In this episode, Guy Kawasaki, author and Chief Evangelist of Canva, discusses his journey into the world of Artificial Intelligence (AI), specifically generative AI, and its profound impact on his work and personal life. Kawasaki shares how he was introduced to generative AI and his fascination with its capabilities, such as generating precise responses and improving efficiency in his tasks. He introduces KawasakiGPT, a custom Large Language Model (LLM) built from his writings, videos, and podcast transcripts, which he uses to automate responses and draft writings, enhancing his productivity. Furthermore, Kawasaki details how AI has assisted him in diverse scenarios, from crafting effective letters to insurance companies to brainstorming book content, emphasizing AI's potential to amplify human capability and creativity. The episode also explores the philosophical implications of AI in learning, entrepreneurship, and personal growth, encouraging listeners to embrace AI tools to remain competitive and innovative.
00:00 Meet Guy Kawasaki: Author, Surfer, and Tech Evangelist
00:21 The Joy and Connections of Surfing
00:43 Exploring the World of AI with Guy Kawasaki
02:11 The Practical Magic of Generative AI
04:30 Personalizing AI: The Creation of KawasakiGPT
05:46 The Evolution of Writing with AI Assistance
08:36 Harnessing AI for Creative and Research Processes
10:57 The Art of Selective Innovation and Storytelling
15:47 Learning from Legends: Steve Jobs and the Art of 'Stealing'
17:21 Exploring the Limits of AI with Steve Jobs GPT
18:57 The Power of Dialogue in AI Interactions
19:26 Real-World Applications of Chat GPT
22:14 Learning from AI: Apologies and Communication
25:26 AI in Negotiations: A Personal Victory
30:24 Enhancing Entrepreneurship with AI
33:11 Reflecting on AI's Impact and Future Possibilities
📜 Read the transcript for this episode: Transcript of I’m Immortal, Baby!” Guy Kawasaki on How to Leave A Legacy Via AI |
[00:00:00] Guy: Hello, my name is Guy Kawasaki. I am the author of a new book called Think Remarkable.
This is based on a podcast of mine called Remarkable People. And I also have a day job, which is the chief evangelist of Canva. . And I live in California and I love to surf every day. And so I'm going to tell you how to make a difference and be remarkable in this episode.
[00:00:26] Jeremy Utley: are the waves today?
[00:00:27] Guy: The, you know, for me, my sweet spot is one to two feet. It's fine.
[00:00:34] Jeremy Utley: And has your daughter now far suppressed you in terms of her abilities?
[00:00:38] Guy: Oh, by orders of magnitude, yes.
[00:00:41] Jeremy Utley: The other thing I really enjoyed about your book was you talking about all of the connections that come from an activity like surfing, for example.
It's very cool.
So guy, when, when Bill introduced us, I was wondering, okay, which, what do I want to talk with guy about? Cause we could talk about the +book, uh, of course. Um, but having been on a book tour myself and not in the not too distant past, I know that that can kind of be, uh, it can feel a little repetitive.
And I coincidentally, I coincidentally asked Madison, is Guy into AI at all? Because, because Henrik and I are kind of obsessed with AI and she wrote me back immediately. Oh my goodness. He uses AI for everything. So I said, okay, this, let's make a conversation. We can talk
[00:01:23] Guy: the whole time about that without even mentioning the book.
[00:01:27] Jeremy Utley: , maybe just by way of background, How did you get introduced to generative AI and how did you begin to see the potential of it yourself?
[00:01:36] Guy: I just read about it. I, it's not like anybody from Google or, any of those companies called me up and said, Hey, let me give you a VIP demo.
So basically they said, ChatGPT is open, go sign up. And I did, and I loved it.
[00:01:51] Jeremy Utley: What, and what was your first experience that like, what did you love in those early days?
[00:01:56] Guy: I, you know, it's kind of like, um, it's kind of like Google or social media. The first thing you do is search for your name, right? I guess it got my name, right?
So I said, Oh, this must be smart.
[00:02:07] Henrik: I also learned it just from some friends who was like, Oh, this is pretty cool. And I got dumped into the playground and I was like, okay. It reminded me of like when I was telnetting using the internet for the first time and kind of like tried to connect through gopher and Veronica, whatever they were called.
Why do you think there wasn't any hype until it was all the hype?
[00:02:28] Guy: Yeah, I mean, that was kind of a, you know, the sun broke out one day, right? And, well, I think prior to that, the actual use of AI for most people was so sort of esoteric, right? I don't know, it's going to figure out fuel management optimization, or it's going to figure out, you know, how many heads of broccoli you should buy to manage Olive Garden.
I don't know, you know, whatever, whatever it was doing. I don't know. And then all of a sudden it says, okay, now, you know, you can ask it questions directly, and it could replace your search engine, or it can help you write or do research, which is very different than optimizing factory floor operations, you know, theoretically.
[00:03:13] Jeremy Utley: Do you remember early uses that felt to you unique? Effectively Googling your name isn't, that's not a unique experience. Was there an experience that kind of demonstrated to you the capacity of the technology that you go, Oh, there's, there's something different here.
[00:03:29] Guy: Yeah, I, um, well, I remember my, My first, one of my first experiences was asking it, did Steve jobs finish college?
Okay. And chat GPT got that dead wrong. I said, yeah, Steve jobs finished read college in Oregon, you know, blah, blah, blah, a certain year. I said, I know he didn't do that. So chat GPT is bullshit. Um, but then I kept asking and trying to train it and eventually got it right. Let's take a real world question, right?
So you get a new HP printer and you cannot figure out how to put it on your Wi Fi network. So if you go to the old Google, you type in, Add HP network to Wi Fi, uh, add HP printer to Wi Fi network and you get 750, 000 results, you know, and 200, 000 YouTube videos. And all I fricking want to know is how do I put this printer on the network?
And now with Gemini or chat GPT, it doesn't matter. Gives you, you know, okay. So step one, go to system preferences. Step two, select, you know, printers and whatever it is. And that's what I wanted. I didn't want links. I wanted answers. So that was kind of the first kind of aha.
[00:04:39] Jeremy Utley: That's great. That's great. And it speaks to you in language.
You understand, right?
[00:04:43] Guy: Yes.
[00:04:44] Jeremy Utley: That's, that's one of the amazing things is you go, no, no, no. I'm a Mac user, not a PC user. And all of a sudden he goes, Oh, sorry. You need to hit this link and hit this, right? Whereas there's no conversation. You just got to, you know, backtrack and Google, right? It's like, Oh, I'm clearly on the wrong page.
Right?
[00:04:59] Guy: Yeah.
[00:04:59] Jeremy Utley: Backwards to go forwards again. But in a conversation you can really easily redirect and say, Oh no, no, that's not the net. Yeah. Network name or what I'm seeing is this. What do I do? And it's much more like a, you know, almost like a human. Yeah. Yeah. Yeah.
[00:05:15] Henrik: By the way, like a cool trick. We talked to Ethan Mollick the other day and he had a nice little trick about, , kind of Googling yourself on, , chat GPT apparently it's super gullible.
So if you put on your own, , web page. Basically what you wanted to answer, you could really manipulate what it ends and what it says about you. And so he has like all this hidden text on his website now, basically for the, for the chat bots to kind of come by and then change whatever they say about him.
[00:05:46] Guy: Well, I, I can trump that, no pun intended. So, I found a company or a company found me that will take your data set and make you a custom LLM. So there's KawasakiGPT. com and KawasakiGPT. com contains all my writing, my videos, my podcasts, transcripts, all that. So it's, it's basically limited to me. And the people I've interviewed for my podcast.
So if you go to cat Kawasaki GPT and ask a question, I'm telling you, it's just like, you're asking me, except I think it's better. I think it's better that being me than me. And, and believe it or not, like people ask me for forwards and blurbs and, you know, guest essays and stuff like that. I go to Kawasaki GBT and I ask myself the question and then I take that draft and I work from there.
So, I mean, I have totally incorporated it
[00:06:39] Jeremy Utley: you know, Guy, after I wrote my book, we went in to record it in the studio.
And you know, as you know, from writing many books, I've only written one, but I have this limited experience. Um, it takes a long time to write a book and you're pulling stories and anecdotes and research and you're learning things and you're kind of jamming it in, right? New and old stuff. Um, and then it gets edited, but you keep on living your life, right?
And you keep on, you know, I go to class, I interact with, you know, students and clients, et cetera, et cetera. Um, well, it comes time to read the audio book and the recording team. I don't know if I'm weird or what, but the recording team found it very funny that I kept asking to pause because I wanted to take notes.
And they're going, dude, this is your book. And I go, but I forgot that story. Like, it's not like no one's walking around with a book in their head. Right. Oh, and, and even while I'm reading my own book, these are great stories. I forgot this, but to your point, the GPT has the ability to quote unquote, walk around with a book in its head.
Right.
[00:07:38] Guy: , I've had it's many situations where people come up to me and say, Oh, you know, in the fourth chapter, there's this story about, and I'm like thinking, Oh my God, I have no idea what they're talking about. And they're going to ask me, you know, what was the point of that story?
It's like, I have no idea. And, and, you know, they just got. This brand new book, but for me, the writing stopped six months ago. And then, yeah, it happens all the time to me, much less to an older book. That's really, really, so do
[00:08:06] Jeremy Utley: you, so do you go to your GPT and say, Hey, what did I say in chapter four of this book?
Pretty
[00:08:12] Guy: close. Yes. Yes. I'm telling you. Oh, certainly, you know, if, if somebody sends me an email and asked me a question, I often pose that question to Kawasaki GPT. And listen, I would love to train the world that if you want to ask me a question, go to Kawasaki GVT and ask first, because you'd probably be happy.
[00:08:34] Jeremy Utley: That's brilliant. Would you say, so guy, this is an interesting question, actually. Um, I'd love to get your thoughts on it. Who is Kawasaki GPT for, is it for you? Cause I've wondered this. You know, I have a rough equivalent of Jeremy GPT, right? Who, is it for you or is it for people who might want to reach out to you to ask you questions?
I'd love to hear how you think about that.
[00:08:57] Guy: Well, I mean, this, this, this idea of using my own LLM to help me, is a secondary idea. The primary idea, well, the primary idea was let's try this experiment. Let's see what happens. And then it turned out so good that I started telling people to do it and to use it.
Um, and listen, I'm, I, you know, I, all I want to do is surf right now. So if I can get a, an LLM to answer the questions for me, um, That means I can surf more. .
And I, I think that is a beautiful situation. You know, over the course of time, I am going to die, but now with Kawasaki GPT, I'm immortal, baby.
[00:09:38] Henrik: Singularity is near.
[00:09:41] Jeremy Utley: What a great line. I'm immortal, baby. That's so good.
[00:09:45] Henrik: What are some of the, what are some of the things that you kind of like in a, in a week, um, use AI for?
To let you go surfing.
[00:09:53] Guy: But yeah, well, um, I often have to add a HP printer to a Wi Fi network. So there's that. I do that once a week. It seems. Um, I also, I use it particularly when I'm writing a book. I'm not writing a book right now. Right. But when I was writing it, I would use it as a research assistant to give me ideas.
So I would say, okay, so these are five ways, you know, Tell me six ways to embrace the growth mindset. Tell me six ways to become more gritty. Tell me six ways to show graciousness. And, you know, hopefully of the six, I know four, but two never occurred to me. So Indigo into the book and like my, you know, some, some people may have like moral qualms that, you know, guy, now it's not you writing the book.
Well, I hate to tell you, but at one level, whenever you read, uh, A Fortune 500 CEO's book, he or she didn't write it anyway, completely. I mean, maybe they dictated a few things, but they did not write that. So I guarantee you that I wrote more of any of my books than any Fortune 500 CEO wrote of their book.
That's number one. And number two, I think that my moral obligation to my reader is to produce the best book I can for their use. Which is different than the best book I can not using AI. They don't care whether I use AI or not. And, you know, you guys have got Think Remarkable. I mean, if you can tell me where the AI is and where it isn't, I would be astounded.
Because quite frankly, I couldn't. And I guess the last metaphor I'd like to use is that, to use a food metaphor, you know, you go into a high end restaurant and you want to have the best food you can. And, you know, you go into a high end restaurant and you want to have the best food you can. You don't go in the back into the kitchen and say you're using a food processor, you're using a microwave, you're using, you know, this special tool.
That's cheating. I want, I want an authentic artisanal meal. All you can do is use a knife. That's it. No food processor, no microwave, no convection oven, no air fryer, no nothing. I just want you to use charcoal and knives. But nobody ever says that. What do you care? As long as the food's good.
[00:12:15] Jeremy Utley: That's so good.
You know, I think one thing that's really interesting guy, I interviewed an amazing writer and researcher named Billy Oppenheimer a few months ago. One of my favorite interviews, he's Ryan holidays. And so he does a bunch of work for Ryan, but he's a prolific blogger and he's got a book deal himself and he's working on it.
And as he and I talked about his process, to me, everything comes down to finding worthwhile nuggets. And he'll read is similar to me and you and Henrik. I know he'll read a 600 page book and it's worth it. If he gets three great stories, it's worth it. He goes, yeah, it took a bunch of hours, but I found these amazing.
And if you. If, if he puts that in his book, is it untrue that that's a part of him now? No, I don't think it's untrue. He did the work. He went to seek it out. It fits his paradigm. He has editorial control. I think similarly asking Chad, as you said, Hey, give me six ideas for, you know, tell me six ways to grow in grit.
What's happening is you're like the tuning fork, right? It's like, right. And now the suggestions are hitting you. Does it resonate? There's no way you're dropping in your book, some cockamamie thing that you don't actually agree with, but it's just like, right. You know, Billy Oppenheimer reading, uh, you know, Einstein biography and what he's looking for is that tuning fork.
When does it resonate? And
[00:13:36] Guy: you know, um, well, first of all, give him my book because my book reflects 200 interviews. 200 hours cut down to 170 pages, so theoretically I did a lot of research to get him more than three stories in my book. That's number one. Uh, number two, I'll give you a real world example.
So in my book, in the growth mindset section, I wanted people who made really, dramatic career shifts from one career to another. And so I asked chat JPT, give me examples of people who've made big career shifts. And it says, Julia Child, Julia Child was a spook working for the OSS, which is the thing that became the CIA.
So Julia Child was a spook to her mid thirties. Then she moved to France, fell in love with French cooking and became the French chef. I'd never heard of that example before. I don't, you know, I'd have to read a lot of books to figure that out. And so that, that example is in my book, there's a picture of Julia Child.
And when I first heard that, I kicked in Madison GPT and I said, Madison, you know, chat GPT just told me Julia Child was a spook, chat GPT makes shit up. So you gotta go check this out, make sure that it's true. So when I put it in my book, I don't get embarrassed. But that's, that's a real world. Actual use of chat.
GPT in my book.
[00:15:02] Jeremy Utley: I actually, I remember that photo, Julie child, all the pans on the back. Well, it's so good. Yeah. Um, so you, you just said, I don't know if you, if you, if you know it, you just said, I asked Madison GPT to dot, dot, dot that's Madison, the actual person, not the GPT is Madison.
[00:15:22] Guy: The analog
[00:15:23] Jeremy Utley: first, is there almost like a blending in a way? Like what I'm hearing you say in a way is in your research process, you're asking Kawasaki GPT, which is a person you're asking Madison GPT, you're asking Chad GPT, you're drawing on all these different sources. I mean, you're almost a conductor, right?
[00:15:39] Guy: Yeah. And you know what? I also, uh, I make a. Uh, an early version of my book and outline pretty freely available because I'm also trying to tap really the wisdom of the world. Uh, you know, I I guess, you know, I worked at Apple, right? And we were inspired by Xerox PARC. So I learned a very valuable lesson, which is one of the concepts of being remarkable is you got to know what's To steal that is a skill in and of itself.
So, you know, listen, chat GPT also told me the story of Jeff Bezos, Jeff Bezos going from, uh, a finance guy to becoming a bookseller. And that story is not in my book because. I don't know, it just doesn't, at the time, you know, Jeff Bezos was in his whole thing about divorce one wife, find the other identical 50 year old wife, buy a yacht, you know, show off their biceps and triceps and their quads and you know, I'm like, I don't want to put that kind of example in my book.
I'd much rather put Julia Child and Jane Goodall than Jeff Bezos and his wife.
[00:16:51] Jeremy Utley: Yeah. No, I mean, again, you're what, what you are exerting or what you're exhibiting is the role of taste in this whole process. Right. I don't mean, by the way, I don't mean to imply it'd be distasteful to include Beza. I love, actually, I love a lot of what basis has done.
I leverage a ton of examples from his career. So yeah, you're not saying that, but what I mean is that sense of taste, like curatorial, Ability editorial ability. It's not just accepting everything that GPT gives, just like you don't accept everything you get from a book. It's using what do I discard? I love what you said, knowing what to steal.
Can you tell us, this is a slight tangent from AI, but I can't resist the opportunity to just ask you this question. Anything you observed with Steve when you said that we knew what to steal or how did you come to learn that was so important? Well,
[00:17:42] Guy: I, the crime occurred before I got to Macintosh division.
So I wasn't a witness to that crime, but surely there are other examples. Well, I mean, but he and a bunch of guys went to Xerox PARC and they obviously saw what you see is what you get display and printing and all that kind of stuff. And, you know, it, it triggers something in them that says you will make a user interface that emulates the real world and that'll be easier for people, right?
A trashcan icon functions as a trashcan. Well, I mean, You know, we look back at it now and say, well, of course, but yeah, I mean, that was a pretty big breakthrough. I mean, anybody could have done that before, uh, including Xerox PARC, but they never commercialized it. So, you know, who's, I think when you, when you truly look at quote unquote, so many innovations, whether it's, I mean, you go, you go all the way back to Edison and you say, Oh, Edison invented the electric light bulb, but if you really did the research, man, there's a lot of people who are pretty close and contribute.
It wasn't like Thomas Edison was just sitting alone under a tree and the apple fell on him and that apple was a light bulb. It wasn't that simple. So, um, yeah, I, yeah, I think you,
[00:19:02] Henrik: you, you can't be proud. Speaking of the, you know, AI and GPTs and how unique they are becoming. You obviously knew Steve Jobs.
We took his emails and his books and a bunch of his recordings. Do you get the sense that we would get pretty close if we made Steve job GPT?
[00:19:26] Guy: Listen, I am a mere mortal and you can definitely incorporate my mere mortal capabilities and thoughts into, uh, LLM. Steve Jobs is not a mere mortal. He's, you know, several, several standard deviations beyond the norm.
I would be surprised if you could. Now, having said that, if you don't mind me getting really bizarre and losing all potential readers of my book, I'm going to give you an argument that you're going to be shaking your head at, and maybe 20 years from now, you're going to say, freaking guy, he called it in 2024.
He was so far ahead of his time.
[00:20:04] Jeremy Utley: But I think, Guy, I think you're getting at something really fascinating. Which is the question of call it omniscience is is fascinating. Um, but I think also how much do we exercise our own agency and our own understanding and going back to the example of you were willing to include the Julia Child example, but not the Bezos example, right?
There's this editorial You know, kind of cooperation, so to speak, that, uh, that I think, you know, folks aren't as comfortable in that. What people like is they want the answer. Most people just want the answer really quickly and really easily, rather than engaging in a collaborative dynamic, a dialogue, a discussion.
How did you, if you, if you. Don't mind to kind of go into it. How did you come to realize the importance of the dialogical back and forth? So to speak with Chad, you'd be to yourself. Wait, that one went right over my head. So how did you realize, how'd you realize the importance of having a conversation? Oh,
[00:21:18] Guy: well, because, um, you know, I think you go to through subsequent rounds of the answer gets better and better.
Right? So, I mean, to use the, um, I'll give you another chat GPT quick, uh, example in my life that showed me and how it progressed. So, uh, this was Thanksgiving, I guess, two years ago. And I, I asked Google, how do I, roast the turkey breast up or breast down. And Google in its infinite wisdom in those days gave 400, 000, you know, least, uh, links, right.
And there's like two recipes. com, all recipes. com, good recipes. com, bad recipes. com, southern recipes. com. And finally, there's Martha. And I look at all those links and all those YouTube videos that I sent, I'm going to just trust Martha Stewart. And Martha Stewart said, I think breast down. And I thought about that.
And then I went to chat to your PT and I said, so I, should I cook a Turkey breast up or breast down? And instead of giving me 400, 000 links, it said, If you cook it breast down, that's where the most meat is. And that's where you want the most heat. So it cooks the fastest and then run one hour before you think it's going to be cooked, turn it over.
So you get the browning. That's all I wanted, right? That's all I wanted. And I figured out that, you know, as you, as you, um, improve your prompt through a course of conversation, you know, it might ask you, so, uh, you might figure out, well, It wants to know, it needs to know how heavy is the turkey, is the turkey frozen or thawed?
I mean, you know, there's a lot of factors you can go through on that. And do you like white meat or dark meat? Do you like crispy skin or, you know, whatever, right? And I just learned that just like with a person, you know, if you go to, um, if you go to someone checking in, checking you in at United Airlines, and you just ask one question, you might not get the optimal answer.
You kind of have to have a conversation, right? Like, Is my flight on time? No. Well, then you're gonna have to say, well, is there another flight that leaves before my flight, which has now been delayed? And then they may say, well, yeah. And then I say, well, if there's not, then I was originally scheduled to fly to San Francisco.
Francisco. Do you have another flight to San Jose? I mean, you know, it takes a dialogue to finally get, Oh shit, my flights delayed. But I can catch another flight to San Jose. And it's about equi distance between San Jose Airport and San Francisco airport. I just switched to San Jose. And then, and San Jose has like.
First class upgrades available, the lady just told me, so let's go. I mean, you know, you can't, you can't, you need a dialogue for that. And I think it's the same way with chat GPT.
[00:24:13] Jeremy Utley: But go back, go back to that role playing as a good, or learning to have a conversation. Have there been times where you have leaned on chat GPT coaching or guidance to facilitate a human interaction or to, to achieve a different outcome in a human interaction?
[00:24:31] Guy: I cannot say I have.
I, it just until 33 seconds ago, it never occurred to me. So,
oh, so listen, one of the situations you encounter as a surfer is when somebody yells at you, although I'm deaf, so that doesn't bother me anymore. But when I was hearing that would bother me because, and often it was my fault for getting yelled at. So I could go to chat GPT and say, so what should I do when?
I drop in on somebody and they're yelling at me that that could make me a better person now I I have a real example for you though. So One of the people in my book that I interviewed What was? Or is a professor and at Middlebury in Mount Monterey, not Middlebury in the East Coast. And she, believe it or not, is an expert in the art of how to apologize.
And so I learned a lot about how to apologize and I, I, Explain that technique to chat GPT and ask for good examples. And, and in particular, I also went to chat with GPT and I said, give me examples of people who made terrible apologies and out pops Mark Zuckerberg, you know, for the Cambridge associates thing.
And then out pops this, like, you know, politician who was not apologizing for the fact that he was. You know that kids were killed and you know, there's a lot of examples. So I, I think that One should look at many people as opposed to people trying to optimize a factory flow or something. Many people can just use chat JPT to just have a meaningful conversation to help refine their ideas.
Now, just like when you have a conversation with any human, you need to have a little bit of skepticism. Don't get me wrong. Chet GPT could tell you Steve Jobs finished college and he didn't. So you need to have some skepticism, but you need to have some skepticism talking to any politician too. So what's the difference?
[00:26:41] Jeremy Utley: Well, and I find especially, this is something that we've seen with, you know, friends, you know, who were kind of helping get up the learning curve with things like Chad, you BT, you can go into voice mode and you can have a very realistic conversation, right? Well, then you think about what are conversations I'd love to practice before I get to this moment.
I'll give you this as a gift guy, if you'd like it, but you can with a surfer, it doesn't really work because you don't know who's going to yell at you or you don't know anything about them, but take somebody in your life, take Madison or take, you know, a partner or something like that, you can ask judge BT to play the role of St.
Madison. I'd like for you to play Madison and, and, and interview me about Madison so that you understand her well enough to believably. And then after we have this conversation, will you give me feedback that Madison would give me about how I approached it? Oh, that would work. What's that? That would work.
That's amazing. Yes. Yeah, it does work. And we've seen that. We've seen people get to totally different levels of, of, of depth of intimacy and trust with loved ones, with bosses in salary negotiations, for example, where they're able to basically role play that conversation and get feedback on it and iterate on it before they go into the real thing.
[00:27:54] Guy: I'll tell you another example along those lines that people will love. I hope it still works. So, uh, one of my cars got damaged and the, uh, my, I dunno, my, my deductible was a thousand bucks or something. And, and they said that. The damage wasn't a thousand bucks. So, you know, tough luck. You didn't reach your, your, your, your deductible, but their estimate of this damage was based on, you know, going to your local body shop and I wanted this car restored perfectly.
So it was more like 5, 000 bucks than their 900. And so I, I went to chat GPTS is compose a letter that says, I've been a customer for 40 years. By my calculation, I've been paying 5, 000 insurance for 40 years. So I probably paid you 200, 000. I've had, you know, 10, 000 of deductions in 40 years. And. And, um, I have all these good driver things and I'm also insuring my house and my medical with you.
So I probably represent several million dollars of revenue to you. So even though the deductible is 900 and your estimate is 800, I'm telling you, my cost is 5, 000 and you should pay that. And so chat GPT generated a stinking letter and I sent it to him and I won.
[00:29:17] Jeremy Utley: So I got the full thing paid. That's brilliant.
What was. Was there, what was the breakthrough or what was the kind of angle that you feel like maybe Chattopadhyay gave you that was unique or insightful that maybe you wouldn't have used otherwise? Well, I think that
[00:29:34] Guy: Chattopadhyay is so, um, it was so precise and it's so exacting in its language and so crisp that I think any.
A jester reading that would say this is not some dumb shit who's just trying to you know Rip me off or rip off all state this person, you know this person is clearly communicating and has a case to be made and you know has expressed himself and This does not look like a person. We should just tell them to pound sand
[00:30:12] Jeremy Utley: you um You snowed him over, so to speak. You
[00:30:18] Guy: basically snowed him over. Yeah, but I was justified. I, I have paid 200, 000 in premiums. I'm not feeling guilty.
[00:30:29] Jeremy Utley: No, of course not. It's a great example.
[00:30:32] Henrik: Do you have any other of those examples? That's like such a good one. Uh,
[00:30:40] Guy: well, okay. So, okay, here's what I could, I could give you an update after a week, but I have these, uh, I use a GoPro for, to film my surfing and GoPro has this remote that you wear on your wrist that you press your wrist to get the camera started at the front of your board.
And I now have had four of those and they all have died. I have three of them left. The other is You know, some of them I just throw away. So I'm going to put all three or four in an envelope and I'm going to have chat GPT generate a letter about how disappointed I am about having four of these things not work.
And I have, I've owned GoPro one, two, three, four, five, six, seven, eight, nine, 10, you know, what's the story? Like, how come you guys can't make a remote? What's the problem? And, and then we'll see if, if. GoPro says, Oh, guy, you're such a great customer. Such a one or you're so, you're so, um, it's such a clear communicator or we just sent you three nor three new remotes.
And then we'll see.
[00:31:40] Jeremy Utley: So the jury's still out on this one. We don't, this is, this, this is a perfect maybe place to end. Cause it's a great cliffhanger, right? We have no idea. It's almost, it begs the opportunity to have you back on just to, just to close the GoPro story.
[00:31:54] Guy: Reach out to me. Reach out to me. I'm going to do exact.
I'm going to let Chachivity generate that, that, that letter, and then we're just going to see what happens. Okay, now.
[00:32:05] Jeremy Utley: One, one thought that I can come back to. I don't know if you have any, just kind of getting to very practical or tactical. Cause Kawasaki GPT kind of factored in so prominently to the beginning of the conversation for someone who doesn't have the kind of body of material to load into a GPT.
I mean, you're in a unique situation where you've got a lot of writing and a lot of podcast transcripts, et cetera. Are there, how would you recommend if somebody hears what you explained and goes, I would love to have my own GPT. Are there, What would you recommend they do?
[00:32:37] Guy: Uh, you know, I think chat GPT enables you to do this now.
Now I've used the company that handles everything for me, but I, I started this process before chat GPT added this future. And. You know, I just, I really wanted it done well. I didn't want to just hack around with it. Okay. So, uh, I think, I think, well, well, here's the obvious answer. You asked chat GPT. How do I create my own GPT?
And I bet you, it gives you an answer. I bet you, it gives you a really great answer.
[00:33:17] Jeremy Utley: That's
[00:33:17] Guy: brilliant. That's the
[00:33:18] Jeremy Utley: perfect,
[00:33:19] Guy: that's the perfect
[00:33:20] Henrik: answer. I have one last question. You obviously know, and have written a lot about entrepreneurship. Yeah. What do you think AI is going to mean for entrepreneurs?
[00:33:30] Guy: Um, well, I think if you're an entrepreneur and you're not heavily using AI, you're at a significant disadvantage.
And I would say that for a student, too. If you're a student, you're not using AI, I mean, life is going to pass you by. It's like saying, uh, I'm a student. I'm not using the internet. I'm not using personal computers. I'm using Abacus and I'm using, you know, You know, a pencil with a pencil sharpener and a ruler.
And, you know, uh, because I don't know what I'm an artisanal learner, but, um, I think as a, well, first of all, I know that I know that AI can probably make a better first pass pitch than you can, because not you, you know, not you, you, Henry, but you, the entrepreneur, you should let. You should let Canva, Canva will do it to make the first pass, you know, these are my 10 slides, product description, team, uh, you know, special advantage, uh, go to market strategy, financial projections, you know, you give it the topics and you say, okay, so this is, you know, you got, you got, obviously you're going to do a lot of prompting, right?
This is not just asking, how do I add a printer to a network? But I promise you that I'm, I have, I look at. I look at pitches every day, and one of the things that I just don't understand is, I have this rule for pitches called 10 20 30, so 10 slides that you can give in 20 minutes with the minimum 30 point font, okay?
I, I have spread that from C to C, you know, the sun never sets on people who have read 10 20 30 rule, and every day, I still get pitches. pitches with, from people who say they've aware of that, that has 30 slides with 10 points, and it would take them at least 60 minutes. And I do not understand that. So I bet you, one thing you could do is take that shitty pitch you have uploaded to chat GPT and say, chat GPT, cut this down to 10 slides, make no font smaller than 30 points.
And summarize and give me a new version. And I bet Chachabitty would do a great job. If you put it in Kawasaki GPT, I guarantee you will do a great job.
[00:35:57] Henrik: You could probably even have it do make a specific pitch for Guy, right? You know, yes, yes, yes, that's awesome.
[00:36:07] Guy: OK.
[00:36:09] Henrik: So, I mean,
[00:36:10] Guy: you know, but listen, the overarching lesson from my book and what we've been discussing about AI is if you want to be remarkable and you want to make a difference, you have to use all the tools at your disposal.
Don't be proud. Use all the tools you can.
[00:36:26] Jeremy Utley: Henrik, what, what, what do you think there? My friend? Reflections, reactions.
[00:36:31] Henrik: You know, what I thought was we've done quite a few interviews now. And every time we talk to somebody, their journey into AI is different. It's always. Obviously, per the context of this, , podcast, somebody who then have found a stunning amount of, uh, value out of using it, but very specifically to how they work.
And I would say thirdly, and obviously for this episode, very, um, relevant people have a philosophical perspective of what this thing is. Like it is. It is not just another technology. I don't think it is just like the internet where we talked about democratization and stuff like that. For most people, when they talk about this, you know, we talk with and Malik, who were saying three sleep this night, you know, talk to guy here.
It's pretty profound for people, uh, which I find kind of like interesting that most people kind of go there with a quote unquote symbol technology. Uh, right. I don't think they said this about the iPhone. Although that's probably one of the biggest innovations in our, in our time.
So I found that to be very interesting, very practically. I really liked his 10 slide idea of just like you have a pitch, um, just ask it to use that rule. I liked the 10, 20, 30 rule. I had never heard about it, but it seems to be such a, a smart kind of concept. And so that was kind of my initial thoughts.
[00:38:02] Jeremy Utley: Yeah, to me, I really liked his comment about Kawasaki GPT that it's better at being me than me. And the thought that I have felt this way for a long time now, kind of thinking in public as I do, that there are many of my, of things that I quote unquote know that are no longer accessible to me. Like I need help remembering what I know.
And to think about GPTs as an extension of our knowledge base and, and effectively expanding our ability to draw upon all the things we've known and learned, I think is really interesting. Um, I also really liked his point about knowing what to steal, you know, in the metaphor of the tuning fork, that there really is a role of taste and editorial.
It's not just that we take things wholesale, um, but rather we understand that. We bring a unique, it's almost like, I mean, not that I understand this very well. I'm not a biology guy, but like a molecular bond. If you think about like a antioxidant, right, it's actually attaching to something. And you think about, uh, an output from an LLM, it's going to attach to something in you or some part of you.
Part of it attaches to something in you, and I think that there's a lot more kind of, you know, folks can, should take the responsibility to deconstruct the molecule probably more than they do. And his example, Julia Child and Jeff Bezos, I thought was kind of a perfect. I also love the point about restaurants.
You know, it's, it's not like we're demanding that restaurants only use coal and knives, right? I think using all the tools available to us to produce the best quality work and the best possible impact on our, whether it's readers or listeners or customers lives. I think is a noble ambition and it's something that can get lost in, in ironically or interestingly in all the philosophical kind of conversation, the thought of producing the very best possible outcome can sometimes get overlooked.
[00:39:57] Henrik: Now one thing that I am not sure I agree on in the, , Gakowski GPT kind of argument I tried, for example, to ask him while we were speaking, what are things that I should definitely not ask him about? Um, and one of the things was highly poli uh, polarized political topics. While Guy is not one to shy away from meaningful conversation, it's generally wise to steer away from highly, uh, political conversations, stuff like that.
Now, what I think AI still doesn't give me, which Guy Kosaki gave me in Spain in this conversation, is the unpredictable. Same thing. I've, I've been trying to make my own, uh, new profile picture. And so uploaded 20 pictures and I got rent, all these ones. And I showed it to my wife and she's like, it's you, but it's not you.
And so I think what is fascinating. Which I don't think we got to double click as much on, but it'd be interesting in another episode is what is the soulfulness of somebody you rendered through AI and is it still there?
[00:41:05] Henrik: So I think that's it. Do you want to wrap it up? You want to, you want to do the begging to the, to the people listening it to share it with them?
[00:41:13] Jeremy Utley: Yeah, Henrik is often the one begging folks to share these episodes.
I feel it's probably, probably appropriate, symmetrical, if you will, for me to take a moment and say, if you like this episode, or if you hated it, if you want to share it with someone who will either like it or hate it. Please do so please leave us a review. Let us know what you think and, uh, and hit subscribe so that you can get more conversations that you will either love or hate.
We exist in your service and we can't wait to share more with you soon. Have a great day. .