Riding the AI Wave: Jonathan Green's Success Amazing Story

What do you get when you combine a best-selling author, a celebrity ghostwriter, and a high ticket affiliate marketer? An intriguing conversation with Jonathan Green, that's what. Jonathan found himself navigating the icy waters of the job market after being fired during a blizzard and turned this devastating setback into a thriving online business. Hear his captivating story about discovering his artistic voice, stumbling into the world of writing, and transforming his penchant for words into a lucrative career.

This episode goes beyond just writing. It peels back the layers of technology evolution and resistance, with Jonathan drawing from his own experiences. From calculators to Napster, we journey through how technology has revolutionized our culture and mindset, the phases of technology adoption, and its profound effects on various industries. Get Jonathan's unique perspective on the current state of AI, and why it's often designed to be an unpleasant experience. Prepare to be amazed, as we unravel the dubious tactics employed by major AI companies.

But that's not all. We also explore the often overlooked user experience aspect of AI tools, and their potential in opening doors for entrepreneurs. Discover the promise of AI image generators and the vital role played by AI chatbots in enhancing communication. Jonathan also shares his invaluable insights on the importance of good prompts and the pitfalls of using outdated ones. Learn from his experiences in using AI tools to optimize content generation, and achieving the best results. So, get ready for a thought-provoking conversation that promises to inspire and guide you on your entrepreneurial journey.

GETIDA Amazon Owes You Money!   Get $400 in FREE reimbursements done for you, follow the link below.

Helium10   50% OFF first month OR 10% OFF LIFETIME subscription = PROMO CODE “FTM”

SoStocked

Start Your 30-Day Free Trial

Your 1st Month Is Free For Any Plan You Choose!

If You receive value from this content please SUPPORT The Podcast

Paypal → CLICK HERE
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
🗣️ TALK TO US ON SOCIAL MEDIA 👇

Instagram ► https://www.instagram.com/firingtheman/

Facebook ► https://www.facebook.com/FiringTheMan

Website ► https://firingtheman.com/
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💥LISTEN TO THE PODCAST 👇

On Apple Podcasts ►https://podcasts.apple.com/us/podcast/firingtheman/id1493680004

On Spotify 
► https://open.spotify.com/show/2mE9YcE5gWtMwsmZUTS84M

On Stitcher 
► https://www.stitcher.com/podcast/firingtheman?refid=stpr
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💻 COACHING 👇
https://firingtheman.com/coaching/
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬

The Digital Revolution Podcast
Welcome to The Digital Revolution Podcast, where marketing experts share their expertise.

Listen on: Apple Podcasts   Spotify

Transcript

Speaker 1: 0:41

Welcome everyone to the Firing the man podcast, a show for anyone who wants to be their own boss. If you sit in a cubicle every day and know you are capable of more, then join us. This show will help you build a business and grow your passive income streams in just a few short hours per day. And now your host serial entrepreneurs David Shomer and Ken Wilson.

Speaker 2: 1:06

Welcome everyone to the Firing the man podcast. On today's episode, we have the privilege to interview Jonathan Green. Jonathan is the best-selling author of 300 plus books, a celebrity ghostwriter and a high ticket affiliate marketer who now lives on a tropical island in the South Pacific. He has turned being fired during a blizzard into a thriving online business. Jonathan is an expert at using artificial intelligence tools to accelerate your online business. He is the best-selling author of chat GPT profits. He has a mailing list of over 100,000 subscribers and hosts a podcast with over 250 episodes. He has been running an online business full-time since February 2010 and is an expert in using chat, gpt, mid-journey Claude and Leonardo to grow your online business. He helps new entrepreneurs leverage AI to generate real revenue quickly and start replacing their 9-5 revenue. He has two successful AI training programs, the first being AI freedom for entrepreneurs and the second fractional AIO for businesses. We're very excited to have Jonathan as part of the Firing the man podcast. Welcome. Thank you so much for having me.

Speaker 3: 2:17

I'm really excited to hang out with you guys.

Speaker 2: 2:19

Absolutely, we're looking forward to this and it could not be more relevant. So a lot of good stuff here we're looking forward to diving into. Before we get into technical AI, as in relating that back to our e-commerce businesses, can you share a little bit with our listeners about your background and your path to becoming an author and AI expert?

Speaker 3: 2:40

Yeah, it's always been kind of a roundabout path. I wish I was one of those people that could tell you I was an entrepreneur at 15. I made first billion by 18, 10 million by 23. Like, some of the entrepreneurs have that story, I was one of those people that was always doing entrepreneurial things and didn't realize it till later. So much of my journey was really trying to be artistic. So I was acting a lot in high school and then in college I dabbled at painting and almost my main instructor of painting was almost like you should think about making this your major and it was really awesome to hear. But at the same time it's like just experimenting at different things. And then I tried a lot of careers in my 20s, as everyone does, and you're trying to figure out what you want to do with your life and you have so much bad information. Like the first thing you find out when you graduate from college is that they're not actually going to help you get a job. They're just going to take credit, no matter where you land. You work at Starbucks, it's our job. You work in a garbage truck? They're going to say the ones you got either. It doesn't matter what you do, they just want to say what percentage people are employed and it's just luck. So that was like a really rude awakening for me, because I went to the hiring office right when I was graduating. I was like, hey guys, I need to find a job, what am I going to do? And they go here's a link to monstercom and that was the hiring website of them. I don't even know if they're still around, but it was like so unhelpful. And then I said how many of you guys graduated from this university? And they all had. And I was like, how many of you guys have worked here since you graduated? And I was like, oh no, I'm in a room full of losers. So they didn't want to help me. After I said that I really burned that. But it's like sometimes you just have to say what you see. And I was really shocked because it's like it was when I was really. I was in college for eight years. It's like you were still a freshman, right. It's like you're not a doctor. That means you haven't really used the time well. So I struck out on my own, tried a lot of different ideas and started. You know, I went through. I was teaching for 10 years around the world and teaching at universities. And then I got my master's degree and some publishing house was like, hey, we want to publish your dissertation and it's probably a scam, I don't know. But it's also like who's going to buy a dissertation, I don't know. They probably got it put in a couple of libraries. But either way, it was like my first taste of that. Just was really passionate about writing as a way of expressing myself. I started blogging around 2007, not knowing that people could find your blog. I thought it was a private diary you do on your computer. And that's not true, because my blog became one of the most famous blogs in the world in the dating space because I would just write about my eternal struggle to get a girlfriend and that made me the number one search result in the entire world on Google for get a girlfriend for like three years. So every guy who typed in get a girlfriend would read all of my stories about how I couldn't do that. And it became this really epic journey that a publisher called me. They just cold called me one day and said listen, would you write a book for us? And I was like that sounds amazing. And that was my first taste of making money from writing something because my blog was making no money. There's this. There's a lot of misinformation that like, oh, my blog has, every blog that comes in search of it makes a bunch of money. It's not true, especially in dating. And it was this great opportunity. I started, got a taste and I was like, wow, people will pay to read the stuff I've written. That's amazing to me. And once you have that revelation, then so many things become possible. And then you know I thought I had my dream job, which I hated so much. It's like, once you get your wish, it's the worst thing that can happen to you sometimes. And I was working at one of the top 20 universities in America running a multimillion dollar program at 29, in charge of all these people that are older than me, and I hated it. There's a skill people get, and you guys have been in corporates. You may have encountered this. You definitely encountered a bureaucracy and government, which is people who can really make a small task, take months and really impressively slow roll, and I don't have that skill. So they gave me a task like, listen, you need to work on this for the next six months. And I finished in two hours and I almost started crying because I'm like I don't know how to do this. I don't know how to stretch. I can stretch a two hour job in eight hours. I couldn't do that but I don't know how to stretch it into 180 days. And that's really what it's all about. Especially, higher education is mass. I mean that's why they keep raising intuitions. There's a huge amount of just, let's say, unethicalness going on in there. The quality of education is really plummeting. The quality of everything is really down and everything you work on there's so much politics and I hated it so much. So when they fired me, they were like we're destroying your life. And I was like you guys just freed me because you forced me to do what I was always meant to do and I've been given that advice. Like, I wonder the people, I was a private tutor in London for a Kazakhstanian billionaire, and if you become a billionaire in Kazakhstan, let's just say you do some stuff. And he was like you should switch and go to the MBA program because you're not going to be a teacher forever. And then the second thing was do you want to come visit Kazakhstan with me to hunt wolves? And I said like with the spear, he goes no, you dumb American, with a snowmobile and AK-47. So that's what I realized. Like people see the world in a different way. I like I didn't even know that's how you hunt wolves. I thought I like that doesn't seem very fair. And but he was right. I just didn't listen to him and he was the kind of guy that would tip you with cartons of cigarettes. That was also a big part of how he did business. He just had just like a bag of cigarettes with him at all times. So I just didn't listen Because I always thought no, it's for other people, because we're so taught in our culture that you can't do it. Only people that start companies me was parents starting companies Only people that can succeed or like generation of wealth. No one can start a business. Now, and you hear that enough times, you just believe it. You think, oh, I'm always meant to be an employee, I meant to be a cog and machine and you have this idea in your head that if you work for a large enough organization you'll be protected. And, as everyone found out over the last couple of years, it's not true. Large companies all the time have higher firing things where they just fire 10% of people performance or element Right, large companies go out of business because the guy at the top did something naughty. It doesn't matter how much loyalty you have for your company. At some point you're going to find out. For some people it's at retirement Right, when there's a retirement party. The next day they never hear from any of their coworkers again. You find out oh, they're not my friends. Or for some people, it's when you're in your 20s. I've had so many friends that get fired on Christmas or the day their child was born or when they were in the hospital, and it's. You just realize that companies and the organizations are always looking out for themselves. So you have to start to see the world as it is. And so much of education is like this. It's like a carrot on a stick in front of a turtle. Right, you're going to work really hard in grammar school to get into good middle school. Work really hard in middle school to get into good high school. Work really hard in high school to get into good college. Work really hard in college to get a good degree. I have a really high degree. I have a master's in applied linguistics with merit, which is a bit equivalent of honors in America. Guess what? I never use that thing. I also never use anything from college. Everything I use to run my business I learned in high school, so I spent hundreds of thousands of dollars and years of my life all learning stuff that's relevant. Like most majors, are trash. Most people who teach at universities could never survive in the real world. You meet all these people that are teaching. You know, like I was working on a program and I met people who were like, oh, they teach my masters in marketing. I'm like oh, what kind of marketing campaigns do you work on? You found out they've never. They've never left the university. They've never actually done it Right. They're just people who watched. And it's like learning to swim from someone who read a book about swimming. If you do that, you're going to die. So that was really my journey starting to realize that so much what we've been taught is just a big ol' lie and that you're capable of more than you realize. That's really the core message I want to give people and that I've learned from my journey is that whatever you think you can do, you can probably do 10 times more.

Speaker 4: 9:26

No, that's awesome, Excited to have you on the show, jonathan, I'm excited to dive into AI. I can tell you, a great storyteller. I think it comes from your writing and from all of that, and so that's really cool. A funny story there was a guy that used to work with David and I and he lived in Pakistan and he was talking about hunting. I had the same image like oh, we're going to like Rambo style, like we're going to get like a spear and we're going to, and he's like no, with a gun. So here in that spot I was going to be like yeah, yeah, right, so pretty funny. So to get into AI, I do research on the show. I looked around on some of your books and some of your background and I saw a quote that I just kind of want to go for first and have you dissect it for the audience to like I think it's very impactful. And the quote is within two years, there will be those who have mastered chat, gpt and those who are out of work, and I think that's a very true statement and I think it's very impactful. So can you kind of dissect that? And what do you mean by that, by that statement?

Speaker 3: 10:18

So at certain points over the last 50 years, technologies have come out and people resisted against them. In the early 70s, when calculators became affordable, teachers were like kids will never learn math and we can't imagine a school child not being allowed to use a calculator. Now, right, it becomes ubiquitous. After those, it was graphing calculators when you have to use sign and co-sign and tangent, and when I was in school you had to learn how to use those all the time. And of course, then there was someone to figure out how to design games that would go inside of those and those would go around. And I had a friend who wrote an entire video game inside of a graphing calculator, like, amazing, there's all these things. But we say, oh no. And I remember, as I was in college, in 1999, napster came out, which was this music sharing platform, and everyone was fighting against it. They sued everyone. People got sued for downloading music by record companies. Every band fought against it and there was a huge struggle to fight against the technology and they failed because now nobody owns music anymore. Right, they were like, no, everyone wants to own their collection. I felt the same way. Now you're just so used to renting and streaming that our mindset and the way culture has completely shifted. Once the ability to download music became there, the businesses that succeeded were the ones who said okay, how can we monetize downloading music? That's really where the future becomes, and it's the same thing. We just saw this huge writer strike, and they you know they got some things that said, okay, you have to. You can't use AI to write movie scripts, and if you have make a writer use AI, you have to tell them that you're giving them something from AI, and it's like that's not going to work. You can't put the genie back in the bottle. Imagine, though, you meet someone you're trying to hire a new accountant you guys know this and you meet an accountant who goes yeah, I don't use calculators. You know what? I don't use spreadsheets. You would never hire that account, right? Even though 50 years ago, that was all accountants, right? Nobody had it calculators, nobody had spreadsheets. What happens is, once a technology becomes ubiquitous, it goes through several phases. First is the phase where everyone's afraid of it, then there's the phase where it's optional, and then there's the phase where it's mandatory. You literally cannot be an accountant who doesn't use spreadsheets, right? You can't be an accountant who goes. I don't use accounting software, I just do it all in a book, unless you're working for a criminal enterprise who wants no records. That's not going to work. You could such a sh. It's a change. What's happening now is that there's so many people who think this is a fad. They go oh, this AI thing will go away. And there's there's legitimate reasons why they think that. Right, there's a lot of bad information, fake AI companies and what. All the companies have kind of gotten together and they had a meeting and they said what if we just create an amazing tool with no onboarding? You log into chat GVT and it's a blank page. It's literally the worst thing you can find it. As a writer. That's my nightmare. You lock me through with a blank piece of paper to type, write a book about what I'm not going to tell you, but you better write the right book. That's what it feels like. There's no welcome video, there's no onboarding video, there's no email sequence to tell you different things that you can do. If you want to know what chat GVT can do, open AI will not tell you, which is such a crazy business model. It's why everyone uses a tool once gets insemmulated and leaves they forced a necessary market, which is the training market, because they were used to teach people what their tool could do. They refused to tell you its limitations. They won't even tell you how many words you can put into a question. They came up with this idea which is like oh, we don't use words, we use tokens. And my experience of tokens is, every time you go to the arcade they tell you it's token, so you don't realize you're paying 27 sets of games instead of 25, right, they're just messing with you. So they've created this entire system which is so unpleasant and such an unpleasant experience. And then the other AI companies saw this and go we're going to do the same thing. They didn't go let's make an amazing onboarding process and steal their customers. They said no, let's do price fixing. Let's all give everyone a really bad customer experience. And I imagine this meeting where the CEO of OpenAI was like all right, number one, this is going to be what people see just a blank page. And someone's like why don't we tell them what chat GVT can do? And he goes you're fired. And the guy goes what about a welcome video? He's you're fired. What about just an email that lets them know their language use, like what a prompt is, what pre-prompting is? And he goes listen, you're all fired, there will be no onboarding. And that's what they've done. They're literally the worst onboarding product in the entire world. And then Reflexi AI was testing them the other day. I was like, oh my gosh, these guys thought that was a cool idea. They go. We're also not doing onboarding. In fact, when you joined, we won't even send you a welcome email because that might accidentally educate you. So they've all done this. All four of the big AI's are playing this game and it's so weird. And it's the reason when people are like, oh, I'm intimidated by AI. You should be. It's designed that way. It's designed to be an unpleasant experience. Now, ai, once you get in there, it's an amazing tool, but for some reason, they've made the learning curve so steep. So when I say that it's not optional, it's that all, despite all of that stuff, it's so powerful and you can do so many things. They did. A study came out today that said people use AI, are 40% faster. I've always thought it was 20% and I'm like, oh, I got to get faster. They're accomplishing tasks, they're learning things, they're being able to get to their work. It's not a replacement for people, but it's an accelerant. It's like putting on the Iron man suit. So right now we're in the phase where it's optional and there's still some people like going on strikes saying don't make me use AI. All that's going to happen is a bunch of non union riders are going to steal their jobs and then the union won't matter. Right, like, oh, this guy's not in union, they can write 50 movies a week. Of course that's what they're going to do, because business is business. Business will always follow profitability, just like a river follows wherever the easiest way for the water to get to the ocean is. So you could resist against it and you could rage against, and you can say I'm too old, I'm not going to learn this. But here's the problem. Let's say, in three years, you guys are thinking about hiring someone to edit this podcast. You have two choices. Well, one guy goes hey, I can edit 10 episodes a day. The guy goes I can edit 10 episodes a week. I don't use AI. There's no way for you to have this conversation. You go let's hire the guy who doesn't use AI because we're going to pay 10 times as much money, but he seems like a nice guy. That's. The problem is that you can't compete because it's just someone who does the same job faster. It's not as though there's a drop in quality. It's the opposite. The quality maintains because there's a consistent element. Ai is consistent whenever level it's operating. So you have this really strange situation where people, because there's no onboarding, because there's no welcome materials, because there's no training, don't know what to do with an AI and so they think, oh, it doesn't work. The reason I wrote my book I read a couple of books that were selling really well about AI and I read it. I go oh, these books are full of lies. There's a huge amount of prompts in both these books probably 80% of them that don't work. They're from two generations ago and they didn't even work. Then I can tell they had JatchEVT write the book. So there's a huge quality issue and they provide a bad impression. It really frustrated me. That's why I wrote my book, not because there needs to be another JatchEVT book, but I was like so mad that the number of people whose lives will be destroyed because they'll read this book, try a problem, get a bad response go, I guess AI doesn't work. And then two years from now they're fired, but they go. No, I tried, and because some guy who has a lot of fake reviews on his book on Amazon also gave me this bad information. I was tricked and that kind of drives me crazy. It's like there's a lot of companies that just add AI to the title. Now they're not AI companies, though there's a lot of distractions. It's totally legitimate for people to go AI doesn't work or these tools aren't really AI, or my experience was bad and they kind of reject it. The problem is that it doesn't matter, because the market isn't going through a major shift right now and the shift is going to happen, whether you like it or not. You can go on strike and resist it and say it's not going to happen, but doesn't every McDonald's in the world have one of those TV menus now that people were afraid was going to replace workers, like they've already started doing this?

Speaker 4: 17:02

stuff. Why do you think there's? No, I agree with you. You go to open AI or there's very little training, there's no onboarding. So why do you think that's? Is that by design? Is that something they skipped? Why do you think that's? Why do you think it's like that?

Speaker 3: 17:14

I think everybody works there's good at one thing and not good at anything else, like most people who work at AI companies you wouldn't want to be friends with, like they're great at programming, not great at anything else Like. So I think it's that. I think literally. Here's another example If you register a business, right, you have to register your local business on Google business, so you start to appear on the map. There's no option for AI. Google has a huge AI division. They have AI software. You could be someone who works at Google in the AI department and you can't list your business because it's not a category they have. There's AI website builders. None of them have the category of their own company. So it's not that it's intentional, it's that every one of these companies has. I guess I mean I don't know the right word for it idiots or dummies like they don't even consider. Oh, what can we do? We have the category of our own business available for our website builder. Do we have our own? Can Google list their own business using their own like location tool? No, so they aren't even adding in these categories and it's not crossing their minds. I think it's really that they go. It's one of the things where, like the tool is so powerful it speaks for itself. And for most people it's like getting in a plane and all the switches are white. So it's even worse. Right, all the switches are black. You know there's switches ever. You have to find them to fit and then you have to guess what each one does and they're like all right, fly the plane. Yeah, I know the plane is powerful, but I'm going to figure out by guesswork and so it's not intentional, it's the exact opposite. It's what you get with group think. Right, you get a bunch of nerds that are all really good at it, and I'm in that category, so I'm including myself where it's like sometimes we forget and this is very common for me that all explaining something and it's so advanced ever in the room is like you lost us like 15 minutes ago. That's the problem, and it's not just that you go to AI image generators, like, especially in our those guilty of this. The Naruto has a default it doesn't. It never makes a woman that you can use for a business picture they put it that way for some and you go. Why does every woman look like an anime woman who works there, like you're quick? That tells you a lot about people who work there, because my assistant was making images and I was like what do you? I was like would you stop prompting? What do you? Whatever you're writing, I need a picture of a woman at work. Why is she so unbuttoned? He's like he was and he was like it's not me. Look at the prompts. Because I thought he was doing it because we're putting Pinterest, which is mostly women platform, and I was like we don't use that kind of content. And then I looked at the feed. I go, oh, they've programmed it that way and I was like is this only, is this software only for 14 year old boys? So they a lot of these tools tell you. So it doesn't tell you about the user. Tell us about the people that work there that they only fed the model certain types of images, or they only better start to say, the data. So you always have to remember that their thumb is on the scale and they create things that match the way they see the world. And so they're like, oh, I know how to program. And when you're programming, you start with like a blank page, you start writing code, so they all do it, and that none of them have thought of it in six months and it's kind of crazy. But I don't think it's intentional. I'm not going to match like it's like Occam's razor. It's either they're a bunch of idiots who are really good at programming and didn't think about the user experience, or they're a bunch of geniuses who are playing 5D chess and thought, hey, what if 86% of people never use our tool and another 10% use our tool once and disappear? That's the perfect business model. I can't imagine that's intentional. To me that spells opportunity.

Speaker 2: 20:37

Absolutely. Now, Jonathan, I want to address the learning curve piece of this. What are some recommendations that you have on how entrepreneurs can learn and use AI tools?

Speaker 3: 20:47

So the first thing is that 90% of AI tools are using the chat GPT API and the other 10% are using the cloud API, and then all the other ones are doing art or video generation, are using one of the two image generators. So there's not really a bunch of different AI tools. There's about six. There's one from Facebook, one from Google, one from Microsoft, who owns open AI for text generation, and then for image generation. There's two. There's Dolly, which is owned by open AI, again Microsoft, and then there's stable diffusion, which is open source. Every single image generator out there, all these different companies. They're stable diffusion. They're using a free tool and they're adding something paying on top of it. So it's easy to get distracted and go. I need 50 tools. I need 30 tools. You don't you need two. You need one text one, which can either be or Plexi AI, claude Lama, which is the Facebook one, or chat GPT. You just pick one and then you'd want image generator. With those two things, you can do just about everything, so you don't need to spend a lot of money. What's really been interesting to see one of the things I like is that a lot of software that was really expensive a year ago has really the prices gone down because open, because chat GPT is 20 bucks a month for the pro version. All these other softwares that were charged like $100 a month, $200 a month for larger accounts, charging you for the number of words they generate their AI Now they can't do that anymore. So it's done a really good thing to the market, which is forced overpriced products to become affordable for normal people Very, very cool. The second thing is that when you're interacting with chat tbt, here's this will change the game for you. This is how you become a top 1% user in less than five minutes. Whatever you want, you tell chat tbt and then you tell it to ask you questions. So instead of saying, instead of trying to give it information, and say, tell me my ideal customer, create my customer avatar, here's my info. Instead of saying, chat tbt, I want to create my customer avatar, please ask me questions one at a time until you have enough information to give it to me. And when you ask it, in that structure, you remove garbage in, garbage out. The first thing I learned as a kid was that when you write a bad question, you get a bad answer, but if you could switch it to conversational. It removes the risk of a mistake. So you can say I'm working on a new e-commerce store and I want to write really good product descriptions. Please ask me questions so I can develop a product description template that matches my industry. It sounds like kind of crazy at first, but you're just telling it what you're trying to get to. It's almost like something that's jeopardy, right? You always start in the form of a question. I always tell it one at a time, because it will ask me 10 questions at once and they're always overwhelming. But if you get one at a time, it feels you can handle it and usually it doesn't need to get through all 10. It will show me 10 questions and by one at a time usually the question is seven or eight it gives me a really good answer. So that process is really the process of mastering. Prompting is to just start with a question, because when you start with a command, you really limit it to doing one thing and if you've made a mistake in your command, it's not allowed to tell you that. It's not allowed to go dumb prompt. It's not allowed to hurt your feelings. So it can't tell you that you've made a mistake. So you have to start with the question. The start of the question will eliminate 99% errors and it will put you in the top 1% of users. Just one change.

Speaker 4: 23:23

Yeah, that's high level. I've not heard that. In our company we use a tool called AI PRM, which I really like but what you just said kind of negates. So AI PRM for the listeners is it's basically a prompting generator tool that has like hundreds or thousands of prompt free program prompts that people use and then bubble up and then they select them.

Speaker 3: 23:43

And so you're saying, is that tool okay to use or would you do you prefer to use to go with the question first- so when I craft a prompt that I give to my customers, you give to someone in one of my books, it starts off as a question and then eventually you can make it into a really good problem. But it takes a huge amount of research, a lot of testing, that what you get is a consistent result. And the cool thing about the question format is that eliminates the possibility of there being an error in a prompt. One of the things that can happen is someone writes a great prompt for chat ABT for then they release version 4.5 and the response is modified and if the problem doesn't get updated you might not notice. But it's like 3% off or 10% off. I'm not against using pre-written prompts I do it all the time but it's hard for someone is a novice user to detect Rift, which is where chat to start to slide off track. If it only drifts 5 or 10%, you might not notice. That's the danger. However, having pre-written problems and having it helpful is good. There's a problem in the market which is prompt collections. I saw there's an arms race for prompt collections. I saw one that now comes with a million prompts which mathematically, you can't test them all in your lifetime. So what you have is a collection of prompts that you'll never use, but just they're just in case and it's like oh, I have a parachute, what's inside the backpack, I don't know. I hope it's a parachute. And then, of course, there's a good chance you find out it's an anvil too late. So I really believe most people probably need 20 to 30 prompts in their collection that are just they use every single day, that are very efficient for them. Like you guys might have, we do show notes for every episode. You just have an AI. You have one prompt for that right, so you're going to use it all the time because it makes sense for a podcast. But it doesn't mean you also need 58 prompts about SEO, because SEO changes every three months, so any prompt from three months ago is going to be out of date very quickly anyway. So there's much, much better to get a smaller collection of really useful prompts to use all the time and know they work and start to get a master with them, and I always tweak prompts. My challenge is that this is how I got really into chat Every time someone, I would see someone else's prompt, I go, oh, I can fix that, I can make that better, I can take. They do it in three prompts. I bet I can do it in one. I'm always chasing the one prompt solution, like can I give someone a single prompt that does everything they want, so that they don't have to do anything? That removes that part? But the question is the best format because it just eliminates the possibility of mistakes. So you can still start from those prompts, because it kind of helps you craft what you want and so you're just really saying this is what I want. What information do you need to get there so that we can get a really good result? So you're just kind of making it cooperative. Most people you talk to AI's like they do on Star Trek. Right, you go do, do, do, do AI, do this, do, do, do, do, do, do that. But if you use what it says in the name, like the chat is in the name, but we all ignore that, right, we still shout at the AI and kind of berate the AI and order it to do things there's just this challenge and communicate with it in a more casual way. And the more friendly you are and the more casual you are, the better the results you get.

Speaker 4: 26:20

Two follow-on questions. First one is I think this will be a quick one what is the difference between, say, chat GPT-3, chat GPT-3.5, chat GPT-4.0? Are they adding a larger database or are they improving response times? What are, do you know, the updates for those?

Speaker 3: 26:38

There's a lot of secrets, you know they consider them to be that they're precise. They don't want you to know the recipe. What I can tell you is that the language model, the number of data it's trained on, is larger each time they release a new model. So to train chat GPT-4, I think it took like 44 days. So the AI is just reading data for a huge amount of time. Obviously, that's just a bit and you don't. You don't want that. It's like when you're like uploading a video. You don't want it on day 43 for it to glitch out like that's the nightmare. You have to start over again. So that's how it works. So when AI model is being trained, it's trained in a single session. So, hey, chat GPT-3.5 is faster because it's using a smaller dataset. So when you use four, you know it writes a little bit slower. It's a slower answer because it's more methodical as pulling for more data. So there's a lot of open source AIs and all of those are always named with a number at the end, like 7B, 170b, and that stands for how many pieces of data. So it's 170 billion, 180 billion, it's in those ranges. You know. 7 billion, 180 billion is kind of what people do with open source AIs and basically one each billion is about a terabyte of data, so it tells you how much, or a gigabyte of data I mean not terabyte so it tells you how much data you need on your hard drive to be able to run the AI. But as you have a larger dataset, it takes it longer to scan everything and pull through. So the challenge with the larger datasets is that, of course, you need either a faster computer, which is always hard right the race between data and computer speed, or you have to find some way to create efficiency. So a lot of what secondary tools and a lot of what they're chasing now is not a smarter AI, but a faster AI. So I think that's what they're trying to do. They're dabbling, they've released some things that I think are version 4.5, where they're testing some ideas with oh, let's add some efficiency models, and those are a lot of people are creating those in the open source world. So there's some interesting things happening there, but I believe that the most important thing is chat. Tpd4 is better. It's more accurate, it does better coding, it has a higher success rate, but since they launched it, the quality has gone down almost every single day, because every time someone figures out something bad to do with the AI, they have to add in a new rule to stop that. So it's like at first, 40% of the time it wrote code, the code would work on the first try. Now it's at 25%. So they have to release a new iteration as they find out what people do that's bad, and they're just trying to figure that out because it's very challenging for them if something bad happens because someone learned something bad from their instruction. So they're in this very complicated place. That is a challenge because we want on Saturday, we want freedom of speech. We don't want it to actually participate in a crime. So that's kind of what's happening with the different iterations.

Speaker 4: 29:07

Yeah, no, that's first time I've heard an in-depth answer like that. I appreciate it. One more follow-up question. I'll kick it to David. So this will be specifically for our listeners that sell on Amazon and so we use chat GPT to write our Amazon listings and copy. What I've found is that the prompts that I give it, it gets confused if I give it too much. Now you said earlier like, hey, I try to take something that took three or four prompts and make it into one. How do you do that? Let's just say for us. I ask it like hey, I tell the AI hey, go and research this other competitor's listing, or maybe two other competitor's listings, and write one. That's better than write a title, that's better than that. And then I'll let it do that, and then I'll tell it to do the same thing for a description, and then I'll tell it to do the same thing for bullet points, because if I give it all in one prompt it seems to get confused. And so maybe it's the way I'm setting the prompt up, or how do you like? Is it syntax or what do you think?

Speaker 3: 30:02

about that. There's a couple of things that you can do that'll make a big difference. The first is that the way the AI works is that it waits all data equally. So when you ask it about an Amazon listing, you're saying you approve this Amazon listing, you're going okay. The first somebody is looking at every Amazon listing for every product that exists. That's in my database and it scanned everything on Amazon for September 2021. That's, it's correct. Most listings on Amazon are bad, right, most products sells zero. Most websites are bad, most sales pages are bad, but it can't tell. It can't tell a sales page that has sold a million dollars as a next to one that sold zero. So what it does is it waits everything equally. That's the first thing. So if you limit it down and there's ways to limit, the first way to limit it is to kind of limit by category. You start with a pre-prompt, which is before I give it information. I'm setting the AI limitation. So I don't want to increase the amount of data. I want to decrease it first so that it just becomes someone who's a specialist in Amazon listing. So you say something like you're an Amazon listing expert who's been working in e-commerce and writing copyrighted for e-commerce for the past 30 years. That sets it as in I'm a category expert. Even better is if you narrow it down to a specific person. So Chatchabee knows who I am. I was testing with it a few months ago and I said write in the style of Jonathan Green. And I was like, well, that sounds like me. That's weird, because I thought it would use because I'm like the seventh most famous Jonathan Green. There's the guy, a famous painter. A fault in our stars was written by Jonathan Green. I assumed that's who it would use, right, the guy who writes these teen angst movies where one of the kids always has a fatal disease. No, and then I thought I would use the science fiction author, jonathan Green. It didn't. It went to me, maybe because I was writing in my space. It recognized me and it kind of gave me this revelation that you can, I can say, oh, just write in my style and it immediately starts to sound more like me. So if you're already in there, it's like a really good benefit. But it also means anyone else can sound like you too. So it's the strange place. That's why a bunch of authors are suing it, but that's kind of a rabbit hole. But what's important is that once you give it a narrowing, to even go more narrow, is you say, hey, this is right, listings like this specific copyright, you're this person, right, you're this particular copywriter. Now you've really narrowed it down to a single data set of someone, because it has all their data and it goes oh, and all their sales pages are good, right. So, like if you limit it to just David Ogilvy or some, like any specific copyright, now it's limited to a specific person who's actually good, rather than the category which has good and bad people, right. So you narrowed it even more. And then, when you ask your question, if you want to get a really good answer, you always want to make multiple choice. So what I always do is say make a table and give me five titles and five descriptions and I'll mix and match. That gives me something that I can mix and match as a master. It's much better to get five headlines and you choose the best one, and then you get five descriptions and choose the best one. That will always give you a superior result. And if you tell it to organize it in a table, then it won't have a problem with larger amount of data that you're asking for. So that's how you would two prompt it.