Ep250: The Creative’s Guide to Understanding AI and Leveraging AI to Build Workflows & Optimize Your Creativity | with Rob Howard

Zack Arnold

I'm here today with Rob Howard. And usually the way that these introductions work is they're very formulaic. I'm here today with the following guests that has the following job title. They had the following accolades, they worked for this company in this company, and they have the following awards. They've written this book. And here we are. But something tells me that for you, Rob, the answer to the following is, it's complicated, which is, who are you? And what do you do? So I'm gonna let you do your own introduction today.

Rob Howard

All right, great. So I'm Rob and I am the founder of innovating with AI, which is a relatively new company and new line of business. So obviously, since the world of AI has been rapidly, sort of expanding and growing over the last year or so. And my core business is hdc, which is a website development firm. I've been doing that for over 15 years, really over 20 years as a professional software developer. And that is no traditional client services business, if you need, you know, custom website development, you come to us, we work with lots of big brands, including like Harvard and MIT, and the World Bank, as well as small businesses and agencies that do, you know, maybe dozens of websites per year with us that sort of thing. So really, I got my start in software development. One of the reasons that people enjoy working with me as a software developer, though, is the communication and the like, sort of journalism advertising marketing background that I bring to that. So as you probably can guess, there's a lot of people out there who are very good on the technical side, but really struggle with things like client communication, or, you know, teaching somebody how to use software once we built it for them, right. So that's an area where my team and I really try to excel in and go above and beyond. And, you know, that's where sort of the value is for for a lot of our clients. And in addition to that, you know, I have an education and sort of technology, learning and innovation and entrepreneurship branch of my business. These days, we call that innovating with AI. So that's innovating with ai.com. And we do a bunch of different programs, you know, ranging from free toolkits to, you know, advanced mentorships, and incubators and stuff like that all around, basically figuring out what to do with the rapidly advancing world of artificial intelligence, which is, you know, kind of almost come out of nowhere, like there was a little bit of it before 2022. But with ChaCha Beatty coming out, and now all the other stuff around image, video, audio, and all that. It's really been a seismic shift in everything, it affects the work that we do as software developers. But the bigger thing is that I really enjoy being an entrepreneur, building businesses, and creating those new things at that very sort of, like, bootstrapped startup level. So that has been a really exciting journey, not only doing my own stuff, but also now helping lots of students do that stuff with AI as sort of the background as well.

Zack Arnold

But you would be the perfect example of what I talked about often, which would be a multi hyphenate jack of all trades. And I just I wrote about recently how I would I'd use the term jack of all trades in a Facebook post, and I got this scathing comment, like, oh, how insulting Is this, like, you need to remove this post immediately to make judgments about people thinking only have one craft or they're not good at a whole lot. It's just like, Dude, chill pill, slow down. First of all, I acknowledge, and I encouraged the value of being a jack of all trades. And you're the perfect example of when somebody says, What do you do you're like, how much time do you have? Right? It's complicated.

Rob Howard

And it really helps when there's a major disruption in the industry that you're in, to have at least some of that jack of all trades spirit, right? So I want to be really good at many things, right. And I sort of, I guess, I don't believe in the sort of axiom that you have to be a master of one thing, and you can't focus on anything else. Like, I actually think that there is a skill set around being very good at being agile in the way that you build your businesses, right. So you know, that actually kind of runs counter to the idea that like, you have to be a master of one thing. And if you're a jack of all trades, you're a master of nothing, right? Like that's kind of the same, right? I think the way that I look at it is like being a jack of all trades, creates a robustness or an anti fragility in everything that you do in life in business. And that especially is pronounced and noticeable when something disrupts your entire industry, and all of a sudden, that thing that you mastered, maybe, like rapidly going out of style. Yeah, and

Zack Arnold

I would say that the, the term disruption is a very kind of nice way to put what's going on right now in a multitude of industries. In my industry of entertainment, it's it's a giant shitshow of a disruption. It's also happening in technology. It's happening in finance, like it's kind of happening globe Really, so it's far bigger than just creative fields or your industry or mine. But disruption would be a nice way to put it. And I've been saying for well over a year, and it's gone from scathing attacks to more Oh, shit, you might be on to something, and maybe I need to start listening to it. But this is where we're seeing the transition from specialized society built for the Industrial Revolution, to a more generalized society, we're having a multitude of different skills, like you said, being agile is going to allow you to not only survive, but thrive in the future of whatever this crazy future looks like. And you had mentioned the the saying jack of all trades is a master of none. What people don't realize is, there's another sentence to that that nobody shares. And it really goes a jack of all trades is a master of none, but oftentimes better than a master of one. And never has that been more true than in this disruption or other periods of disruption. But I've added another layer on top of that, which I want to dig into a little deeper with both you personally, but also the different things that you do, then I want to dig a little bit more into just understanding AI. But my belief is that you can be a jack of all trades and a Master of one. It's about ideas, what is the ideas? What is this one true specialty that I have? And how do I diversified amongst so many different things. So I may just be putting you on a hot seat here for a second. But whether you're the you run hdc, your website development company, whether you're a coder, whether you're a software developer, whether you're an entrepreneur innovating with AI, whether you're a dad, all of these things wrapped up, I bet there's one thing that's really your specialty, what do you think it is that makes you great at all of those?

Rob Howard

Yeah, I think that you are kind of putting me in the hot seat. But I think I have an okay answer to that question at least. So, for me, I think that the thing that connects all those things, right, is something around the communication and simplifying complex ideas. Right. So that's something that I think, you know, is you can draw that line through pretty much everything that I do, you know, even having a son who's in fourth grade like that, you know, his life is basically based on me simplifying complex ideas, like so he can understand stuff, right? And now he's, he's almost at the point where he can do that, for me for stuff that I don't even know what he's talking about. He's like, Oh, let me let me tell you about this video game, and I'll explain it to you, right, but you know, that that skill, you can draw that through everything, right. And it really helps when a lot of this stuff I do in tech, half of it started as me playing with with computers as toys, and then wanting to learn how to use them, right video games, just like messing around as a kid. But the other half of it is, people rapidly would start to come to me for for help with stuff, right? So I quickly had to learn well, okay, this person who, you know, maybe they're like a holder relative, or they're a client or whoever, like, they need me to explain this thing to them in a way that makes sense, without, you know, the years of playing with computers as the underlying framework, right? So, you know, I think that is something that I almost inadvertently created, right, or inadvertently honed as a skill. And now you can really draw that, you know, parallel to pretty much everything that I'm doing. And I enjoy it. And, you know, I think it works for for my clients and students as well. Well,

Zack Arnold

just for full disclosure, we haven't gotten to how you and I ended up on this call together. But you and I were both students in a small group coaching program to learn how do we take our different ideas and leverage them into building and growing and scaling our businesses. And it started with, Oh, crap, I need a webpage by next week. And I need, you know, like a professional company that can do this, but do it fast. And of course, I'm gonna pay you a 10th of what you're worth. But how right? And I'm like, Well, let me reach out to rob, he seems to know what he's talking about. And I can attest to firsthand that your your real specialty goes far beyond coding and developing a website and whatnot. It's, you're asking for things that are far more complex than you know how to communicate them. So let me take what you're saying. And simplifying it, but then I can relay it to my team. So I can speak their language, but I can also speak yours. So your ability to be a liaison between a very high level coder, that's amazing at what he does. But sometimes the communication is a little bumpy. Like, let me step in here for a second, let me translate all the things that this crazy high level coder developers telling you, but let's bring it down to simpler turns were like, Oh my God, this totally makes sense. And we've now gotten to the point with at least the the website of your company, where you have both a designer and a developer that are hooked in right here to the epicenter of the matrix, where I share a few simple things and what they spit back is exactly the image that I have in my mind. So your ability to communicate this complexity to somebody that doesn't know how to speak your language, I can attest to that firsthand. That's very much your specialty. And I think it's why you and I gravitated to each other because in a different space with different technology. I feel like that's my specialty

Rob Howard

too. Yeah. Nice. Yeah, agree.

Zack Arnold

So having said that, the main reason that you're here is to help us simplify this infinitely complex world of artificial intelligence specifically for creative people that I don't want to learn the ins and outs of how to code and develop and all these other things. I just, I want to know how it works. I want to know what I should be learning what I shouldn't be learning. But I think the other area where you can be tremendously beneficial, is that if I if I know what my specialty is, and how I help people, how can I leverage that with AI to not just learn the tools that are out there, but maybe develop my own tools to augment that specialization and develop a new source of income because the jobs are not raining from the sky right now? So having said that, I want to start with kind of the the most basic level, which is understanding how AI and specifically, you know, maybe chat, GBT are kind of these models that have become popularized how they work. And I want to know, What in the world that has to do with the prices? Right?

Rob Howard

That's a great question. So if you've ever watched the prices, right, which I've probably watched 1000s and 1000s of hours of on, you know, snow days and sick days as a kid, right? You're

Zack Arnold

such a child of the 80s. You're so Gen X, oh, my gosh, there

Rob Howard

you go. Right, they have a thing called the Plinko. Board, right. And when you play the Plinko game, what you do is you get up on top of a gigantic structure, and you drop a ball into it. And the Plinko board is just full of pegs, right? Little circular pegs. It's kind of like a pinball thing, right? And you drop the ball in and it bounces randomly left and right. Every time it hits a peg, when it gets to the bottom, you end up with either $10,000 or $0, or whichever sort of bucket it dropped into. So if you don't have that image, you can go Google Plinko board, and that's

Zack Arnold

definitely going to be on our show notes. Is anybody that's over the age of 30. Yes, 35. They're like, Dude, I totally get Plinko. Everybody like early 30s and younger, they're like, What the hell is Plinko? Let me google it. Well, we'll have a link in the show notes. But continue. Yeah,

Rob Howard

my wife and I often debate whether we're Gen X or millennials, because we're like, right on the cusp, right? And she's like me to Gen X. We can't be millennials, we can like I think technically we're millennials. But either way, we allow just a lot of prices, right? We're so you're dropping your ball and at the top, and what's happening is you're getting a random 5050 bounce in either direction, right. So just keep that in your head for a second. Because what chat should be T and the image models, and most of the other machine learning and AI that's out there is doing is creating an extremely complex version of that sort of random bounce mechanism. So what's happening is, when we think about chat, UBT, the technical term for that is a large language model, right. So what they have done is they have gone out and collected all the words basically that are on the internet, you know, like as much as they possibly can ingest. And then what they do is, they break those sentences down into words, and they break those words down into what are called tokens. So a token can be a word, right, like the word word would be a token. But it can also be a part of a word. So for example, podcasting, you might actually have a token for being a token for cast and a token for pod, because each one of those things kind of has its own unique meaning. And the result of that is that those essentially get coded into a database, right? And you know, they're assigned numbers, or like, there's weird algebra going on. But you can think of it as basically every token is a word or a word part that can potentially be meaningful. The word parts, by the way, are why you can ask ChaCha Beatty to invent new words for you, right? So you might say, Hey, I'm on the Optimize Yourself podcast, invent a new word. And they might say something like, podcast iser, or something, which is not a word that maybe I'd ever learned before. But it can put those pieces together, right? So you've got this giant database of tokens, then what's happening is, is, and this is where like the millions of dollars of machine learning go into the process before you even touch it, right? They are then looking at all the tokens that exist and all the written language everywhere that they can find out in the world. And they are determining the probability that one token goes with another token, right? So pod and cast are gonna be two tokens that often appear together, for example, right? But P and pod also appear together as tokens, right? There's all sorts of like, interesting nuance, and there's a lot of like, deep sort of math, and they call it linear algebra that basically is how they figure out like, how all these things fit together. But in the end, when you've analyzed enough of that stuff, you have like a reasonably good probability in terms of, hey, if I say the word pod, what should come after it, right? And then I can compare that to the other words that are in the the sentence that I'm creating. And what you get is, once you get that to an extremely high level, you get something that actually sounds like reasonably good as like written Cuban language, right. And then where the Plinko board comes in, is, after we have all those connections, we start to do training. And we say, Okay, I have a input that I'm going to put into the system. And I know what a reasonable output should be. I also know that I can change the nature of every single little peg, right, which is called a parameter or function in the world of language models. So that rather than in them all being 5050 bounces, you actually can create a peg of any shape, right? And what the training is, is, hey, I want to create pegs of the right shape, or I want to create functions that look for the right, you know, attributes of these tokens that are coming in. So that I get something that makes sense coming out, right? And that's where people say, Oh, well, it's a black box, because no one really knows like, what all the parameters are. Basically, it's like a Plinko board where you have a different shaped peg, and you're not bouncing randomly, but you're bouncing in a way that is thoughtfully created by a human a certain level. But then eventually, what happens is, the computer can actually learn to train itself, right? So for example, if you switch over to image models, because images are a little bit easier to train for, for a reason, that will become obvious in a moment. Let's say that you have 20 images of a cat, right? And I train my AI model to detect Is this a cat? Yes or no. And maybe there's 4000 parameters that helped me determine, you know, bounds this way, if the pixel is this color, if this pixel exists over here, if there's a, you know, pixel of this color next to pixel of that color, like that's the kind of stuff that it's interpreting, just at a very high scale or high level, right? Then what you can do is say, Okay, I think I have a pretty good idea of what a cat is, in terms of the parameters that I need to detect a picture of a cat. Now, I'm going to put a bunch of inputs in, and I'm going to know, is this a cat or not? And then every time it gets one, right, I'm gonna basically reward the computer and say, good job, you did it right. And every time it gets one wrong, I'm gonna say, Okay, now switch some parameters at random and see if you get a better version of that, right? So it's almost like evolutionary in a way, because you're saying basically, like, hey, that didn't work, change some stuff, try again, and see if that works, right. And when you think about a computer doing that, it can do that at nearly infinite speed, right? Much faster than if you or I were trying to, like, randomly, you know, turn the proverbial dials and, you know, change the proverbial, like, pegs sizes, right? So what happens is, you put in a bunch of training data where, you know, if it's going to, you know, what the answer should be? And then by determining if it's giving you the right, or the wrong answer, you can actually tune up that model, just with the computer, looking at the stuff without you really doing any further intervention at that point. So that's what eventually happens. And that's why people say, well, we don't even know exactly, like what decisions the AI is making, because we've done so much of this self training or machine learning, right, that it is almost training itself to give better answers. Right. So now we're at that level, and you know, that's, you know, obviously, can be sort of sci fi scary in a way when you think about, like the computers teaching itself, but the result is like, you can now type in hey, I want a picture of a cat on a on Santa's sleigh, like with a bunch of geese in the background flying, and it's gonna be like, you know, just crazy stuff. And you get the exact thing that you want out of, you know, the the image creation models. And then you can also do image recognition, right? Like, the image recognition is just unbelievably good. These days, like what's in this image? It'll be very specific. It'll read the text out of it, it'll understand the context. Yeah. So it all basically comes back, though, to this idea of randomness that is then. So you know, making the connections between tokens, whether they be, you know, word parts, or pixels or anything else that you're using as your sort of baseline content. And then understanding the probability of the around the connections between tokens, and then adding like just the right amount of randomness to it, not too much, but not, you know, too little. And then you get sentences that that sound, basically, human.

Zack Arnold

This is the perfect example of taking a year's worth of what the hell is going on to, oh, that kind of makes sense. I can kind of visualize that. I don't think I could build my own machine learning AI model tomorrow, but I kind of get it. There's one further piece that I want to add to this that in your use, like 120 minute workshop that you did, where it was a year's worth of me trying to figure it out in 20 minutes. I was like, I get it now. And there's one component that I want to add that you haven't added yet. The Plinko board It was an amazing start. And I totally understand the idea of the different pegs. And I love the analogy of, well, maybe it's not a square peg, maybe it's a round one, or maybe if it's even a slide. So there's a 90% probability the ball is gonna go this way, right? Here's what changed everything. For me. It was when it went from the Plinko board being 2d to being 3d, right? Because you could argue that, yeah, there's an X and a Y axis, there's a Z axis, because the Plinko board is this deep, right? But imagine an infinite Z access on top of x and y. And it's this three dimensional Plinko board in space. That's when I went from I think I get it to Holy shit, I understand what's going on. So go a little bit deeper and explain a little bit better than I did, how this three dimensional Plinko board is really the world that we're living in now with AI?

Rob Howard

Yeah, absolutely. So basically, you know, when you think about all the connections, right, every connection between every word part essentially creates a multi dimensional set of like connections within the computer. And that's where kind of the, like, crazy linear algebra stuff comes in. And a lot of people will say, Oh, it's, it's 17 dimensions is impossible for you to even comprehend this. And that sort of thing. I say the opposite, I say, actually, you can comprehend this. And it is important for you to figure out how to not only comprehend this, but also teach it to others, because it's going to be a big deal, right? It already is a big deal, but it's only going to get bigger. So the way we think about this is imagine you're you know, seeing a you know, in an open space, looking at the night sky, you're on Earth, you're in the middle, and there's millions of stars all around you, if the millions of stars are each a token or a word part, then what you want to imagine is a path or a line being drawn from every single star to every single other star, right, so now you have like this exponential growth of the different possible paths that you can take. And imagine that you're, you know, on a little spaceship, and you get dropped into your 3d Plinko board. And every time you hit one star, you bounce at random to another star and another star and another star. Now you have this very large and very complex 3d web of connections between your tokens in your in your AI model. And that is really what is happening. Like, if you look at some of the more technical documentation, they'll even have diagrams that essentially evoke that idea without, you know, necessarily using like the star metaphor. But I think if you think, you know, it's relatively easy to imagine, you know, there's paths that are going behind me, because those stars are on the other side of Earth, there's paths that are going, you know, left and right, that are far that are going long distances that are going short distances, right. And basically, what's happening in the computer, is the 3d connections are being stored as numbers, right, which then can then be retrieved and compared to create the results that you're seeing from AI. And the other cool thing is that because this is three dimensional calculations, right, the hardware that's being used for this is actually video game hardware, because the first use case for rapidly developing 3d graphics or 3d connections was like the PlayStation two, right? And the Nintendo 64. And they were like, hey, we want these to be 3d. Well, how do you do that? You do basically the same exact thing that we're describing, and you connect all these different points, but you convert them into math, you convert them into numbers, right. And what was discovered, you know, decades later was that actually, these graphical processing units are the perfect hardware for this, you know, it's not visually 3d, but in, you know, the, in what really needs to be done. Like, it's, it's a very similar concept of connecting all these things in a 3d space. So I think that's cool and fun, but also, you know, harkens back to my, my early days as a computer gamer, and stuff like that. So kind of kind of all comes full circle. And it's fascinating that, you know, a company like Nvidia, who basically started to, you know, make PlayStation hardware is now you know, one of the leaders in in AI hardware that is sort of rapidly changing everything about the world.

Zack Arnold

And it's probably no coincidence and at least as of recording this, and I hope that people can listen to this years from now, and it's still useful, but at least as a recording this one of the biggest pieces of news is the acquisition or partnership with Epic Games, which creates fortnight, and that goes far beyond their games are really fun, and they look cool. It's all of the engines and the math underneath and how they relate to artificial intelligence. That to me is the story that isn't as much the headline, but like you said, if it's starts with GPUs, and graphical processing units, nobody has better looking games that function better on a network than epic. I mean, they've they have cracked that code with fortnight, but it's way deeper than just, oh, this is a cool game with a user base, right? Yeah.

Rob Howard

And there's so much that, you know, the gaming industry is one of the industries that usually is I accidentally on the cutting edge of stuff just because, you know, they're always working with a relatively relatively tech savvy audience, they always are, it's highly competitive, they want to kind of like push the cutting edge a little bit, even with things like wearables, right? Which are becoming more and more common now, like you look back in like, early, you know, Nintendo products where you could, you know, swing the tennis racket or, you know, even like, shoot the screen with a little orange.com gun and stuff like that, like, you really see a bit of pioneering coming out of the video game world. And I've also seen video games that use AI to generate the actual interactions between, you know, the person playing and the in game characters, which is still very early, but is a super cool, sort of little experiment that's going on right now.

Zack Arnold

Yeah, and so I want to go back now that we have this visual understanding in three dimensional or 17 dimensional space, whichever it might be, of all these different connections, and I want to go back to understanding chat GPT a little bit and how it's able to make the predictions or even, you know, like Dolly, which is now part of chat, GPT. And all these are going to kind of start to come together. But what I want to better understand and help everybody else understand. And I'm going to steal from one of our shared mentors, you and I have a shared common mentor, and we're meets at, and he has a saying not about technology, but he talks about this with finances and with business. It's not magic, it's math. And what you helped me understand is that there isn't magic behind me saying, write a five page paper on the American Revolution, and oh, my God, how do you know this? And how did you construct this paper? Or if I say, write a five page paper about the American Revolution, from the perspective of a third grader versus a high schooler, that blew my mind and terrified me. And then it was like, Oh, it's just math, it's just parameters, right? So if we want to go even a little bit deeper, we have these infinite level of paths or parameters, which, when you now that I have that language, when I look at the different GP T's or different models, they always say it has billions of parameters or 17 trillion parameters, and like, Oh, I get that it's just the number of permutations or the the quality of the probability of the answers that I'm gonna get, right. So how does it actually choose some of these paths? And how has it gotten better, because Chechi Beatty was spitting out a lot of garbage a year ago, and it's still spitting out some garbage? And like so as generative AI with the seven fingers on hands and whatnot. But ultimately, how does it know? And what is using to determine to actually get us a decent answer?

Rob Howard

Yeah, great question. So

Zack Arnold

how does this let me add a second part? How and why shouldn't this terrify us?

Rob Howard

So, I mean, is it is going to be a big challenge, right? To be the person, you know, you know, I'll start with a terrifying question. Right? So I think our goal, everyone's goal for the next 10 years, should be the person who figures out how to harness these new tools and use them to advance the goals that you're already pursuing. Right? So whether that's in coding, whether that's in entertainment, when writing in any other anything else, right? You know, these are fundamentally new tools, right. And just like any new tool, like a tractor, or a loom, or anything else that you know, was invented, at some point in the past, it is going to make some manual labor less necessary or unnecessary, but it's also going to empower people to accomplish more in less time, you know, when they master those tools, right. So, from that standpoint, you know, it does not seem like there is a realistic possibility that this advancement will stop anytime in the near future. And I think what's more likely is that it's going to improve way faster than you might even estimate at this point, right? Like, you know, you start to see like an exponential improvement curve on a lot of this stuff, which is just hard to fathom for, for the human brain. Like, it's like, wow, this didn't get doubly good in the last six months. He got 100x Good in the last six months, or however you want to put that right. So basically, what's happening is, if you look back at like GPT, two, for example, which nobody knew about, because it wasn't very good, right? The big difference between GPT two, and three and GPT, three and four, is there's two big differences. One is that there are simply more parameters, which means more pegs on the Plinko board more stars in the sky, that results in essentially, output that seems smarter, right? And you can see the same thing with image models, you know, the more parameters you put into that, the more expensive it is to build and train but also the better results that you get because it's kind of like filtering by, you know, filtering billions of extra times right to get to the exact right output or For an output that feels as smart or feels human to us, so some of that is a scale, right. And then some of that is also the so there's the scale of the parameters. And there's the scale of the training material, right? So what GPT for is duty for is better than GPT. Three, because it has more parameters, but also because it's trained on more stuff, right. And they are just as funneling as much content as you possibly can into this. To give you some really kind of wild examples, I'm doing a project with a mentorship student right now, who is doing it all in Romanian. So what we're doing is actually taking Romanian input from any language input, getting English language output, parsing that English language output, and then turning it back into Romanian. So he can share it with his own clients and customers in their, you know, native language, that is a function of just like the massive amount of content that's being brought in, it can speak pretty much every language, it actually does translations better than Google Translate, even though that was never like one of the intended outcomes, or even as far as I know, like a particularly like big focus of the projects, you know. And it also does stuff like corrects your grammar, or like, I can send it a bunch of stuff that's, like misspelled, and it will understand it and spit back good stuff to me. So one of the weird features of bringing in this much training content, is that it actually tends to learn grammar, better than the level of grammar in the training content, right. So the output is going to be nearly perfect. Grammatically, obviously, it does come off as a little bit kind of like, bland sometimes as a result of that. And, you know, there's stuff that people are working on to sort of, like, improve upon that. However, you know, basically what's happening is more data going in, and it's bouncing through more parameters, that makes the output better and better. The flip side of that is that that requires billions of dollars of computing power, like this stuff is not cheap to do. Right. So that's one of the challenges, right, is that it's actually kind of hard for the little guys to compete in the large language model space, because it is so expensive, right? Like, not only is there a limited number of GPUs in the world, in terms of the hardware, but you know, owning and operating that hardware is very expensive. It's electricity intensive, like there's a lot of infrastructure costs there. So that being said, you are seeing Google Mehta, Amazon, Apple, open AI, anthropic, and a lot of others, like they have the money to do this. And they are just going top speed, because they saw how exciting ChaCha Beatty was when it came out a year ago, and they're saying, Well, I mean, we have billions of dollars, like, why don't we build one of these, right, so now you, you're gonna have maybe three to six of those next year, I would guess that several of those will be as good or better than the open API product. So we're going to be seeing not just the open AI products that are out there, but there's gonna be an Amazon model, there's going to be a Google model. And what's happening is, they're getting better, because the investment now makes sense, right at the the investment pencils for these companies to put billions of dollars of hardware and electricity costs into this, because, you know, they're creating something that is very unique to them, you know, they could become the infrastructure providers for the whole world of knowledge at some point in the future. And obviously, there's a lot of other sort of conflicts and challenges that come out of that. But, you know, the big reason that things are gonna get rapidly better is number one, computers just have just tend to get better exponentially when you put this sort of like, you know, system together. But number two, exponentially more money is going in now that this concept has been proven, it makes sense for literally everybody who possibly can to invest in this. And that's what we're seeing right now.

Zack Arnold

You can even make the argument that nobody has the money because at least as of recently, Sam Altman, the CEO of open AI said I need $7 trillion dollars, which is more than the entire crazy Democrats, right. So I don't think anybody really has the budget to do it properly, or the way that they want it to. But clearly, it's a very, very small space of people that are competing. And again, we could we could make this a multi part series just talking about the political implications or the societal implications, but it's what I call the Skynet become self aware argument. We're definitely not going to go there. I want to more look at it from a practical standpoint, tools resources, like you said, over the next 10 years, I've been saying for well over a year and viciously attacked that we have to embrace this and learn it because there's no world in which we're gonna say nope, AI is not allowed in our industry. We must outlaw like good luck with that, you know, let's talk to all the companies that raised horses and made buggy whips, right that everybody have one of the biggest arguments that I've heard as well. But this is like the transition from film to digital. I'm like, No, it's not. This is the transition from horses to cars that change the entire world. It's not this, you know, this novel, like, oh, there's going to be different workflows, but it's still the same basic process. This is horses to cars completely changed the landscape of the world, how we acquired knowledge, how we communicated the speed at which we could do things, nobody in our lifetime has experienced this, nobody was alive. When we went from horses to cars. At this point, maybe there's one or two people.

Rob Howard

And I think the other thing is, you know, I often also compare it to, you know, the printing presses, right, going from scribes copying books by hand to being able to produce books and mass to electricity, being able to power commercial devices, right. So one of the things that, you know, we talk a lot about in our innovating with AI programs is that most of us were raised on what they call the iron triangle. And it's a Venn diagram, where you have good fast and cheap, and you can only pick two, right? So if you want a good and fast, it's not going to be cheap, etc, etc, right. And what AI has done is created a lot of opportunities for having good, fast and cheap all three at the same time. And that is what is really messing people up because they as service providers, whether it's coding or film or anything else, like, you know, we've built our entire careers around like, well, I'm good at fast and expensive, right. But now there's stuff that's out there, that's potentially good and fast and cheap, or at least good enough, and fast and cheap. And that is, you know, a major source of not only disruption to your, you know, career opportunities, but also ego disruption, which I think is, you know, what we're seeing, you know, in my industry and your industry all over the place around this, like, Okay, well, I thought that, you know, like, like, I really wish that this computer didn't exist and wasn't doing like an okay job at my job, right? Because now, there's all these, like, unknowns that are coming up. And like, you know, nobody really knows how it's gonna shake out are what the, you know, ultimate sort of new landscape is going to be but as you said, I think the one thing that I actually can predict with a lot of certainty is that people who are get good at using these new tools are going to be able to excel beyond the people who ignore the new tools, right? And that, I mean, that sucks, like, that's the work that you have to do, right? It's a detour that you have to take right in what you were planning to do. It's like you said, like, if you were scribe in the 1400s, and you're like, hey, this new machine is doing my job a lot faster. Like, it's, it's basically that type of situation. Even though we're all more advanced, and, you know, more digitally savvy and all that, then you know, the people in the examples from the past, like, it's a big deal. And it is extremely challenging to kind of get the psychology, right. I think you and I probably benefit from just having certain personality traits, and having honed certain things in our own business lives where we're like, Oh, hey, a new disruption. This is gonna be fun, right? But like, not everybody thinks that way, obviously. You know, so I think we have a little bit of a head start, just from a temperament standpoint, one of the things that, that I, you know, want to do is give everybody the equipment and the knowledge they need to make that transition when the time comes. Because, you know, we're getting kind of farther down this adoption curve now. And people who are not weirdos like us who enjoy like, crazy, new challenges, like are going to have to learn this stuff soon, you know? So, yeah, it's, uh, but you know, as you said, words like disruption, don't really do it. Justice, it's, it's going to dramatically change pretty much everything about how we use computers, how we, you know, do the vast majority of work and learning and that sort of stuff. I

Zack Arnold

wish that I could say I was in the exact same club as you, which is, ooh, disruption. Cool. This is going to be fun. I have more in the club of HA. More disruption. All right, bring it. Let's do this. I don't know how much of it's fun. Like I'm prepared for the adversity, bring it on. And I like a challenge. I don't know if I'm quite in your camp of oh, this is gonna be fun and exciting. I'm like, really? Alright, let's do this. That's that's kind of where I'm where I'm not. I'm not afraid of it. I'm not shying away from it. But I'm like, really? All right, fine. Let's figure out this AI thing. And let's teach people how to use it. And, but hopefully, I'll transition to your camp. One of the things that I want to bring up from what you said before that I think is so key that I want to get back a little bit more into the weeds of the creative process, quote, unquote, of GPT understanding decision making and how we can embrace it. But you said something about the ego. And I think it's really important for people to understand this, because for the last year, year and a half, there's been a technology crisis. There's been well, it's a financial crisis, or it's a political crisis or whatever. Right? And I say what we're experiencing is an identity crisis, my identity of what I du is who I am. And now machine is doing what I see myself as a living. I really think that's where most of the fear is coming from with AI is an identity crisis more than anything else.

Rob Howard

Yeah, 100%. And, you know, to give you a personal example, you know, I'm, I'm at a place with my software development business where I don't do a lot of coding day to day anymore, we have a team. So I'm really on more of the sort of, like, management sales side of things. However, I coded for decades, every day, I enjoy it, you know, I enjoy building stuff, right? So one of the things that I've been doing lately is coding with GitHub copilot as my literal assistant. Right? And, you know, that's, that's the AI assistant that, basically, you start to type, hey, I want to write a function that determines the day of the week, 14 days ago or whatever, and it just goes right into there, right? It's just like this, like really fast, smart, you know, 80, to 90%, correct? Coding assistant, right. And to some degree, I'm like, Oh, well, I don't have to do this anymore. And that, you know, could be kind of scary, right. But also, it allows me to do stuff that I want to do a lot faster. And, you know, one of the things that I've discovered is that for a senior coder using copilot, it's like 3x, to 10x, productivity increase pretty much immediately, and potentially even more than that, as they get better and better at it. And it's opened up the opportunity for me to do stuff that I just literally didn't have time to do before. So you know, I would say, Well, I really want to build this, like dashboard, just report for my staff, because I want them to see this data differently. But like, I just don't have the like eight hours of focused time that it would take for me to do that. And it's not my priority, right. But with get up copilot, it's like, two hours of like, leisurely work, you know, and it's like that, you know, opens up a lot of opportunities for me while also simultaneously literally replacing a lot of coding time, right? So for those of us whose job it is to do that sort of construction level coding, or that scaffolding level coding, like, it literally does replace a lot of that, right. So there is a huge disruption and also opportunities for like tons more efficiency. So it is a very tricky balance, obviously, I mean, the same is true, like, you know, even for things like editing, you know, simple videos and podcasts for the courses that I put together, like, I'm no longer sending that out to a service to do that, like, we just have AI tools that do 90% of that job really well now. So there's a lot of stuff like that, where people are not wrong to have that, like identity crisis and that negative reflex about it. In fact, people who are going to be in the most trouble are the ones who aren't even having the identity crisis yet, right. So if you're having an identity crisis, rest assured that at least you're moving in the right direction, you know, like you're noticing the problem, because there's a lot of folks, even in the tech industry, who are like, Oh, this is this isn't gonna take off, like, it's gonna have too many copyright issues, you know, they, they find all these edge cases that they think are gonna take the whole thing down. And like, yeah, those issues are all problems. However, they're not big enough problems to stop the like forward motion of what we're seeing.

Zack Arnold

What I want to dig into is something that you said not once, but twice that, I think is the key to managing the identity crisis and learning how to leverage AI and not be afraid of it. You said, it'll get you to 80 to 90%. And then you say, to get you about 90% of the way there. If your job now as a specialist is to do the first 80% in whatever the field or craft is, you're screwed. It's you being a generalist and a problem solver that gets in from 80% to 100%. And I'll give you the perfect example of this, from my perspective, where I almost instantly went from this is scary to Oh, I'm not worried at all. And I want to use this again, to help us better understand the decisions and the parameters and kind of the basic architecture. So I was experimenting with creating images and dollar using chat GPT for because I wanted a very specific image for one of my newsletters. And it used to just start with let me put this in Google kind of get something that's not great, but it's okay enough, right. So the example would be and anybody that wants to go onto the website and go to this newsletter, it was either one of the last newsletters of 2023, or one of the first of 2024. But I've always used this analogy that was very apropos at the end of last year, that we all just feel like this zombie carcass that's at the end of an ultra marathon crawling our way to the finish line at the end of the year. Right. And I'm like, I now have the tools to actually create a visual image of what's in my mind. So I said create image of a zombie at the end of a marathon crawling across the finish line. And the very first version wasn't bad. Like, holy shit, this is kind of scary. This is good. Now I want you to do this, this and this and the whole thing fell apart. Its ability to interpret feedback and iterate. If that were a real person Since they would have been fired on version two is like, Did you listen to anything that I asked you for? Have you like, as somebody that's good at communication, and either a creative world or a coding world, you know that somebody can get you a first version. But if they can't interpret feedback, they're gonna get fired. Right? And you know that from working with me personally, my feedback is almost always, this is fantastic. Now I've got 87 notes, right? Yeah, their team does amazing work. But I know exactly what I want. And I'm very detail oriented, and thank you and your designers for putting up with me. But I know what I want. And you guys get, you probably got it to 95%. Most people get it to seven, you got it to 95. But for me, we're gonna get it from 95 to 100. And I've got notes. So your ability to use tools to fill that gap is invaluable. GPT, and Dali, and all these other tools, there are at least as of right now, a disaster with interpreting both notes, but the notes underneath the notes. And my belief is that's what keeps our jobs safe. Our ability to communicate, interpret feedback, we can use the tools to get us much further, much faster. But it doesn't replace us, we still need people that can interpret the feedback and provide that feedback.

Rob Howard

Yeah, and you know, I don't know how much of a Star Trek fan you are. But if you've ever seen them interact with the Star Trek computer on the show is very similar to that, right? Like the Star Trek computer has all the knowledge of the universe instinct built into it. But it really needs like Geordi and data and all the other engineer characters to kind of pull out like the right pieces and put the puzzle together, and that sort of thing. So you're absolutely right, though, to say like, if your job is that first 80%, then that's a serious problem, right? Like, you really want to try to rapidly level up and find that place where you are really the master of the computer puppets underneath you basically, right, because you need to get to a place where like, you know, what we're seeing is, you know, senior developers, senior coders are dramatically more efficient, because it takes a lot of the busy work out of it, right. But a junior developer might struggle because they're not really able to take it from 90 to 100%. Because that is sort of the hard part that we haven't figured out how to automate yet. And I think we see that in every area. AI is great for things like eliminating writer's block writing outlines, you know, that sort of thing. But as you mentioned, it fails when you try to get it to that level of like, very serious prose or like masterful like art and stuff like that, you know. So, if you can figure out how to use it to speed up your first 80%. That means that you can spend more time on your last 20% and become even better and better and more efficient at producing that like, amazing, final product. That being said, like, even so like it is still going to negatively affect the job market as we know it today, it's just gonna be a question of who pivots and how to get to a place where they are now the like, AI enhanced masters of their domain.

Zack Arnold

And to go back to something you shared earlier, which I've talked about endlessly ad nauseam. And I can't remember the name that you gave it, but that triangle of fast, cheap and good, right? Well, you actually had a name for it. I've never known it had a name. It's called the Iron Triangle, idea, Iron Triangle I started doing I don't know that either. I've talked about it for 15 years. But now I'm glad that I have a term for it. Right, where up until now fast, cheap and good, didn't exist. And now we're at the point where we can have fast, cheap and good enough all at the same time. Right? We're we're gonna make a living. And we're going to excel is making things fast, cheap and great. So if we focus on making things fast, cheap and great, and let the technology get us to fast, cheap and good enough, that to me is where we're going to see those that can learn and grow and pivot and survive and thrive in this rather than I'm going to be replaced. And the there's nothing that angers me more than like you said, those people are like avice is never going to be adopted are the ones that make me even more mad as we must legislate and make sure in our contracts that this technology isn't allowed. I'm like how myopically stupid are you to think that a contract is going to stop this level of evolutionary progress amongst the entire globe, it's so asinine to think that that would be the strategy versus let's protect the areas of ourselves where we do need boundaries, but then really embrace this to make all of our lives better. Write one of those areas for me, and I'm gonna, I promise not to turn this into a TED talk, because I really want to learn more about how you are leveraging and innovating with AI. But the short version of this is that when we become infinitely more productive, it's not a matter of think of all the extra time that I have to be creative and go home to see my family. Like, none of that is going to happen. The deadlines are just going to go, Well, you've got the tools and you can do this fast. That to me is what we protect. Those are the boundaries that we fight for mercilessly not trying to eradicate or we can't use this technology and it should be outlawed. So it's not that there shouldn't be protections but it's around copyright. It's around so stealing people's ideas that's around plagiarism. It's around safety. It's around boundaries. So we still have time to do our work and not be exploited, protecting us from the technology itself. That is just such a losing battle. And a year ago when I shared that, I mean, we're talking the torches and pitchforks. Bob came after me now they're like Richard Gere. Oh, yeah, there's still a little bit of that, but it's gone down a lot. But it's really embracing this idea of like you said, you know, those that are using AI are going to excel versus those that aren't, but there's no version of at least for now, I think AI is going to replace an entire field AI is going to replace people using AI will replace people not using AI. Right. So

Rob Howard

there's a significant significant open question, you know, industry by industry about, like, how much compression there will be, right? So, really, the question is, you know, how elastic is the demand for the thing that you're creating. So, if you're creating Netflix shows, there may be a very elastic demand. And we may just say, hey, guess what, like, if we can create Netflix, Netflix shows three times faster, and they're just as good, we're just going to have triple the amount of stuff to watch, right? And that way, you actually don't see a lot of compression in terms of the total employment in that industry. Whereas if you see something like, Hey, it now takes, you know, 1/3 of the time to make a coffee, because the robots are doing it for us, right? Then there's only so many people who are coffee customers, you're not gonna necessarily see a desire to triple the number of Starbucks in the world or triple the number of coffees being sold. So there, you actually will see employment compression, right. And I also think that tech encoding is somewhere in between those two, like, I think that there is more demand, there is unfilled demand to some degree, but I don't think that there's enough unfilled demand for new software to be developed that we will not see any compression in the, you know, software development zone. Right. So, you know, thinking about that, you know, those are, there's no, like, clean answer to that question. But you can think about your industry and be like, well, like, if suddenly, I tripled myself, right? If suddenly, there were like, two more Zack clones, right? Which is kind of what's happening, you know, if we're working

Zack Arnold

on that behind the scenes, by the way, and that's one conversation I was having with you, but go on.

Rob Howard

Yeah, future state? You know, if there were two clones of me, like, Would I be able to bring in triple the work and get paid for it? I mean, I think for all people, the answer is yes. And they would feel like pretty confident that they could do that. It's not the case for every person in every position. So that's where we're going to see the disparity between like, jobs getting eliminated or reduced versus people who are doing great work just producing more great work, right. So you know, I think that's, that's kind of a, you know, that's kind of a prompt for people to go become really great at a thing that is valuable, right, which is always a good idea. But especially now, when you're in a position where you could then potentially double or triple or quadruple yourself that is, you know, even more valuable. Well,

Zack Arnold

that's the perfect segue to learning more about how we can innovate with AI, see what I did there. Um, and I'm gonna start with the example that you just shared, which is me very much as a student learning this as well, when you said, you know, tripling myself, we're literally figuring this out behind the scenes right now. Because I've realized that there's a version of fast, cheap and good enough, where if you need basic advice for me, let's say the firt, you know, up until a year ago with chat, GBT, somebody reached out and said, I'm struggling to figure out how to build my professional network and right outreach, right? The best I could do was me personally, or my assistant saying, why don't you check out these five resources on the site links? ABCD, right. And that's helpful to a point, then somebody comes with more nuanced needs, like, Hey, I've read these articles. I've listened to these podcasts. But I'm really struggling to do this one thing with my outreach, my only option is, well, why don't you join my program, and then you and I can do a one on one session, whether that's in a group session to make it cheaper, whether it's one on one, which is a my time, and it's going to be more valuable. That was the limit, right. But now I'm seeing this exponential limit where I can grow and change setting the right expectations and boundaries, where I can create what we're calling Zack GPT, which a lot of content creators are creating now, where based on the 1000s and 1000s of hours of audio recordings, coaching sessions, articles, classes, it can interpret to a good enough level. This is how I might answer a basic question. And this is how I might curate the following podcasts you should listen to or the articles that you should read. So as long as there's transparency, I can say listen, I can help exponentially more people as long as you understand. It's not as good as working with me in person. But this will get you started. Those are the kinds of things that I'm learning how to do to innovate with AI. So I can take what I'm already good at and I can scale it. So now help people better understand how they can identify that for themselves. innovate with AI, knowing that you don't have to be a coder or a developer, or you don't need to have $7 trillion to take advantage of this new wave. So now let's really get into the heart of what you're actually doing with this latest business. Yeah,

Rob Howard

absolutely. So I'll give you kind of the high level overview of what we teach in terms of idea generation and building out, basically rapid prototypes, right. So some of the core pieces of that are number one, as you mentioned, there's just a ton of no code and low code tools out there. So it's really not necessary for you to like go hire a skilled developer, or be a developer yourself, like most of this stuff can be done. Basically, with drag and drop at this point, especially at the prototype level, you don't really need to go hire custom software developers until you validate your thing, you have some customers for it at a very base level. And then maybe you want to go out and build like a really fancy like website interface or take E commerce transactions or something. But the first thing is that the no code tools that existed even before AI, have already pushed that threshold farther out into the future where you need to bring on a coder. And one of the things that I do with our mentorship students is I literally just like, walk through the no code builds with them, and they see what I'm doing. I know more about it than them. But like, they can do this too. Like I'm not coding in Python, or JavaScript or anything like that, as part of this, you know, process, you can do it that way, but it's not necessary. The second thing is that there are really two different paths you can take. And you can take both of these paths at different points as you go through. But one is a product, which is going to be a customer facing thing that someone else pays you for whether it's 20 bucks a month, or 1000 bucks a pop or something like that, right? So the sort of Zack chatbot idea is ultimately going to be a product, right. But there's also an option to do a process, which is defined as an internal thing that makes everything faster and better. Right. So for example, our designer might have an internal process, where she takes all of your feedback, and has AI coded into categories, and maybe even generates, like different color schemes, or imagery or whatever, for sort of the vibe of what you're looking for. That is never something that she's going to go directly sell to a client or a customer. But it dramatically speeds up the work that she's already doing for pay, right. So the process is actually, for most of us an easier area to start, because it doesn't require you to go find an audience, it doesn't require you to think about pricing. It's just saying, Hey, I'm already selling something, whether it's web design, or coaching, or, you know, we have people who are doing this around client services, or that are outside of the world of tech, for example, wellness and health consultants, right, where they're saying, hey, you know, my process is going to be my clients are doing this big intake questionnaire. And then I am in turn reading the whole questionnaire and writing a 30 page report with recommendations on it. Well, that takes me 10 hours. But what if I just built something that had all the information that I have in my head on paper, and asked questions about what the person had submitted. And what they found is like, now they can do those recommendations in like an hour instead of eight or 10 hours, which is insane, you know, productivity increase, obviously. So thinking about product versus process is one sort of like, place where your path might diverge, you can also have a process that becomes a product someday, right? Or you can have multiple of these things happening over the course of the year. But for all of them, we have something that we call the AI business model matrix that has kind of four components that you have to have together. The goal being I want to create something that is special to me, and that creates the proverbial moat around your idea, right? So that it's not just so generic that like anybody can do it. If you ever see like the lists of like, Hey, here's 1001, chat up tea ideas, like, those are fun, but like, the problem is, those are for anybody to do. And I've even seen things like, hey, what you should do is go on Fiverr and then charge $100, to write an ad for somebody, and then just as chat should be to write the ad and then copy and paste it back to the person who made $100. Right? And it's like, that literally is the thing that they're promoting as, like the good idea, right? So what you want to do is move away from the very generic stuff, and instead, figure out what your unique subject matter expertise is. I mean, this is kind of like what we've been alluding to throughout our entire discussion, right is, you already have something that you're really good at, like, if you've been in your career for 10 or 20 or 30 years, like you're really good at probably multiple things, right? You probably even have something in your personal life that you're really good at, right that you have honed your skills on already. So take that thing and combine it with some of the AI stuff that's out there and some of the traditional stuff that's out there to create something that is essentially more than the sum of its parts, right? And well, when we do this, it usually, like, sets off the light bulb over people's heads. And they're like, oh, you know, I actually am, like, really good with nutrition, or I am really good with, you know, we did one around website accessibility, because that's something that we do for every single client. And you know, there's things that are very tricky about it. So we started teaching AI tools, how to help us with that, right? You know, there's stuff like, I mean, really anything, right? We have people who are doing, somebody created a historical biographical video of their great, great, great grandfather recently. So there's like a genealogy, I'll ask her that there's a video aspect of that. So you know, that idea generation is something that I think a lot of us miss, because there's a lot of these like, very, there's a lot of very generic, big stuff out there, right. So what you want to think about is like, well, I don't want to do the thing that Google also wants to do, because that's way too big of a market. It's way too expensive and challenging for me. But what I do want to do is say, hey, yeah, I'm really good at this niche task. So now I'm going to get a figure out how to get AI to help me with that task. And then I'm going to use that to make myself more efficient. And then eventually, I may also sell that to other people. We're also seeing a lot of this with like, physicians, you know, like, I'm really good at these clinical reports that I do for my patients. But I want to get faster, because a lot of this stuff is just be essentially like reading back test results, right. And then there's a little bit of like, a consult consultative recommendations at the end. So I want to get faster at that. But also, I can then go turn around and sell that system to other physicians who are in a similar place, right. And there's all sorts of stuff like that, where we're seeing people say, like, hey, lightbulb, like, I can get better at this. But then also, I can sell the thing that made me better to others. So that's where you have, you know, both a process and a product, right. And it doesn't have to be like these, like, you know, if you think about the last 2030 years, it's mostly been like these, like mass market, you know, Facebook, Instagram, stuff like that, or Airbnb, like, these are the startups that we think about as like successful tech startups, what I think is that, there's actually going to be a huge opportunity for people to create really great niche, AI powered businesses that don't need to be as big as Facebook don't need to be as big as Airbnb, they don't need to be like, mass audience, that's where a lot of the investment has gone over the last couple of decades, because that's what you can like profitably scale up, but that equation is going to change significantly. And there still will be these big infrastructure providers, like the Googles and metas of the world. But, you know, we're in a place now where we have access to this infrastructure, and it can 10x one person's productivity, but then that person can also turn around and share that with others, which becomes its own line of business. Right? So that's kind of the framework that that we're teaching. And, you know, the the other thing that is not really new, but it's kind of like old fashioned agile entrepreneurship is that you don't really know what's going to be a hit until you get out there. Right. So one of the things that we do is, we really encourage people to rapidly prototype we set a limit of like, if you do 10 hours of work, you have to stop and you have to try to sell it to someone, right? And if you don't get the sale, then you need to, like pick your next idea and keep iterating, right? Because you really don't know, like, even you know, for me, I've iterated through probably five or 10 ideas in the last two years. And a handful of them have been good, but I've also kind of just, like, thrown some of them away and been like, yeah, that was a cool thing to build, but it's not going to be the thing that I want to scale up, you know. So just having that iterative and agile philosophy, I think is more valuable than ever, because of the extremely rapid pace of the change. And because as much as it seems like everybody's talking about AI, like, only like 1% of the world is thinking about this right now. You know, like we're in that you know, zone. So it seems like everybody around us is obsessed with it. But that is definitely not the case. And we still are, you know, very early adopters in this. So anybody who is interested in this now is going to have that early mover advantage for a significant period of time. So it's kind of a nice, you know, it's a nice moment, if you are interested in this and you are willing to invest some time and energy and feel stressed out by the like unknowns of the future. But if you can do all those things, then, you know, it puts you in a really good position to kind of reshape your career or lines of business around these new opportunities. I

Zack Arnold

agree with all that except I would argue and I have no data to back it up. It's far less than 1% of people that are thinking about this. I think it's about 99% of people are like oh my god, this is terrifying. What the hell is going on. And there's point 1%, that are saying amidst all this disruption and terror and chaos, there is a once in a generation of not a once in a lifetime opportunity. So I think it's less than 1%. But you and I are both in the intersection of that 10th of 1% 1%, whatever it is. And this just, it just goes to one of the other concepts that I teach where I think it's so aligned with what you do, which is understanding what is it about our human intelligence that makes us future proof from artificial intelligence, and it is this unique intersection of my life experience, my work experiences, my skills, my abilities, my passion, my knowledge, and how can I package and integrate that either as a product or as a process or both. So I want to, I want to give an example, that's very, very relevant to me just to bring it from concept to execution, right. And I kind of talked about this idea of Zack GPT already. But if we talk about taking what I believe is one of my unique specializations and turning it into something I can scale similar to you, it's my ability to take a lot of complex information and simplify it. And I've been saying for years, and I truly believe that the value of information is now worthless, because everybody has access to it. We think it's all about let's get more information out there. But everybody has access to information, the value is now and the ability to curate that information. So what I've been doing for years is I do a 30 minute zoom call, and somebody shares, here's what's going on in my life right now in my career, or my whatever it might be. This is where I'm really hoping to go where I'd like to go next. And here's what's stopping. And with those three inputs in less than 30 minutes, I can say, I think this is the one thing you need to work on next. And it's different for everybody. Some people need to work on cold outreach, some people need to work on a resume, some people need to work on time management, some people need to get better sleep. So there isn't a one, one stop fits all, here's the curriculum in this order. And I thought, what if I can build a GPT that curates this customized learning path based on these simple inputs? Now, it's just not how many 30 minute zoom calls can I fit into a week, which by the way, is about 30, or 35, which is exhausting. And now no longer scaling. But imagine if I could do 1000 of those a week, knowing that there's going to be a percentage of them where it's not enough. And they want to get on the call. And we want to figure it out one on one, great, but that's me where I'm taking a process. But then imagine I take that process of information curation, and I take it to a whole bunch of other coaches that have really vast and deep libraries of content, let's take our, again, our mutual mentor remains at, he's got a bunch of products. The only way I know if it's right for me or not, is I read a sales email. And it's like, oh, that's me. And I have that problem. But what if I miss that day sales email, and I say, I want to use all your products, but I have no idea where to start, boom, 15 minute process, answer these questions. Here's your curated customized learning path, that's me taking a process to make what I'm doing more efficient, that then becomes a product for other people where that makes them more efficient. So that's one of the things that I'm looking into right now. And I've got a whole host of other ideas as far as like the podcast space, too. But I want to, I want to come back down out of the clouds, for those that don't already have 1000s of hours of podcast content and articles and newsletters and coaching sessions. If you're like, I have none of that. What am I supposed to do? So I want to I want to come back to the basics of I've really only been doing one job for most of my career, I wouldn't even know where to start. How do we start figuring out how we can innovate with AI? And what are these other kind of three areas of this matrix beyond just finding out what our unique talent might be? Yeah,

Rob Howard

great question. So to think about, you know, a few unique talents. The other components are a Stefan effect, essentially, which models and tools you're gonna use, and then which traditional thing you're going to combine them with, right? So for example, one of the things that I'm working on combines ChaCha beauty with Gmail, right. And it does a much better job of sorting email, and filtering out cold email pitches, sorry, to everybody who is sending out cold email pitches, but it's pretty good at it. And, you know, this is like, you know, you have something new your something old school. And then you think about the fourth component is which platform you're gonna use, whether it's web application, mobile app, Chrome extension, you just do it for the manually as a consultants, right. So those are the four components, the new thing makes the old thing mixed with your expertise mixed with your delivery mechanism, mechanism or platform. What comes out of that is if you have your expertise, but you're struggling to sort of figure out like, where to begin, right, like you've been a professional on a certain career for 10 or 20 or 30 years. The way to do that is kind of similar to what you might do in like a time management exercise, where I want you to start breaking down what you're doing into the smallest pieces possible, because where people fail with AI ideation, is they say stuff like hey, I really want to Find the right AI tool to like write me an award winning screenplay, right, I have this idea. And I really want to turn it into a beautiful screenplay. And it's like, this is not possible right now, I think that's way too big of a task, what you want to do instead is find the small tasks, and then systematically build those out. So you don't have to do them anymore, or they take like 1/3 or 1/10 of the time. So that's where we see our students succeeding is breaking what they're doing down into smaller pieces, obviously, that can be valuable in a variety of ways, whether it's for scheduling, or just time management, or efficiency or delegating, right. But if you do that, and you imagine yourself, delegating tasks to AI as if you would delegate them to an assistant, right, then you're putting yourself in a place where you actually have feasible things that you can create, that are that fit within those sort of bite sized chunks of your overall process, or the overall thing that you're creating, with your, you know, week of work or your month of work. So, for example, there might be elements of sales proposals that you don't really need to write yourself anymore, right? Like, you can say, hey, I want project description for XY and Z, you put it together, and it's done 80% of it right there. You know, I had an experience the other day, where a client asked for a specific type of documentation for a task. And I had the brainpower to write like, sloppy notes about it. But then I just put the sloppy notes into AI. And I said, Hey, please write a two page document for a non technical audience explaining this, right. So if you start to flux, stuff like that, out of your workflow, especially stuff that you're doing over and over again, that's gonna be an area that's ripe for you to build an AI process into. Now, that's not necessarily going to change your entire life overnight. But those things start to compound pretty quickly, right? Where all of a sudden, you're saving four or eight hours a week, but you're getting the same amount of work done, right. Or you are developing processes that can be shared or sold to other people who are in the same position as you, right. So not every single one is going to be a hit. But the experimentation is really where the, you know, where you open up opportunities to kind of win big and discover really cool stuff with this. And also keep in mind that like, you don't have to go out and build a billion dollar business, right? Like, if you built $100,000 business, using your existing expertise, and no code stuff that's already out there. That would be a huge, like, close to career replacement for, you know, the vast majority of people, right? So those types of opportunities are there because you're, you're moving early, and you're you're ahead of the game. And you know, that's where the experimentation comes in. And you know, the ability to sort of like, pluck out the small pieces that can be really nicely automated or improved or sped up with

Zack Arnold

AI. Alright, so I realize we're running out of time. But there's one more area that I think would be tremendously beneficial to make this a little bit more practical. Because it's really important that in any coaching call, any article, any newsletter, and most importantly, any podcast. It's not just that was fascinating. What the hell do I do next? Like I want a real sense of concrete action steps. So let me break this down into something fairly simple to make sure that I understood what you said. And then I want to figure out what my next step would be. Right? So one of the ways that I use chat GPT the most is as a writer, anybody is thinking, Oh, I knew it. His newsletters or AI. It's such bullshit. Never. And by the way, I've actually tried, I said, I want you to write me a 1500 word newsletter, here's the name of the newsletter, here are the topics that I cover, I like it to be about these ideas, gave a really detailed notes it spit it out, I'm like, I'm using none of this, this is garbage. But you just gave me some ideas, right? There are some phrases that you use that sound like me, and I start to pluck it out of there. So I love it as a writer's assistant, where I'll say I'm developing a new course. And here are the units and the modules. Give me three different versions of a summarization paragraph, or give me 10 different product names, and it spits all of them out. And usually individually, the ideas aren't good, but the combination of them and like I wouldn't use any of these and copy paste, but versions one and four together, I never would have thought of that. And that's brilliant, right? So that

Rob Howard

to me, that's a brainstorming writer's block tool, basically. Exactly.

Zack Arnold

So that's, that's a process that's helping me become a better writer and newsletter creator, content creator, right. But so far, I haven't seen a business model that I could turn that into because I'm just using chat GPT. So if I found a very efficient process, where every time I brainstormed a newsletter, I put together a podcast show notes for our conversation. I'm using somebody else's tool and I'm paying them. So how would I turn that process that makes me efficient into something that I can make that process a product? Yeah,

Rob Howard

great question. So there's a few different ways to do it. You know, I think one element of that is Probably it needs to be niched down a little farther, right? Because I'm not hearing necessarily the unique to Zack expertise in the idea yet, right. But I do think that it is in there somewhere or we could we could do deeper, right. And then what we do with a lot of our students is we actually create, you know, automations, using tools like Zapier and make.com, that combine these things together and do it fast, right. So in a way that you might be copying and pasting from ChaCha beauty into another platform early on. But when you kind of nail that sequence of things that you want to do, you then put that into an automation tool. And now you're doing that every time like a certain trigger happens, right. So like, we were doing one where I uploaded three files, that are just raw PDFs. Now those are going to then get processed by seven different prompts. We're going to repopulate that in this database, we're going to send it over to a Google document. So now you're starting to look at something that is more than the sum of its parts, and then has kind of your, like, special sauce in there. That being said, I think that there's even probably stuff you could pluck out that's even smaller than what you just described, like just doing shownotes, for example, or just doing like, outlines or one elements of that. Or things like the course curriculums, right, we're seeing some of the course platforms try to do that. Now. I've also had success with things like, hey, I want to create a quiz based on this module, right? For some reason, I get a lot of writer's block when I do that, but if I say hey, here's all the content, the module quiz, make a quick a 10 question quiz, like, that stuff can be very valuable, right? I guess my big picture advice would be, if you find yourself stuck, you want to go narrower and smaller, until you find the thing that is like, oh, yeah, that's me, like, people are gonna be like, Oh, hey, Zak's really good at that. And now he's making AI that's really good at that. So I want to pay attention to that. So I think that that's kind of the ticket for the smaller and more niche stuff. And where people stumble, is they say, Oh, hey, I wanted to create an outline generator. But there's a lot of those out there. And I don't have any sort of unique value proposition or anything that's unique to me. So we really try to hit that special sauce as hard as we can. When when people are thinking about those those new ideas. Yeah,

Zack Arnold

and I would say that a lot of that is just good general business advice where you know, don't don't try to build the next Facebook, figure out where can I be uniquely valuable to a very small audience. And then when you find that intersection, and that specialty, then you can broaden. I mean, that's exactly where I am in my business I started with exactly isn't health to TV editors, like very, very niche, and then growth, you know, very slowly grew and expanded. And now it's, I feel confident that if you work in a highly demanding creative job in any industry that requires creativity, I think we've got the tools that can help you. I wasn't there when I started, but I've been able to very slowly and gradually and meticulously broaden out knowing the niche and understanding my customer. So I didn't even beyond AI. I think that's really good advice. Last quick thing, which is really more of a follow up to the previous one. On a very practical level. If I wanted to experiment with some of the tech, I know that like Chachi btw, you can literally build your own GPT is kind of like an app store. There's Zapier. If I want to just get started and dabble with the basics of how would I build a processor, kind of what are the tools, I should pay attention to knowing that the tools are all there's this massive competition for who's going to kind of take those sectors of the market. But if I just wanted to just understand the playing field, what should I be looking for if I want to dabble with tools?

Rob Howard

Yeah, so we actually start people off. We have a no code toolkit that's free through integrating with AI. And what we do is we start them off with two old school tools, Zapier and Google Sheets. And the exercises that we do to show you how it works is you have a Google sheet and your Zapier automation, is watching that Google Sheet for changes. And then it's plucking information from relay it's sending us ChaCha btw. And then it's putting that result back into Roby, or rhoc, or whatever. And you would be amazed how far you can get with those simple tools, right? It is also kind of a trap. Like there's all sorts of websites out there that are like, Hey, here's the 1000 AI tools that are available for any particular task, right? It is a huge rabbit hole that people are going down and they're like, Hey, Rob, you know, I was just considering these 17 different voice generation tools. Like how do I choose between them? And I'm like, literally just roll the dice. Like there's no way for you to really judge these at this point. Like a lot of them are very early stage. So a lot of people get actually a form of writer's block from just picking the tools right, which is why we start you off, like super simple. Zapier has been around forever. Google Sheets is obviously like, Microsoft Excel is like one of the oldest pieces of software right that we're still using today. So you Using those is gonna give you a sense of what you can accomplish and how much you can accomplish. And additionally, just, you know, get the, you know, $20 a month charge CBT Plus subscription and just play with it in a very manual way. Like I think people instinctively tend to jump into, let's make this all scalable and automated, like way too early. And what we try to coach people to do is, hey, like, if you can do this manually for a month, and it's still useful for you, after doing it manually for a month, then you should think about building it out as like a faster automation, but you really need to be the, like, artist experimenting with the like, sort of blank canvas that is this AI system, right? Because it's not obvious what's going to work. And you can actually waste a lot of time going down the rabbit hole of things like tool selection, or, you know, building out a complex automation and then just scrapping it, because it turns out, you're not even using it a week later, right. So actually, the simpler you can start, the better off you're going to be in and as you mentioned, you know, most of this advice isn't like novel or specific to AI, like this is mostly like sort of entrepreneurship or productivity advice. But what we're seeing is that we've just opened up this whole new world of people who maybe have never been in this entrepreneurial space before. But now as early adopters, you sort of are forced to think this way, because of the nature of the rapid advancement. And you know, just how many options there are, and how many, like weird little traps you can fall into and waste a lot of your time. So, you know, starting small and simple is really going to be a very valuable tweak for you if you're finding yourself overwhelmed. Yeah,

Zack Arnold

I would say at all, that's invaluable advice. There's one quick thing I want to add, then we're gonna wrap it up. Because I get similar questions a lot, which is there 27 tools to organize my dailies or, you know, Dad label, whatever it might be right, like, which one should I choose? My advice? Wait. That's it, because eventually one of them is going to take over the market right there. The question was, should I be building my platform on MySpace or Facebook? Nobody knew at the time, well, if you'd waited a year or two would have been very clear where to put your effort, right. So if you have the time, and you really want to find the industry leader, let it find itself and wait. Otherwise, like you said, roll the dice just learned the process, right? If you spent all of your time and energy, building a compelling story and profile on MySpace, you have to redo your work. But you learned how to build a compelling story and tell you know how to build a nice profile, you just have to take that work from MySpace and put it on the Facebook, but you still develop the fundamental and meta skills of communicating who you are. So ultimately, the key is that you learn how to leverage the technology and learn the meta skills necessary before you worry about the actual tools, because the tools aren't going to change the game. It's how you use the tools to change the game. So thank you for coming to my very brief TED talk on that subject. Having said that, I want to at least wrap up for now and be very conscious of your time. But if people want to find out more about you, they want to find the free resources, I highly, highly encourage them PS no affiliate Commission's here. But if you want to learn how to leverage AI at the basic level of not having to know Python, and JavaScript, and all these other things, Rob is the guy Rob is my guy. And Rob should be your guy too. So if they want to find you, and they want to start going down this rabbit hole responsibly, how do they do that?

Rob Howard

Sure. Well, thank you, and it is innovating with ai.com. And then you'll find our no code API Toolkit, which is free. And that's kind of what I just touched on that gets started, gets you, you know, the tutorials that you need, and also the framework and philosophy, right, which I think is we try to be very different than that and really avoid the, like 372 business ideas that you can use today. Like, we don't do that stuff, we really focus on, you know, making real progress and being really personalized and real and authentic for you. So innovating with ai.com is where you can find all that.

Zack Arnold

Excellent. I'm going to make sure to put a link to that. And if I'd asked you how to leverage AI, and you said create a cheat sheet prompt with 738 Chat GBT prompts and build a Facebook funnel around it. I would have said, nevermind, we don't need to do this podcast. So I'm very glad that wasn't that your approach and your strategy?

Rob Howard

Yeah, I'm sure that's working for maybe one or two people somewhere in the world. But it was very well

Zack Arnold

for a little while. But yeah, it's not gonna last. So Rob, it's been a long time coming years and years. If you and I being you know, behind the scenes, learning from each other exchanging ideas, you saving my asked multiple times with my website. And I'm glad that I can at least start to return the favor and get you on the program and hopefully send some students your way so they can learn how to innovate with AI. So on that note, yeah, well, thank

Rob Howard

you so much. It was wonderful to join. Yeah,

Zack Arnold

you bet. And I'm pretty sure this conversation is going to continue for a long time to come. So appreciate you being here today. Thanks so much. Thank you.

Transcribed by https://otter.ai