be old school yeah it's going to be history repeating itself understanding marketing
- Speaker #2
It's just like Lutz was saying about community. about the brand. You have to have a brand. You have to have that community. And this reminds me, back in 2013, but 10 times faster right now, Amazon. Amazon, all of a sudden, all the service providers came out, you know, around 13 to 2017, and everybody, well, lots of people had shiny object system, or shiny object syndrome. People were getting hit left, right, and center, and it was the same thing. Like, I was talking back in the day to a beginner who had invested $3,400 a month in multiple things. What was he doing investing in all this? He just wasted, you know, a ton of money. You have to learn what, and I think Lutz said it exactly, was, you know, you take your core group and understand it, you know, understand the value. But anyways.
- Speaker #1
Speaking of understanding value, if you listen to the Marketing Misfits podcast, you understand value because every Tuesday we come out with a new episode. Isn't that right, Norm?
- Speaker #2
Yes, we do. And if you want to check out those episodes, you can go to marketingmisfits.co. That's the website. If you want to go and check out the long-form videos, go to Marketing Misfits podcast on YouTube. And we also do something special for you guys. It's three minutes and under clips where you can find them on YouTube. They've got their own channel. It's Marketing Misfits Clips.
- Speaker #1
That's right.
- Speaker #2
That's about it.
- Speaker #1
I think that's it, Norm. That's a wrap. We'll be back again next week. See you guys then.
- Speaker #2
All right. See you later.
- Speaker #1
Ciao.
