Transcription
Hi Jeremy, welcome to Just Curious. Thanks for having me. Today we're going to discuss how to unlock creative potential using AI and the role of experimentation and fostering innovation.
Before we start, a quick introduction. Jeremy Eli is an adjunct professor at Stanford University, where he's been a favorite professor multiple times. He's also the co-host of Beyond the Prompt with Henrik Verdlin, a friend and former boss of mine at Prehipe, where they explore how AI can be used to create impact on both business and daily work. There's a lot more to Jeremy than just that, as we'll soon hear, but Jeremy, maybe you can start by sharing a little bit more about your work and your own AI journey.
Sure. Yeah, I I've been teaching innovation to both grad students and executives at Stanford since 2009, so about 15 years now. And I wrote a book called Idea Flow, which I think is a canonical book on idea generation, but ironically it was published one month before Chad GPT was launched. And so instead of kind of going on the professorial world tour to support my book, I basically made a decision to strap myself into the front row of the classroom and become a learner so that I could help my students, my former and current students, navigate how do they leverage this technology to suit their purpose of innovation. So for the last two and a half years, I've been I really feel more of a student even than a professor, whether it's through the podcast or through research efforts, and I've been really interested in this intersection of call it creativity or innovation and AI and how does AI amplify human creativity and innovation?
Yeah. What was it that you saw initially that inspired you to sort of lean into um AI as an area of exploration?
You know, I I had a chance to play over that Thanksgiving week of 2022. I think just like everybody else, it's like you take a break. You know, buddies start sending a text with this link to this website, and you go to it and you're going, "Wait, what does this do and how does this work?" And I was actually taking a coding class mostly just for fun at that time. And I started really putting some instructions into Chad GPT. It was just 3.5 and kind of blown away by what it could do, playing around in the sandbox. And I thought, there's really something here. Uh, and I I played basically all that fall. And then early in 23, a good friend of mine and a professor at Singularity University, Kan Gohar, reached out to me and said, "Hey, have you been playing with this?" That's all good. Keon Gohar from Singularity University, reached out to me and he said, "Hey, have you been playing with this chat GPT thing?" And I said, "Yeah, a little bit." He goes, "Do you feel like there are implications for your work and expertise and innovation?" I said, "Definitely." He goes, "How about we do a study?" And I joked, I said, "Kon, I've been teaching at Stanford for 14 years, and I've never done a research study." So it's fitting that we do it now. So we actually we launched a research study to look at the implications of generative AI on problem-solving and innovation inside of enterprises. And that's what really kind of thrust me into this both learning the technology but also learning about how people interact with it and leverage it inside of businesses.
And one of your focuses now is on helping executives better understand and use this technology. I'd love to share an example or two or three of um how you go about doing that. Um, and maybe we could start um at the beginning with who came to you, what was their problem, and we can go from there.
Yeah, absolutely. For the last five, six years or so, I've been meeting with a small group of CEOs. We call it our explorers Junto. Named after Ben Franklin's mechanism for sparking his own imagination. He met with the same tradesman every week for 30 years. That's how he came up with things like the lightning rod and the bifocal and mapping the Gulf Stream and the Continental Congress, right? He had this kind of group of people that he spoke with regularly. I had a couple CEOs in my life who had been former students running large mostly public companies who had I had had one-on-one relationships with, and at some point probably right at the beginning of the pandemic, one of them just said to me, you know, I can't talk to my board about this stuff and I can't talk to my employees about this stuff. And that struck me. And so I asked another CEO that I was talking with uh later, a different industry, hey, who else do you talk to about this stuff? And they said, you that's it. I can't talk to my board. I can't talk to my employees. And that theme kind of started emerging. And so me and a couple of those folks basically reached out to folks in our network and said, "What if we gather together for the purpose of mutual benefit and learning and sharing in a safe environment?" So I've been meeting with that group at least monthly and with most of them, there's about, you know, 12 of them now, and most of them every month independently as well. And so there are lots of questions related to driving a growth mindset in the organization or agile ways of working or pursuing a horizon 2 horizon 3 innovation, all that stuff, right? So we're making space to talk about that, well, of course, when GNAI comes out, everybody wants to wants to know what are the implications for our business, and so a lot of our meetings started naturally gravitating towards the topic of AI, and a lot of my one-on-one meetings really started to concern AI. And I'll never forget there was one moment where a private equity operator who had had to step into one of the portfolio companies as a CEO called me and said, "I got a I got a serious issue and I wonder if you could help me." I said, "Yeah, what's up?" And he said, "I've had to step into this company and we've got a couple of really senior leaders who've been here for a while. They've had a liquidity event and they just they don't have the drive that we need, and it's been placed upon me to have the termination conversations." and you know HR has given me a binder full of materials, but I feel like there's some there's an opportunity for AI here, and I said, oh you bet. I said, let me make you a video real quick, so I quickly went to ChadBt, I basically got Chad PT to interview me to create a psychological profile of a person with whom I needed to have a difficult conversation. I then fed that profile into a new Chad GBT window and said, hey, play this role, and then I took it to voice and I role-played a difficult conversation, and then I got the transcript and I took it to a third Chad GBT window and I said, "You're Kim Scott, author of Radical Cander. I want you to give me brutally honest feedback about my approach to this conversation." Like all in, call it literally 10 minutes. I made a Loom video. I sent it to this guy, and he goes, "I'll be right back." And 15 minutes later, he's like, "Okay, what should I do next?" And I said, "Well, ask Chad GBT to now it's played the same role, but it's had a bad day and your conversation partner is less likely to offer concessions and is going to be a lot more difficult in this discussion than they were last time." And then 15 minutes later, he wrote me back, "Okay, what do I do next?" And I realized, "Wow, there's really something here." And what's what's funny is I said, "Hey, tell me how it goes." He wrote me the next Tuesday and he said, "You'll never believe it." He goes, "I had the termination conversation. It was a really productive, fabulous conversation. We're walking out in the hallway, and the head of HR looked at me and said, 'That was the most effective conversation I've ever witnessed in the company. Which of our training materials did you find most helpful?'" And he said, he said, "Jeremy, I didn't have the heart to tell her that Chad GBT helped me."
Right. Right. That's amazing. Is that something he's been able to use subsequently in other HR interpersonal like conversations?
Well, that's I mean that's we actually wrote a Wall Street Journal oped about the topic. How to use AI to prepare for conversations you dread because it's such a nonobvious and unexpected use case that people find delight in. For sure this individual has continued to use AI to prepare for difficult conversations, both at the operating partner level and then also, of course, you know, operators in businesses portfolio companies, but that's something that I've seen, and then of course that story kind of got shared throughout our group, and then others kind of go, hey, what's happened here, tell us about it, and I've ended up many times I mean I'll never forget one day I'm driving down the road, and a CEO of a multi-billion dollar company calls me and says I need you to join a Zoom right now. So, I pull over on the on the shoulder. I say, "Sure, what's up?" This person says, "I've got our head legal counsel in the office. Can you walk through what you told me about about this roleplay thing because we got a serious scenario we got to deal with and our chief legal counsel wants to know if it's legit?" And so, I'm like, I'm doing a personal hotspot from my phone. Um, you sitting in the passenger seat of my car on the side of the road role-playing a new conversation, chaining together a series of interactions with GPTs. And then the chief legal counsel's like, "I've got what I need. That's great. Thank you so much." And the line goes dead. And I'm sitting there. I'm like sweating in the front seat of my car. And I text him. I go, "Did that work?" He goes, "It's exactly what we needed to know. Thanks so much." But you see that kind of interaction. Most folks most folks think about content generation. They don't think about AI's ability to imitate and AI's ability to synthesize and AI's ability to process. And so where you might love to have an objective perspective, you might love a sparring partner, very few people are thinking about AI as a sparring partner and then just another AI window as an objective perspective giver, right? And so just kind of opening the aperture of the the mindset that folks are bringing to the technology and the kinds of use cases that they derive value from. It's very different than a frontline worker, right? Crafting marketing copy or a social media post or a performance review, right? These are really delicate conversations a lot of times, but I find that AI is certainly fit to the task so long as your imagination is sufficiently sparked. And what most folks really need is they actually need to be placing themselves in the stream of information where their imagination can get sparked, like a podcast like yours, right? Where you're hearing use cases that make you go, could it? That is actually the most important question somebody could ask. Could it...? The the problem is for most people, they don't have a question.
Yeah. Yeah. What's amazing about that example, one is how impactful it was, and two, how it really was about just using an off-the-shelf tool that doesn't cost very much. The unlock is opening up the aperture and asking what can be done. And so, how do you how do you how do you help people like understand what is possible and open up their aperture of thinking? I assume most people would love to explore all the different ways they could use this technology, but they just don't really know how to open their minds.
Yeah. Well, it really depends on what your what your role in an organization is. You know, if I'm working with a senior leader, one of the things I love to do is just leverage AI as a strategic thought partner. I think it's a it's a phrase that Greg Schauve from Section has used. I got to give him credit. Using AI as a thought partner is this incredible unlock where you basically can come to, you know, chat or Claude or whatever and just say, "Hey, I'm working through this issue. I want you to play the role of this particular expert." Right? So if you've got a you know if you've got a you're trying to raise money and you need you need like a equity expert or a or a a debt raising expert right then I want then you tell the AI I'd like for you to ask me one question at a time until you have sufficient context about my situation and my business such that you could recommend three to five non-obvious strategies for ways I could resolve this challenge. Please with each solution you provide give me a detailed implementation plan as well as play the role of devil's advocate and poke holes in your own thinking. You give a simple prompt like that which I just have as a text replacement. Like I use that prompt so much that I've just got an Apple text replacement for it. If I say thought partner, it populates that entire string of text into my chat. Right? But the point is a lot of times like I mean I was I was working with a Japanese CEO just earlier this week and we were talking about on the podcast Henrik and I had Brad Anderson the president of Qualtrics on uh last week and Brad reported to Satya at Microsoft for a bunch of years. He's an incredible leader, and one of the things we asked Brad in the beginning was how do you think about AI first culture? And Brad's response was telling. He said, "The only way you can claim you're an AI-driven company is if three things: Every employee uses AI in every function every day. If every employee in every function isn't using AI every day, you're not an AI first company." So, he had said that and that that kind of that quote was on my mind. I'm on a call with a CEO of a Japanese company, huge mega company, and I said they were talking to me about an AI-driven transformation because as a teacher and as a consultant, a lot of what I'm doing is creating capacity. And a lot of what I'm helping organizations do is create capacity to collaborate with this technology and their people. So, I'm working with this Japanese company to help them build capacity. And I said to the CEO, our goal is nothing short of every person in every function using AI every day. And he kind of clarified. He said, 'But Jeremy, I don't want people using it tactically. I want people using it strategically.' And I said, 'Great.' And he said, 'How would how would you help me help people use AI strategically?' And I said, you know what I'm going to do? I'm the AI expert, and rather than answer your question right now, which is what everybody thinks the expert is supposed to do, I'm gonna share my screen and I'm gonna pull up Chad GPT and I'm going to ask it to be my thought partner because this your question right now is the most important question I've gotten today. Why on earth would I provide an artisanal handcrafted response when I have augmented intelligence available to me? Why don't I show you how I'd use my augmented intelligence to answer your question, which respectfully is the most important question I've gotten today? And so, real-time, I'm sharing my screen and I'm showing him how I'm using AI to answer his question. And then I said, what you need to do, respectfully, I would suggest, is take I'm going to copy-paste this same prompt into chat for you. I want you to copy-paste it, and I want you to make a Loom video of yourself using this framework for whatever strategic question is on your mind right now, and I want you to send it to all your employees, and then I want you to paste the framework into the email and tell them the first thousand people who send me a screenshot of how they've used this prompt are going to get invited to a special group. There you go. Right. It's just super meta, right? But the point is there can be a lot of talk, but there's no walk. There's no rock. Sorry, I'm I don't mean to be like dropping bars, but the point is you got to actually start doing stuff and showing people how you do it. And at some point, the AI if it doesn't actually enter the conversation or like slide into the DMs itself, it remains this ethereal nebulous topic rather than no no like this. It's like do it like this is what a lot of folks need. And you can really do that at any level of the organization. I just happen to be interacting with folks who who are responsible for more strategy than they are the tactics.
Yeah. Yeah. And it's the first I guess step for someone who comes to you and wants to become an AI first company or they want to design this AI strategy to become a better company in the future than they are today is just to start like using it, and then where do you go from where do you go from there?
Yeah. Well, for sure leaders have to be using it, right? So that's unequivocal. This is not about you telling someone else to do something. It's about you taking action yourself and then learning from the front. You know, just like I've done in my classrooms, right? I'm diving in as a student and then sharing my kind of crib notes with other students who happen to be my students, right? But the leaders that I think are most exceptional are people who have practical pragmatic examples of how they're using AI. I'll give you four rapid examples. One, Bryce Shalomal. He's the head of AI at Madrona. He's got an amazing story about how he used AI as as assistant, as coach, as expert, and as creative partner to reinvent his kids' parent-teacher conferences; that's incredibly compelling and rich, right? That culminated in a series of table games that he and his wife play with their kids to help them develop in accordance with the goals that their teachers have given them in paired teacher conferences. Really cool, right? Brad Anderson tells a story at Qualtrics about how he has committed every single time he's got an all-hands meeting or a meeting with more than 10 people. In the three or four minutes inevitably where it's green room time where they're waiting on everybody to show up, he shares his screen and shows people the latest tool he's using it and how he's using it. Greg Schauve, the CEO of Section, says he always feeds his board deck into a series of AI tools to ask what questions each of his board members are going to ask. and he says he's seen 91% overlap between what AI suggests his board members are going to ask and what his board members actually ask. Right? Diara Bousos, the CEO of Diara Blue, amazing Sinagalles fashion designer, says she's taken AI taken processes in fashion design that historically take two months and she's compressed them down to 2 minutes. But she says one thing she always does is make a video of herself using AI tools because she quote doesn't want to be leveling up privately when she's expecting her team to be leveling up publicly. But in all of those cases, I just gave you four radically different examples, the leader can say, "Here's what I'm doing." And when and I can't imagine in any organization that there's a more important task than preparing the board deck, right? And if you want to make sure that people only use AI for the trivial tactical 80% of work that only drives 20% of outcomes, then what you do is give them examples that are trivial and tactical and that that don't drive much outcome. But if you want them to be strategic and to look for really impactful use cases, then you have to demonstrate how impactfully it can be used for you in your role. Which means if you're the CEO, Greg Schauve, the most important thing you can be doing is updating your board on your progress and your problems and your priorities, and you're showing your team how you're using AI to prepare for that conversation. It gives everybody in the organization not just permission to use it tactically, but to permission to use it strategically as well.
Yeah. Yeah. That's really interesting because it really is like it's inspiring people but it's also giving them permission, you know, and at the intersection of that hopefully is that opening up of the aperture and the experimentation. I think that word experimentation is really important, and one thing I would recommend is something I've talked with a lot of leaders about. If it's okay to not use AI, people are going to not use AI or there's going to be a lot of people who don't. Right? Which is to say what if inaction is permissible, folks won't take action. And I think increasingly in organizations, leaders have to make inaction the punishable offense. And that's a total paradigm shift, right? Most people inaction's fine, and sitting on the sidelines is fine. And I think you got to bake it into KPIs. Do you need to tell any employee if you have, you know, senior leadership team, I want to hear from each department, what are the 10 AI-driven experiments that you've run? Success is not important. I don't care if you've succeeded or failed. In fact, perhaps all 10 fail. But if you show up and you have less than 10 experiments, that's a problem. And right now, showing up and saying, "Yeah, we haven't really known if we're allowed to try," is permissible because you haven't said it's the punishable offense. And you just have to you have to totally change the paradigm.
Yeah. And that comes from the top. Um, let me ask you about creativity because that's been a big focus of your work. How does AI change the creative process?
It amplifies the creative process significantly. There's a lot of research obviously that suggests not only the Wharton study at BCG of the speed of work, the quality of work, the the volume of work all goes up a lot. There's also studies from Wharton that show that the volume of ideas and the variation of ideas go up as well when you leverage AI. I would say there's kind of a big asterisk with a lot of that stuff because some of that research I saw in the early days, it clearly intersects with my area of expertise, which is part of what made me curious about all this stuff. And as I study it in organizations where folks aren't primed or aren't equipped, basically what I observed, what Keon and I observed as we studied this was most teams fail to outperform with the help of AI. Most teams actually underperform when they're using AI because of this pernicious cognitive bias that Herbert Simon won the Nobel Prize for back in 1954. I think satisficing humans are just looking for the good enough answer. And so AI effectively, one simple way to think about it is AI gets us to good enough much more quickly. The challenge is knowing when your goal isn't good enough, it's the best. Granted, it's too expensive calorie-wise, resource-wise, etc. to be finding the best answer to every problem. But when it comes to, you know, new products, services, groundbreaking innovation, solutions to problems, you really do want to identify the best solution. And furthermore, when the effective incremental or marginal cost of a solution is just like tokens, it's not even human effort. Why would you not push farther? And yet, what we found is time and time again, folks inside of organizations ask, you know, give mediocre prompts, get mediocre output, and go, "That's pretty great. I guess we're done, right?" Rather than pushing. And so what we saw was the teams that outperformed treated AI fundamentally differently than the teams that underperformed. The underperformers treat AI like it's an oracle. Like it's got the right answers. You just got to like a genie in a bottle. Just summon it. Right. The outperformers didn't treat AI like an oracle. They treated it like a coworker, like a collaborator.
Yeah. And so my obsession became, how do you develop good collaboration skills with this new technology? And that's that's kind of that's the key to unleashing the creative capacity of AI is working with it like you would a colleague. Not just asking it a question expecting the right answer, but pushing and prodding and guiding and giving feedback and testing and iteration is really important obviously. And that that might sound obvious, but the truth is if you ask most people right now listening to this show, if they go into their, you know, chat history and chat GBT or Claude or whatever, look at how much back and forth there is in your conversations. And you can do that's a very quick litmus test for the degree of the collaborative dynamic you have established with AI. And the truth is most people it's indistinguishable from a Google search for them, right? Which is like they aren't they haven't even scratched the tip of the iceberg. Uh, what about angry back and forth? Angry. I feel like
Sometimes I get like, like frustrated. Yeah, great. Of course. Sure. I mean, and how, how great.
By the way, you get angry with an intern, and what happens? Like they cry. You know, you get angry with AI, what happens? You're right. That wasn't a great idea. Let me try again. You know, it's indefatigable and it's indifferent to your emotion. Not that you can't use kind of emotional manipulation to affect outcomes, which you can, but the point is I'd much, I'd rather you much more be getting angry with an AI than with a human.
Most certainly. Who do you think is, for the people listening to this? Who's the best person for the type of work you do? Like, what's the right kind of criteria? I guess it's leadership. It's, are there certain like, um, like industries that you tend to focus on more than another?
You know, for whatever reason, this is strange, but I, I've somehow gotten involved in the sports world quite a bit. I don't know why. Maybe it's just a function of my network, but I've, I've ended up in a number of deep collaborations with professional sports teams. So, that's one. I don't think that it's the technology or that my expertise is particularly suited to sports teams. I just think I've kind of gotten there, and I happen to really enjoy sports. So, it's kind of, it's gratifying and enjoyable. But I, I think people who are spending their time in front of a screen, people who are spending their time working with language, which is increasingly kind of all of us.
I'll give you one example that's pretty radical as a, as a totally different thing. And here maybe the criteria is folks who don't consider themselves technical in nature because I think AI, the connotation of AI is it's a technology, and I got to be digital or I got to be, you know, tech-savvy, whatever. And I think the some of the biggest gains actually come from folks who don't consider themselves tech-savvy.
So, as an example, I was working with the National Park Service. They, I'm, I'm a huge outdoorsman. I love being in the wilderness, camping, things like that. And the National Park Service reached out to me and said, "Hey, we'd love to develop AI fluency among some of our park rangers and helicopter pilots and facilities managers. Would you help us?" And I said, "Totally, let's do it." We gather about 60 folks from around the US park services. You know, there's probably 450ish parks. And we get about 60 folks who are curious. I heard from many of them that their colleagues made fun of them for joining the session. So to kind of give you a hint of the demographic, these are people whose the default culture is it's so anti-technology that if they stay for the AI lecture, they get made fun of as the dork. Okay.
So I get these people, and they're in a class. I do a couple of sessions; one really focused on just good prompt hygiene. Basic things like chain of thought reasoning and few-shot prompting and custom instructions and getting AI to question you and getting AI to rate its own work. Simple kind of foundational techniques that lead to empirically better outcomes. Right? So that's the first session. And then the second session I do focused on, at least in this module because I was kind of doing a pro bono offering, I skipped some content and went straight to how do you build a custom GPT? Very simple instruction sets to codify workflows, right? And I just asked the question, which you know well from your prehype days, is what sucks? Yeah. Yeah. And I ask, hey, what sucks in your work? Or as prehype folks would say, it sucks that dot dot dot. Right. And I just asked folks, you know, what part of your work do you dread? What part of your work do you wish could be easier, etc.
One of the facilities managers on the call raised his hand and said, "Oh, we've always got to fill out these statements of work. You know, if I got to replace the carpet tiles in the lodge, I got to fill out all this paperwork for federal funding and make sure that OSHA compliance standards and ANC regulations and RS means data and all this stuff. You know, that's so annoying." I said, okay, well, you could build a GPT to help you with that. So, after the call, in 45 minutes, this guy builds a GPT that in about an hour saves him two days worth of work that he would have spent filling out paperwork. He did it in about an hour. Wow. So, and he, he sent me an email and said, "Hey, this is super cool. I did this. What do you think?" I said, "Hey, on our next call, you should share with the group." So, we have another call. I did office hours just for people to share kind of questions, things like that, just for fun. I said, "Hey, Adam, why don't you share the tool you built?" And he started, you know, he, he asked somebody, "Hey, anybody got a statement of work they need to complete?" Somebody, one guy, I think his name is Kevin, said, "I do." You know, deep southern draw. So, Kevin, what's your, uh, what's your thing? I got to put a new crane on my boat. He goes, "Okay. Um, tell me about your boat." I'm going, "They put cranes on boats." And, uh, and, and Adam said, "Tell me about the crane." He's like, "I got a 110-footer. Need a new crane. It's got a," you know, he had all these specifications. Adam just typed it into his GPT. And then this guy's his camera's off, but he's not on mute. So his, his, his camera, his, uh, microphone's live. And I hear this guy as the GPT starts cranking out this S.O.W. So, I hear this guy going, "My gosh. Oh my good." Oh. And then he goes, "Hey, Adam, could you, uh, could you send me a copy of that?" Right? And I said, "Hang on a second. How long would that have taken you to complete yourself, Kevin?" He goes, "I'm a boat man, dude. It taken me four hours just to get the bullet points straight in Word." And I said, "Wow." And, and then a woman spoke up and she said, "Wait, Adam, you made that, that, that link." And he said, "Yeah." And she goes, "Somebody was passing that around the other day." She goes, "I blocked my calendar Monday through Wednesday to, to fill out a statement of work for, we've got to replace the front door and all the windows at the lodge at, I don't know if it's Yellowstone or US7, something like that." I had blocked three days because that's how much time it typically takes. She said, "Somebody sent me your link and I just tried it. I was done before lunch on Monday." And I go, and of course, the management consultant in me goes, "Hang on a second. How many times, Ashley, do you perform a function like this?" She goes, "I don't know, like 20 times a year." I go, "Wait, Adam, did you know you just saved Ashley 40 days this year?" He goes, "I didn't know people were sharing my link."
So, this is a guy who, by the way, I go to tag him on LinkedIn, like to brag about him. Homeboy doesn't even have a LinkedIn account. Okay, so this is the kind of person, no LinkedIn presence, no technical ability, who's creating a tool that, I mean, if you scale it across the park service, you do a little bit of back of the envelope math, it's like 10,000 days of labor that could be saved this year alone because he had basic training in how to get an AI to consistently deliver a reliable output, which is basically the purpose of a GPT. Yeah, what a great story. Just really highlights the impact and leverage and the transformative potential of this technology. I really appreciate all the examples you shared; like they've definitely inspired me, and I'm sure they will others.
Let me ask you one last question, even though you've, I think given us a lot to chew on. What's one piece of advice you give someone who's looking to become an AI-first company or an AI-first person and wants to know, yeah, what's the first step? This sounds silly or recursive, but it's, I can't tell you how often I say it. Use AI to use AI. So, you wonder how could I use it? Don't wonder by yourself. Yeah. Open up ChatGPT and say, "Hey, here's who I am. Here's what I'm trying to do. How could I leverage generative AI to do what I'm doing?" Right? That's one example. People go, "What kind of custom instructions do I write?" I go, "Use AI to use AI." Say, "I don't know how to write custom instructions. Can you interview me till you compose them for me that I can copy paste?" Right? This is a true story for me, which just is illustrative of this example. I was building a custom GPT to perform a certain task that I need performed, and it wasn't delivering the output that I was seeking. It's really, and as you know, like you, you experience this frustration and anger. It's maddening when you're like, my instructions are totally clear. Why isn't this working? And I spent probably 45 minutes, honestly Stuart, just looking at this thing, like looking at my instruction set in one window, looking at the output in the other, and I'm doing like line-by-line comparison going, what is unclear, and literally 45 minutes in, it dawned on me I could use AI for this, and I, I opened 01, yeah, I pay 200 bucks a month for, for the pro subscription obviously, which I think is the greatest deal in the world right now, and I said to 01, "Hey, I'm trying to create a GPT, here's my instructions, please tell me what you understand of my intent." Okay. And then it, you know, did it. Okay. Great. Thank you. That's correct. Or I edit it, whatever. Here's the output the GPT is giving me. Can you tell me what the disconnect is? Why is my intent not being translated in the output? And 01, you know, thought for 5 minutes and spits out this incredible specific critique of my instruction set and then revise the instruction set so that I could copy and paste it into the GPT. I then take that, I copy it, paste it into GPT and retry it, you know, in the preview window. It delivers exactly, and by the way, it's very complicated. So, it's not like easy to figure this out, but it delivered exactly what I wanted it to. And the point is even as, and I hear this all the time from people. It's like the number of times I hear AI leaders and AI experts go, I was doing this for two hours before I hit my forehead and go, why am I not using AI? Because the truth is it's so deeply ingrained to rely upon our own intellect and our own instincts and stuff that it's just, it's a, it's a deep-rooted behavior change that's got to occur. And I go all the way back to the beginning of our conversation. This idea of a Junto, this idea of a learning collective, this idea of getting in places where you can get your imagination sparked. It's so important because it will just increase the number of times that you slap your own forehead and go, why didn't I think of that? And we need a lot more of those. I think as a, as a species, we need a lot more of those "why didn't I think of that" moments in order to answer your question, take that first step to using AI and figuring out how you could use AI.
Awesome. What a great place to end, Jeremy. I really appreciate, uh, you taking the time to chat today. This was super informative. I got a lot out of it. I'm sure others will, too. And yeah, have a great day.
Likewise, my friend. Thanks so much.