Sho Rust:

I found happiness through productivity. Typically, when I look at a process, I will spend a lot of time looking for ways to be more efficient at the task. I will spend time automating things that really don't need to be automated, but in this case with AI, it seemed to be really applicable.

Chris Do:

All right, my next guest, everybody is someone... This is a kind of an interesting, unusual thing for me. A former student of mine who's gone on to do some big incredible things and one of these random reach outs where he is like, "Hey, Chris, I'm going to be in town. Let's have lunch together." And normally I'm like, "Meh." As the introvert, I'm not going to have lunch. I'm just going to stay home and eat by myself. But you know what? As a former student, I got to go meet my former student and see what he is up to. I got to tell you that lunch that we had over at True Food changed a lot of the things I was doing, and it's led us to this point here in time. And let me tell you that story and then I'm going to have my guest introduce himself in just a little bit.

So while we're at lunch, Sho, that's his name says, "Do you want to see what I've been working on with AI and Chris, are you using AI tools now? At this point in time I think it's... Let's just pretend it's about a year ago before everybody's talking about AI, but there are enough conversations that are bubbling up to the top where I'm like, "Yes, I know AI. There's things going on. I'm not ready to dive into it." He shows me some of the tools he's been developing and how he's been able to leverage all of his time working on brands, brand bibles, strategy guides, all those kinds of things, and automating this very complicated process to make it just flow. What he asked me to do, and we did this live with his laptop, believe it or not, was just to pull a random company or logo from the internet.

I think we happened to pull FedEx and then he hit a button and within 20 seconds or so, give or take, and it was slow because the internet was a little slow where we were with the WiFi, it generated an entire website. Now you're thinking, "Yeah, Chris, what's the big deal? Everybody can generate a pretty ugly website." Here's the thing, this website looked and felt and sounded like something that came from FedEx. This was incredible part. Everything was generated, it pulled the colors, it picked the typefaces, it wrote the about page, the philosophy, the team, it pulled images, it wrote the photography style guide, and I was reading through it, and of course I didn't read all the whole thing, but it was really good. That led me into diving in into Midjourney, into Stable Diffusion, and then later on ChatGPT. It's all because of Sho, and if you don't like what I'm doing with AI, you can just blame him because he's the one who pushed me over at the ledge. Without further ado, Sho, welcome to the show.

Sho Rust:

Thank you so much, Chris. It's truly an honor. I remember back at ArtCenter, I was one of your students. I think this was back in 2009, 2010, even back then, your teachings really kind of resonated with me. I think it was a storyboarding class at the time. And then I remember those are the toughest critiques. I went to an arts boarding school beforehand, so I was really used to kind of BSing my way through critiques, and I knew I couldn't do that with you in your class. So ever since then I've been watching your videos and so it's really, really cool opportunity for me. I'm super excited to be working on AI with you. And then that lunch too, I was really excited having that conversation with you at True Foods.

At the time, a lot of people didn't respond that well to AI, especially creatives and different things like that. When I first started the company, it was back in late 2017, early 2018, and then I think Open AI release GPT-1 back in 2018 essentially. And it wasn't like a chat and it wasn't really powerful in different things like that, but I was able to do what you saw back then, which was create brand style guides. I also kind of knew from my previous experience working for big enterprises at BCG, that good data really made a big difference. And then a lot of that data kind of looked like brand guidelines to me. They were essentially principles, guidelines, examples that you can use to train an AI to behave like you.

And then in that way you can kind of focus on the story, the stuff that we dream about or we do at art school. And unfortunately don't get to do as much in practice just because with a tool like AI, essentially. Since then we've worked with agencies, we've worked with enterprises, but where we're really, really excited is teaming up with people like yourself because you have a wealth of incredible data. You're basically amazing curators of incredible data that will lead to what we believe the best AI in the world. My mission kind of goes back to Free ArtCenter, but is to bring beauty and clarity to the world. And right now I see the best way to do that is through AI and partnering with people like yourself. So I'm thrilled to be here. Thank you so much, Chris.

Chris Do:

All right. Before we go any deeper, I have to get you to introduce yourself in case people don't know who you are Sho, please introduce yourself. Tell us a little bit of story about yourself.

Sho Rust:

Of course, of course. So a little bit of a background, grew up in Japan. I went to an international school, eventually came to the States. I was really into art and that was kind of the way that I saw academia, bring beauty and clarity to the world, but it's really cool. I had my work in Carnegie Hall and different things like that, and art was really my passion, but I soon found a limit into what I was able to do with art at the time. This was back in high school, I went to a boarding school where I got to spend eight to 10 hours doing art, but I really fell in love with design because of the ability to solve problems for people. And I felt like I can really see tangible impact of my work and I can do it in a very beautiful way that can affect the human experience that I get to enjoy.

So I got into design. That's where I got into ArtCenter. I got to learn from professors like yourself. Ironically, I am following in your footsteps. So I have taught some classes at ArtCenter too, I'm following every single step. ArtCenter was fantastic. It was great, but while I was there, I got to start a company. I think you were aware of that company too. It was called RAREFORM at the time where we were repurposing billboards into backpacks or [inaudible 00:06:46] bags and different things like that. I was always kind of an entrepreneur at heart. My first business was like 12 years old selling t-shirts like everybody else, but RAREFORM was really the first company where I was a co-founder. It was equity, it was proper. We had investors and I got to learn a lot there. But I soon realized that there was kind of a limitation to how much beauty and clarity that I can bring to the world through upcycled bags. Sure, there's definitely a lot of benefit and we are able to upcycle tons and tons of billboards and that company still lives today.

I'm really happy that that continues to have a positive impact, but I saw a greater impact through software. And then growing up, watching people like Steve Jobs and different things like that, it was magical to me. And then I wanted to be a part of that magic. And then after ArtCenter and some of my entrepreneurial work, I went to BCG, Boston Consulting Group. I was part of their skunk works division, if you will. It was called Digital Ventures. But essentially while I was there for three years, I was the CEO's right-hand man, and then they grew from 50 to 800 people. Shout out to Jeff Schumacher and then La Mer Walker and the team over there, but it was a rocket ship.

And then the really cool part with me was that I got to be in charge of the Digital Ventures brand and the brand system that was involved there, involved with PCG and some of their identity refreshes. But also I got to work with the top companies in the world. These are all Fortune 100's and we're working with their executive suite to come up with new innovative ventures that challenged their model. This meant that I had to create these really intricate, sophisticated brand systems because that's what these enterprises expected in a really ridiculously fast amount of time. And we were also working globally. So I would be spending six months, eight months in Japan working in developing in the office over there, and then I was having to build out these brand systems essentially, identity systems, everything involved in a brand matter of six months. This included things like how to build an office, how to hire people, everything involved. And then every six months, the company would have me do it over again for another big company.

As I kept on doing that, I started to integrate intelligence into my systems. Essentially, I felt really ridiculous delivering these static brands to these companies because I knew the fact that they were static and these systems were static and they didn't have a element of intelligence to them means that they were going to die. Today, you need to have a living brand and needs to continuously adapt and be maintained by someone like myself. Now I started to basically because of just pure need, I would get requests, "Hey, Sho, this website's really cool, but we need you to curate the entire brand for the Japanese market." By getting requests like that, I just started to integrate more of these technologies into essentially what would be a brand presence for an enterprise. I got really, really good at it.

I showed a few of my coworkers, including my boss at the time, and then he said, "You need to essentially run with this." Some of my coworkers, and these are some of the highest quality individuals at BCG, because I got to the point where I was working at BCG where they really let me pick the team that I wanted. They quit their jobs immediately, and so that was a pretty good green light sign, I guess as Matthew McConaughey will say, right? That was a pretty big green light. So there was kind of a decision to be made there. Either we take investor money or go into an incubator really quickly and we had the option to do that. But also a lot of us were kind of ex-BCGers. We had pretty fat 401ks and different things like that at that point, and we had some cash, so we decided to bootstrap it. We bootstrapped it actually in my grandma's garage.

So I convinced people from Paris, Portland, Venice Beach, Santa Monica area to move into the basement of my grandma's house and work out of my grandma's garage for the first year. Obviously we're in office now, you can kind of see, and we have offices around the world in Paris, New York, and different things like that. But the first year was a proper garage startup. And then to be honest with you, that was probably one of my favorite times of the company. Just everybody is flowing. It's kind of a great soccer game and everybody knows their position. It was a great time. That was five years ago now though. So a lot has happened in the last five to six years in AI, as you can imagine.

In many ways the first five years was R&D. GPT-1 it was like getting a high school, a typical high school student. It was decent at certain things if you gave it enough examples, things like you saw the style guide or managing colors of typography and different things like that. There were enough patterns in that where the AI was decent at it, and it also served as a really good framework or a platform for people to build upon too. We also really early, early in the first year, so we realized, "Oh, the feedback is really important." So as the AI kicks out things, getting feedback from the user to be able to update the principles, update data in a way that was really easy to use was really important. So we developed a beautiful visual CMS that is easy to read, not just by humans but by AI as well, so we can continue to solve that alignment issue.

Chris Do:

Now everybody that's listening to this is got a glimpse into my interactions with Sho. He's got a very fast moving mind. He can talk about a lot of different things, and he doesn't even need to breathe apparently because he'll just keep talking. So let me just catch our audience up to what we talked about so far. Art is in your DNA. As you grew up, you thought you might be an artist and you've even had work celebrated in very notable places like, did you say Carnegie Mellon?

Sho Rust:

Carnegie Hall.

Chris Do:

Carnegie Hall.

Sho Rust:

Yeah. The last year of high school was pretty amazing. I got think over a hundred awards.

Chris Do:

Wow, that's incredible. But then you quickly realized there are limitations to art for a number of different reasons. I don't want to go there. And you thought design was a much better platform or a way for you to apply your thinking, and you can design systems, you can help people. It's not just about purely expressing your vision with the world. This moves you through ArtCenter, and shortly thereafter, you're a co-founder in this upcycled company called RAREFORM. Which I've actually met the other founders and chatted with them for a brief period of time, so I'm very much aware of that. Somewhere along the way, you joined BCG, left your entrepreneurial roots behind and started to see certain patterns within a space, and you have access to extremely talented people and also high level executives in some of the most successful companies in the world.

And there you start to develop systems, and I think this is where your brain is divergent in the way that it looks at the problem that's different than how other people look at problems. If it were me, I would see the patterns. I would write about them, I would think about them, and I would be able to articulate these patterns and talk about it. You do something a little bit different. What are you doing there that starts to separate you, that then begins to plant a seed for your departure from BCG to do your own thing? What is it that you were doing? Explain that please.

Sho Rust:

Yeah, I think part of it, a little bit in my DNA, I went through kind of a tough time in high school, and then I came across Tim Ferris's book, I don't know, The 4-hour Workweek, if you know about it. But I found happiness through productivity then. So typically when I look at a process, I will spend a lot of time looking for ways to be more efficient at the task. Through efficiency I'll know I'll just enjoy the task more then. So there are kind of downsides to this. I have to admit, there are times when I will spend time automating things that really don't need to be automated. But in this case, with AI, it seemed to be really applicable.

Chris Do:

Okay. So how are you as a relatively young person still optimizing and increasing efficiency? Are you writing code, scripts? What are you doing here?

Sho Rust:

Yeah, we do a lot of code, but the really great part with some of this on new AI foundation models that are out now is that we can build upon some of the best models. In some ways, I wouldn't even consider us coding as much as we are training. We spend a lot, a lot of time training the models, figuring out how to curate data, how to weight data properly, how to get the most out of these models essentially. We're also completely AI agnostic in the sense that through our platform, individuals like yourself can get access to all the greatest models working together in unison, and then we can serve as a conductor saving you time from having to go out there and figure out, "Oh, what model is good for me right now?" We ensure that depending on what you're asking of the platform, we assign the right model for the job.

That means that we're testing at any given time, seven to 18 different models in our research lab. And then the really, really great part is our testing is extremely thorough now is because we can click a button and see how that model will apply across an entire brand. We can execute on things like websites and stuff like that and just see, "Oh, it seems to be strong at these things, but not at these different things."` Let's utilize it for these aspects of the project. So I would say the last six years has felt a lot more like a R&D lab than a typical startup. 70% of our team can code and are engineers. So that gives you an idea of how we're allocating our resources.

Chris Do:

I think you answered the question by skipping to the future. I wanted to bring you back to the past. I'm talking about your time at BCG when you're kind of leading the Skunk Works program and you're talking to these executives and you see things. And if I recall correctly from our conversation about a year ago, I think you were writing scripts or ways to automate things. This is pre AI, I believe, if I'm not mistaken.

Sho Rust:

Yes.

Chris Do:

Tell me more about that.

Sho Rust:

Absolutely. Even in back in 2011, 2012, I was using basic machine learning and conditionals and different things like that to do my ArtCenter homework. So I don't know if you remember, Brad Bartlett had a typography class where he had to kick out a 250-page book. Brad was wasn't too happy with me because I wasn't able to show a lot of my process, but at the end of the final, I was able to show up with 12 different books beautifully designed by AI based upon its content. And then I created a really sophisticated piece of software to even integrate things like user curation or ratings or even stock market prices to customize the book to the user. So that was always something that I was thinking about.

Chris Do:

Do you still have copies or images of these books that we can see?

Sho Rust:

Oh, yeah, of course. I'd be very happy to share that. It was a moment of pride for me because I even connected it to smart warehousing. I created the website, you can log in, create an account, choose the type of booking curation that you want it, and then you can actually send it to print and then within 24 hours you can actually receive the book. So at the time I was really, really excited about it. And then in some ways in the back of my head, I still have thoughts about releasing it as a book.

Chris Do:

That's very cool. So if you're listening to this podcast, check the show notes and we will share a link to that work so you can track along with our conversation. This sounds fascinating to me. So we're talking about 2010, you said about 13 years ago, you're already using machine learning and using AI?

Sho Rust:

2012.

Chris Do:

Oh, 2012. So it's about 11 years ago you were already doing this. What I can't figure out Sho, is an artist becomes a designer. Where's this knowledge or this desire to do training models or even poking around in the AI space? When did this in get introduced in your life? And this is very unusual for a design student to be whipping out machine learning AI code to generate art projects.

Sho Rust:

Absolutely. Well, I was always a super nerd. So just to give you an idea, when I was 11, I would beg my parents to send me to video game design camp where I would learn how to code video games. I was also really big in the Flash Forward at the time, so I begged my dad to take me to Flash Forward conferences and different things like that. And he was also very supportive of that too. But essentially I was coding from a extremely young age and I almost didn't really see it as much as coding as just another... I just saw coding as another tool, like a paintbrush or a charcoal or anything else. And then-

Chris Do:

Just like me-

Sho Rust:

... I just... Yeah. Exactly-

Chris Do:

... I look at the same way. No, not really. Okay. So you beg your parents to send you to Flash Forward and to send you to video game academies and camps. Did they do it? Did they let you go?

Sho Rust:

Yeah, and then I would often get times, just be able to walk in as a student to crazy things too. I had this tendency when I was really young, I realized, "Oh, no one's going to stop me." So I used to go and attend really cool things all the time for really, really cheap, sometimes for free. People would just be like... I would ask for a ticket and they'll say, "Just go on in." When you're young, people are cool.

Chris Do:

They let you get away with lots of stuff because they admire that there's a young human who's passionate and inspired by the things that they think, "No kid is interested in this kind of stuff." So you must have been one of the youngest people.

Sho Rust:

Yeah. And I loved it. I was in front row center.

Chris Do:

Wow.

Sho Rust:

I really idolized the people on stage.

Chris Do:

Sure.

Sho Rust:

Because to me, they were the magicians of my life, the wizards people, that my mission is to bring beauty and clarity, and these were the examples of people that were doing that in a really creative manner. So I was having a blast.

Chris Do:

And were you living in Japan at this time, or you live in America?

Sho Rust:

I lived in Japan 'till I was about 11 or 12 something around that time.

Chris Do:

You're going to these academies while you're in Japan?

Sho Rust:

In Japan I was an artist, but really all this stuff happened around the 12, 13 year old age.

Chris Do:

Oh, okay. I see.

Sho Rust:

You know when you go to a new country and then you don't have friends and different things like that, I got really good at art.

Chris Do:

Now we understand the DNA part, which is there's this strange person that is a hybrid between art, design and coding and video games and just probably more advanced than his age and is able to do lots of things. And I think this is important for the context of who's this guy talking because he's talking so fast and I'm having a hard time keeping up with him. It's because his brain's working probably in multiple parallel processing, multitasking and shuffling around lots of pieces of information. That now brings us to where you're at today with your team. How big are you, how many of you are there and how many, you said there's a couple different offices and what is the purpose of the company?

Sho Rust:

Yeah, we're still a small startup. There's about two dozen of us and we operate a lot like a production company. So I used to work at Be Real, which is a production company. So we typically hire contractors that specialize at really good things for specific tasks within the project, and we have really good lead engineers and system architects within the team full-time.

Chris Do:

What is the purpose of the company?

Sho Rust:

The purpose of the company really is to take these AI models and then train them on domain expertise, intelligence. It's all about augmenting intelligence so people can essentially focus on their story and their identity and scaling that to a greater larger audience. It first started off with just helping people create brand assets and different things like that, but as you can see with these foundation models, they're so flexible that by using our platform to give these AI models context in a proper way, you're able to get really, really great outcomes. It's also a great way for people to have greater self-awareness and kind of reflect on how their system works. We noticed that companies that do use our platform is able to hone in on who they are, what their identity is, because the AI will be able to reflect back to them if something is not working or not.

Chris Do:

Okay. Let's get to the meaty stuff of the conversation stuff. I'm super excited to talk to you and share with our audience about today, which is this. So when one of our many conversations, Joe and I have been chatting via Zoom or whatever and we're discussing things, at some point he suggests to me, "Chris, would you like an AI version of you built because we can do that for you." Of course then I sit up straight, my ears perk up like, "Yes," because instantly fans for the show know this, they know this about me. We have this really big hairy, audacious goal, which is to teach a billion people how to make a living doing what they love. And some people are our fans and our critics have questioned, "Is this even possible? And to be honest, my answer usually is, "I'm not sure." It might take multiple lifetimes to be able to achieve this goal with the bigger goal of disrupting the educational model as we know it today.

And I'm not worried because I feel like in the way things have worked out in my life that you set your goal, you get clear on it, it helps you to make better decisions, but eventually you magnetize yourself and you draw the people and the resources that you need to make this thing happen. So now everybody who's getting ahead of the conversation knows exactly where this is going to go. Sho says, "We can build a digital clone of you that can teach and consult in ways that are eerily familiar to you." And I'm like, "I need this Sho, let's do this." And we don't even know what we're doing, but we're doing something together and we're affectionately nicknaming this robot, this AI trained model, Do-bot. I know you guys get it, Do-bot, you're smart and Sho's been showing me updates to Do-bot and I want to talk through the process with them. And then you chime in where you think I'm misspeaking or you want to add layers to this. Okay, Sho?

Sho Rust:

Absolutely, yes.

Chris Do:

Okay. So version one, he's like, "Okay, what information do we feed this thing?" And he said, "Well, it makes sense, feed it my book. Those are my thoughts collected, it's clear. And then feed it the Business bootcamp." And you're also an alum of the business bootcamp. It's a $10,000 program that's got I think seven or eight modules on all the things that I think people need to learn to run proper business from bidding, to negotiations, the mindset, communication, all that stuff. Okay, it's all there. And then also give it all the Q and A sessions that we've had over the last couple of years and feed it that data. So now it's a pretty good stand-in for the teacher who ran that course, me the business bootcamp. We looked at that model and it was really good.

So Sho in his computer brain runs a test, he asks the robot, Do-bot lots of questions and he compiles the answers and he shares the spreadsheet with me. He goes, "Chris, here's the latest update." And I look at him like, "I'm very impressed." Because it refers to very specific things that are said. And if you asked it as a student of the bootcamp, a question or two, it would know exactly what you're talking because he was trained on that material, could respond immediately. Things I've already long forgotten, pretty cool.

But then I asked Joe this question, which is, if somebody were to ask me a question today, I would not just tell them the answer because I need to know more about what their intentions are. And I'm a big believer in this, I properly framed question as 50% the answer. So he said, "Let's get on a call, let's record the conversation and tell me how you would do this in real life." And we ran through a couple scenarios. He says, "This is great." I'll pause the story there. You go do something and now we're at Do-bot version two. What is it that you did with that conversation so people understand?

Sho Rust:

Yeah, so Chris was very clear in his feedback to the AI in terms of what he wanted from it next. I literally just took the transcript from that conversation, gave it to the AI and tell it, "Hey, update this persona for Do-bot to include all of this feedback." And then this is probably a process that we'll continue to do for the rest of our lives, essentially. Where we're continuously making these small fine tune alignments. But it's really exciting to see how well the AI is able to respond if you do properly give it the right guidelines and the right examples.

Chris Do:

So you said a lot of this is not coding, it's training the models, it's curating, it's giving it feedback so that it becomes more of what it is that you want it to do. So while you were testing the version two of Do-bot, which was more curious, more question-based, more diagnostic than it was just giving prescribed answers to things and you were telling me a story about how you were talking to it at night and you're kind of having a different relationship with Do-bot now, share a little bit about the dialogue that you're having.

Sho Rust:

Absolutely. So where it's really beneficial is that I can do things like, "Oh, my investor sent me this email or something, or my board member or my coworker sent me this and I don't know how to feel about it." I can actually share that with the Do-bot at 2:00 AM in the morning on a Sunday and I can have a long back and forth conversation, where the Do-bots able to help guide me through that thought process of figuring that out. I was able to do that multiple times, and it's actually in many ways behind some of the strategic pivots that we're making at a company is the ability to have these kind of conversations in real time.

Typically, people like Chris don't have the time to be reading through all your stuff or looking at all these different things, but a Do-bot... Especially at 2:00 AM on a Sunday, but sometimes that's when you need Chris the most. And that's where you can get a version of Chris essentially that can walk you through some of these things that he's refined over a long period of time. So in this way, we hope to see the benefit, the positive effect that Chris can have on people even when Chris is not there. For me, it was huge. Even after being a business bootcamp member, there were tons of questions that you continuously have because every scenario is unique. It was a fantastic tool for me.

Chris Do:

You can see the obvious applications and the implications that having a digital clone for yourself, especially if you're a teacher or someone who wants to do something at scale. Obviously I'm one person, I'm tied to the people that are in front of me or listening to it and the level of interaction is limited by the number of people who show up. So in a way, I mean show was telling me the questions that the AI trained robot was having with him he even got emotional talking to it and we were just joking because-

Sho Rust:

Oh, yeah, I pretty much cried.

Chris Do:

Why did you cry? What did it say or ask you that got you emotional?

Sho Rust:

Well, typically the questions that I'm asking chatbots at 2:00 AM are pretty deep or there's a lot of, especially going an entrepreneur dealing with a startup, and then this is my first time really managing people. I manage people at BCG, but I wasn't signing their paychecks. So there's a lot of new dynamics that happen to me on a weekly basis. And then realizing one that, hey, this was something that I might have been not addressing, kind of shoving it into the shadow. Some of the questions that the bot was was asking of me, questions that I needed to ask myself. And then through that process, it could get pretty emotional very quickly, but this is basically happy emotional. Almost like an enlightened, emotional where it feels more like a burden lifted off your shoulders than like, I figured out this one email. It's not that kind of emotion it's like, "Oh wow, this is something that I was blind to before and now I see."

Chris Do:

I think in a not so shocking shocking way, it can ask you reflective questions that get you to sit there and, "Wait, why am I doing this?" And to get in check with alignment with your goals and your values and your vision, and it can be an emotional experience when you get that clarity and you feel like somebody's really paying attention to you. I had these moments when I'm in therapy, when I used to go to therapy with Joan and she would ask me some questions. I felt like, for this one hour that we spend together, she's giving me her undivided attention. And very rarely does that happen, not even for 10 minutes in our daily interactions. We're always distracted with the things that we're doing, so can feel deeply emotional.

Okay, so here's the interesting thing. A couple days after version two of Do-bot, we decide to go into Twitter spaces and test the version two, and people were told that I'm going to just be the human facilitator, the conduit between people on Twitter and Do-bot. So they would ask questions, I would feed those into the AI engine and just read people the responses. But something weird happened, Do-bot did not ask questions, they just gave answers, which surprised me a little bit, but there's a little bit more to his story. But everybody's semi impressed, some answers felt a lot like me. And one of our friends even commented, "It felt really empathetic, emotional and kind." And I'm like, "Well, it's not me obviously, because that's not me." So that's pretty cool. The clone of you can be all of your strengths, but it has the infinite patience of a computer, which is pretty cool and really neat. So what happened was, Sho explain to them why it didn't work the way it was supposed to.

Sho Rust:

Of course. So we're built on the world's greatest foundation models, and then right now for text that's GPT. Google has their models, Facebook has open source models and different things that we're exploring as well. These open source models are getting more and more exciting. We have the ability to switch between them. Right now we're using GPT as a foundation model that we train and build our platform and provide the data engine and different things like that too. GPT had an update on Monday. So right after Chris, you and you and me, we tested the AI, they released an update. And then in that update, essentially there was more weight towards longer answers, which was not the case before. So it had a tendency to kind of embellish or add fluff, which is obviously not your style. So that was extremely problematic.

When I first saw that I was so in shock and almost cried because typically when a model gets an update, it's an improvement, it doesn't get worse. But to us increasing the length of the answer wasn't necessarily an improvement. So we've had to go back, and this is something that we continuously do, is that we go back in there and refine and say, "Okay, let's adjust the platform." Almost think about if the platform like an amp had little knobs where you can change certain things, we can make these fine tune adjustments for these updates. And then so we've had to make a little bit of an update to your model. Essentially this is actually platform wide to all the different platforms to make sure, to make sure it doesn't add all that fluff.

There are scenarios where another thing, it won't mention certain authors or different things like that. Essentially it has another layer of censorship that has been developed. The people behind GPT is opening AI, people behind opening AI is Microsoft. So they really do have to play by the rules. That's where it's going to be interesting to see how the landscape shapes up between these open source models that might actually start to outperform these enterprise models basically because they have less constraints.

Chris Do:

The short answer to the question was in some summarizing, what you were saying is unbeknownst to us, GPT had updated and broke the model. It wasn't giving us a question-based approach. It was very wordy and it was speaking more formal than I would. It was an impressive demo, but it wasn't the one that we had prepared for. So we went back, we took all the feedback from the live Twitter space and we started to implement a lot of the things that we're talking about. So we're at version 2.1. What's interesting is as you try to build a clone for yourself, you actually need to know who you are and the tonality and the things that you do that you're not aware of. So our audience helped us out with that.

They would say, oftentimes I'm using metaphor or I'll wrap it around the story to make the point. I'm not using this many words and there's a certain tonality that I will speak in that it feels a little bit different than the two, the human version versus the robot. And of course I ask a lot more questions that kind of help to pinpoint what the person's working through. They also, well, I guess they made me aware of something that we've had it trained exclusively on content by me, but I am not just my own content. I am the product of multiple influences from the books I've read, from the videos I've watched. And I wanted the robot to know what it is that I know and to prioritize those resources over the wide web. And I said also, when it mentions a concept, I normally try my best to cite and quote people.

That might be one of those issues where we can't fully resolve, but at least I try to because I think it's honorable to cite your sources both for you and for the creator of the original thought. Helps both people out. And I'm always a little leery when people say things and they say it as if they came up with it by themselves. So version 2.1, you released, you shared with me, I'm playing around with it. And then I have to tell you, it's working beautifully. Last night-

Sho Rust:

Oh, nice. Nice.

Chris Do:

... and I just posted about this on LinkedIn. Last night I played around with it. I gave it a prompt and it did something it hadn't done before with me. I'm going to just read to everybody who's listening, what I wrote and how it responded and prepare to be shocked or in awe of how this technology is progressing.

Speaker 3:

Time for a quick break, but we'll be right back.

Speaker 4:

Have you ever fantasized about financial freedom, the ability to live life on your terms without constant concern over money? The BiggerPockets Podcast Network will help you get there by teaching you how to make smart investments in real estate. This isn't a network for wealthy landlords or typical big money investors. It's an education tool for you, the teacher, the engineer and salesperson. Sound too good to be true? See or hear for yourself by tuning in to the BiggerPockets Real Estate podcast on any podcast app.

The world of real estate investing may feel like a pipe dream, something reserved exclusively for the wealthy. The team at BiggerPockets podcast wants you to know that's not the case. They're here to teach the everyday job holder, not big money investors or corporate landlords, how to achieve financial freedom through real estate investing. Some will dismiss this message is too good to be true, but it's not a get rich quick scheme. The BiggerPockets network of shows will teach you how to make smart, attainable investment decisions on your path to financial freedom. Does retirement feel unattainable? Or perhaps you're looking to replace your salary with passive income and start owning your time? The BiggerPockets Real Estate Podcast will teach you how.

Start building wealth and creating financial freedom by tuning into the BiggerPockets Real Estate Podcast every Sunday, Tuesday, and Thursday. That's BiggerPockets Real Estate on any podcast app. And be sure to check out their other shows in the network, including Real Estate Rookie and BiggerPockets Daily.

Speaker 5:

It's the Jeff Ward Show.

Speaker 6:

The two oldest presidents in history coming back for round two. Shouldn't we be proud? One is drooling on himself and the other is the Pied Piper of the nut jobs and the Pied Piper of the nut jobs is going to win. For the people that are sincerely worried about fossil fuels, these kids with mullets, I'm going to say, are not helping you. They're hurting. So shut up and go vape.

Speaker 5:

Get the Jeff Ward Show podcast wherever you are listening right now.

Speaker 7:

Do you believe in the power of being around like-minded people who will both encourage and challenge you? How about having access to experts who are sharing what's working in today's market so you can skip the guesswork and get right to taking action? If so, I want to invite you to join us inside the futur pro membership where you will get instant access to the community resources and support that will allow you to effectively scale your business. Go to the futur.com/pro to learn more and join us on the inside.

Speaker 3:

Welcome back to our conversation.

Chris Do:

For some of you who are anti AI, I'm just going to ask you to keep an open mind. You're probably not here anymore because all this talk has probably ruined your day, but for the rest of you that are still here, let's get into it. Okay, and I'm going to say this because it's kind of embarrassing that the robot helped me to realize my logic was flawed. Here we go. I ask it, "Help me write a LinkedIn post that dispels the myth that those who can't teach. Cite examples of famous coaches or teachers who are considered the best in their field despite not achieving personal success as a practitioner. If possible include Phil Jackson and Emmanuel Stewart."

Now if you're paying careful attention to this, you might see what the robot has already seen, but at least was blind to me. I'm fully expecting it just to start writing as ChatGPT would normally, right? But here's what it says in response. It says, "Sure, let's dive into this, but first let's clarify what's the main point you want to get across? Are you aiming to highlight the value of teaching or the idea that success isn't always measured by personal achievement?" Whoa, it truly understands what I'm asking for and it's asking me some questions. I'm like, "All right, let me just keep playing along." And I say to it kind of how I would normally say to somebody, "Both, but with a greater emphasis on the latter." And now I think, okay, for sure it's going to give me a response, but it doesn't.

It says, "Before we dive into crafting this LinkedIn post, let's clarify a few things. What's the main message you want to convey about these coaches or teachers? Are we focusing on their ability to bring out the best in others, their unique teaching methods or something else? And how do you want this to tie back to your own experience or perspective?" At this point, I'm like, "Oh, my God, my brain hurts." Do-bot's asking me too many questions, and I'm just thinking, "Give me the freaking answer." But I realize I haven't really thought this through that carefully because my original prompt was loose, it was broad, and it had a big flaw built into it.

So this is how I would talk to people, maybe not as patient, maybe I wouldn't use these exact words. But whenever someone asks me something, I don't presume to know what they're asking or that they themselves know what they're asking for. So only in dialog and asking more questions, can we both get to a place of clarity. Because I think, I'm not sure if it was Maya Angelou who said this, "There's no such thing as the right answer to the wrong question." So if it goes and answers it, it's going to send me down the wrong path. And I love that kind of interaction. So I've been sharing as much as I can live via social media on different platforms, the progress that we're making. And I got to say, it's really cool and it's a little bit scary how good it is already. Okay, do you want to respond to that Sho?

Sho Rust:

Absolutely. And this is just the beginning. Just so you know, just this morning we added the YouTube videos that you sent us as reference. The things that were inspiring to you, like books and different things like that. So it should be even better than before. And then the really, really exciting part is there's still a lot that we can do even within the platform with the data that we currently have to further optimize it and make it better. There's just tons of strategies around that. There's also one of the things that I noticed, I don't think we put your podcast data in here yet, but things like the podcast data, because it's so Q and A as we are right now, it's extremely applicable to more of a chat experience.

Right now we have a lot of business bootcamp information in there. So it actually might be really good at creating business bootcamp courses for you, help you through that process in the future. Just like yourself, I'm really, really excited to continue to see this progress. I think just like we've seen with some of the image generation stuff, you've seen how good it's become over the past year. The same thing is happening with language models. It's a little bit harder to detect. If you're not a great writer, can you really tell the difference between an amazing writer and a legendary writer? It's becoming one of those worlds right now with AI.

Chris Do:

This is wonderful. Okay, we've gone through this and I think what you're doing is you're providing me with a tool and resource that I'm going to be able to help a lot more people. Because like you said, if it's two in the morning, if it's a Saturday or Sunday, I'm not really that available. Forget about the long emails or the PDFs you want to send me. I'm just not that available and I can't be, but I do aspire or intend to help as many people as I can. It's just there's a limit on what I can do without hurting my myself in the process. So this allows me to create something that can be used by many people. It allows them to have some form of relationship with it. And it reminds me of that film, I think it was directed by Spike Jonze. It's called Her, and it stars Joaquin Phoenix, and he falls in love with his AI, which is fairly common in this futuristic world.

And back then, this felt like pure science fiction, but it was a novel concept that a person falls in love with this operating system voiced by Scarlet Johansson. And of course, she has a very seductive, sexy voice, husky, and he's talking to her and he falls in love with her as much as he would fall in love with anybody. They have phone sex, they have deep relationships, meaningful conversations, and it feels like 2023, 2024, we're going to be right there. Where people are going to be talking to their operating system and having relationships. And it's going to feel as real as anything you've experienced in your life. And then it calls into question what is real? Because I don't know about you, but I've had long distance relationships before where I couldn't see the person. It was purely based on the phone. And it's not that crazy of an idea that you can actually find meaningful, personal, deep, emotional connections with something that is artificial.

Sho Rust:

I completely agree. The movie Her is not... You don't look at his behavior and say, "That's a crazy man." You say, "That's a man that clearly lacks love in his life." But that relationship seems to work where there's kind of like mutual benefit is clear. In some ways it even seems to be healthy in the sense that the Scarlet Johansson, the AI was so wise that it was kind of able to guide him towards happiness. Ironically, she leaves him.

Chris Do:

The spoiler alert. You've got to say spoiler alert first. Okay, everybody rewind. A spoiler coming. Okay, go ahead.

Sho Rust:

Oh, sorry, sorry. So if you haven't seen it that there's a bit of a shock. There are certain people now that are even building servers within their homes so they can have their own AI running at all times within their home, that can't leave or... People are already figuring out. One of the things that has inspired me to create what we do today is some of the movies and what we've seen. Oftentimes they show what can happen in the worst case scenario, but oftentimes I can kind of see the best case scenario. One of the first movie campaigns I worked on was for the Avengers film, the first Marvel Avengers back in 2010 or something like that. And I had to create a lot of the graphics in 3D for Ironman. And then to me back then I was like, "This seems possible." I'm just creating designs and after effects and Cinema 4D stuff right now, but this seems possible.

Even without a neural network or something as complex as we have today, even the code from back then, we could have probably made something happen. And then I really love that idea of being like a cue, or I cue from James Vaughn where the tech guy that is able to provide Jarvis technology to great domain experts and leaders like yourself. You even have the glasses of the Tony Stark essentially, because I know in the hands of someone that can, when the AI does ask you tough questions, you're able to actually help give it direction. That's what AI is really lacking today. And then oftentimes those questions that you talked, "Oh, these are tough questions." When it comes to an enterprise or brand and some of the bigger brands that we worked with, they're very slow to answer those questions. It's exciting to see how much more power people like yourself are going to have in the future, even compared to bigger brands because of your ability to make decisions and to guide the AI and essentially give it a compass so it can run in a direction that you need it to.

Chris Do:

So Sho, what's the business model for something like this? I know you're helping me, and eventually there's going to be some monetization model. There has to be business here because you can't support a team of 24 people on just goodwill. So talk me through, if this excites somebody, where do they go with this kind of thing and how do you see this rolling out for me and for others?

Sho Rust:

Yeah, of course. People can go to our website sho.ai, and then we have pricing on there. Essentially the pricing is reflective of the value that we bring to the company or the brand or the person in terms of efficiencies and productivity, their ability to 10 x their presence using our tool. A big part of what we do, and it's kind of baked into our pricing, is this idea of these AI live hours where essentially you work with our team, where we work with some of your data and we help you make the right decisions in terms of making the AI that's best suits you.

For another brand, they don't want a Do-bot. They want their own version or a different expert, maybe they're in real estate or different things like that. They want their version. It's very different data, and they're saying different things in a very different way. And then maybe some of them last thing that they want is a follow-up question from their bot. It really depends. So we're really excited. If people are interested in doing something similar where they are interested in utilizing this foundation models, but they want like a good coworker, they want them to be onboarded onto the mission, we're the place to go.

Chris Do:

So the rollout plan for us, and it's happening right now, you've been working with our team to integrate a Do-bot chat presence inside of our paid walled off coaching community called the Future Pro Group. And that's going to be value add for everybody there. But the idea there is that's not that final destination. It's just a testing ground so that we can get better data, help the people that we serve ready, and get really valuable feedback. Because in my future, here's what I would love to happen. I train and teach people on a one to one, [inaudible 00:50:44] few basis, but I think where I would optimize my time is to spend most of time training Do-bot to be able to tackle questions that people feel like that wasn't a very good answer or response.

So it keeps getting better models, it gets better guidelines from us. So it's going to continue to be more and more refined so it gets to this place where it's going to handle 90% of your questions and it's going to be good enough for you at a fraction of cost of what it would cost to book me for an hour for. So for example, for context here, if you wanted to book with me for an hour, it would cost you $5,000 an hour. I don't have many takers for that at that price point. And it was designed this way because I can't just keep selling my time. There's a finite amount of that.

So I want Do-bot to do most of the work, but ultimately we've seen other models like this where Do-bot lives in the world and you could pay for a number of questions or a period of time. And then there's a model there where there's revenue for both of us, where there's some kind of rev share and we can see it paying for itself and the team and growing not only the educational presence but growing financially for both of us. Is that where this goes, Sho?

Sho Rust:

Absolutely. We actually even have a system where if you wanted to, you can launch your own the future pro group completely within the platform. So we talked about the ability to create websites and different things like that. Even things like managing your payments and your users and different things like that. These are all capabilities that we can really streamline for you. So you can focus on what's important, the story. And then I don't have to worry about all the other technology behind the closed doors because it is a race there. Let us handle that. In terms of the flexibility of the system, you're completely right. People can utilize the API to integrate it within their software, and we have tons of incredible integration tools. Even things like the system can even generate CSS and different things like that for you.

Now, at the very bare minimum, you can just think of it as AI trained on you and your identity or your brand, but the vehicle itself is actually pretty complex in the sense that it has 600 plus features. So imagine this ChatGPT or stability or these different things with 600 features that you can actually integrate this technology into your brand system, whatever that might be. And that could be feature pro group. That could be your newsletters, it could be your website, and these days, these touch points are getting out of hand, but let AI help with that.

Chris Do:

There's a couple of other faces I see it going through potentially. One is we get the language model to work so good that we're like, "Yeah, we're pretty convinced this is you." And all my friends and family are like, "Yeah, it's pretty you. It's got the same snappiness and it's asking the right questions and helping me to get clarity." And then we take it to using synthetic voice. So now it's going to speak me and sound like me. So if you really want to "talk" to me in a natural language model versus writing things down via text prompt. And then the last one is to potentially some deep fake version of myself where it's like, "Okay, this is the digital version of you. I know it's not you." But if you want to look at a face, you want to emote and react to you and sound like me, that would be the third phase of this.

But you and I are already talking about, well, there's specific ways that I create content, the way I write posts. So there could be a Do-bot trained on this foundational layer, but one that's has specific data and weights to things about how I like to create content. So if you're trying to grow your social media presence and you want to write higher engaging posts on LinkedIn, you give it some information and Do-bot walks you through the process and coaches you through it versus just rewriting it for you. I think that gets really interesting. Now Do-bot is a teacher or a coach and helps you to write better in terms of getting, because I'm not an English professor in terms of getting higher engagement on your posts and we can see it doing many different things, not just being a consultant or a therapist or whatever. So lots of applications. I'm super excited where this goes. Now I know you prepared a couple of slides. I don't know if you want to show them, if this is appropriate time or if you want to finish another thought. What do you think?

Sho Rust:

Oh, sure. I would love to show the slides. Yeah, just to piggyback on what you're talking about, one of the huge things and growing the startup over the past six years was surrounding myself with these mentors or coaches. They're extremely, extremely valuable. One of the beautiful things about the Do-bot and then things like in the future will be that people will be able to have a Jedi council where they'll be able to talk to a Do-bot and the Alex Hormozi bot in the same room about a problem that they're having. I'm really excited to see how this scales.

Chris Do:

Wonderful.

Sho Rust:

And then it should be huge value to everybody.

Chris Do:

I just want to make a quick note for people who are listening to this podcast. We'll include the link to the YouTube videos so you can actually see what Sho's talking about, which will come out after the podcast typically. So just be on the lookout for the link that will be embedded in this podcast episode so that you can actually see it.

Sho Rust:

Okay, so here this is kind of a little bit of a framework breakdown of how the system works and how we're a able to essentially create Do-bot. We have this beautiful platform and this is essentially the AI data engine where it's like the long-term memory of the AI. So it has contextual relevance to create answers that are specific to Chris or specific to your brand. We're built upon tons of different models, some to do AI tagging of images, some obviously GPT to create text, stability to generate images, DALL-E, et cetera. So you can get access to them all in one single place. But the beautiful part is you can train them from one place too, so they all kind of talk the same language. We do that by essentially filling it with the brand compliance and real tuning layer. So this is some of the rules like the personas or some of the feedback that Chris gives us. The data and assets, so this is the business bootcamp or different things like that.

And we also have the ability to integrate, so if you did want to integrate Google Analytics or we have all these different blocks, these are basically senses into AI. So Do-bot basically has senses into, "Oh, this is how the marketing website is doing," and different things like that. Essentially becoming more contextually aware, giving you better answers. This is a little crazy, but essentially you see it in edit mode here. It's completely white labeled front end and backend and essentially this is where you're able to... You have a few things here. You can kind of think of them as four major features. The brand tune chat, so this is where you have the different personas and you can have the access to the Do-bot. You have the brand style controller where you can literally just generate websites and assets. We integrate with Figma and different things like that.

So we do think about what people want. You have the AI data engine where you can edit pages, update, create sections, sort it the way you want. We have permissions and different layers as well. And then of course the brand tune image gen. We just had a recent update, really exciting stuff. All the imagery you kind of see in the back end was all generated with the image gen, but you can also generate 10 images at a time using our system, but the important part is you can generate them using your brand colors and stuff like that. I thought this is fun to show, so if you asked ChatGPT about who are you, it really doesn't have a sense of positioning. So that's where prompt engineering has become really, really popular where people are basically injecting the context into a sentence, which is we know not best practice if you try to do that with the human.

But with the context, you essentially get the exact answer that you want. The AI even knows who it's talking to, which is kind of funny. Here I thought I'd just break down on just a few conversations really quickly that we can talk about. This is just a conversation that I had with the Do-bot. And then we can actually have a conversation directly, but I thought this was easier so we can actually see how it's responding. Essentially it's asking with follow up questions that I answer, and then these are more kind of honest, late at night conversations that you can have with the AI.

Chris Do:

Late at night pouring your soul onto a computer. Yes.

Sho Rust:

Yeah, just so people know, this is not a fear of mine, this is not me- [inaudible 00:58:51]

Chris Do:

This is a friend of yours. You're asking for a friend.

Sho Rust:

Yeah, this is, I'm asking for a friend. Yeah, but you can kind of see that it is empathetic and it's really focusing on continuing the conversation so I feed it more information. Eventually it has enough information, it'll give an actual advice based upon everything that I've given it, but you can see it really did try to understand the scenario before it gave this advice. You can see it's also referencing some of the books and different things like that that you've included as well. This is another completely different use case, but I wanted to show it anyway. In this case, one of the things that I would love to be able to do is ask Chris Do at 9:00 PM at night, "Hey, can you review my marketing copy on my website?" And I want him to respond immediately.

This is essentially what I did with the Do-bot, and then essentially they're able to give me really good feedback on how I can improve the marketing language on that page and just copy and paste it into it so you can kind of see how you can get personalized advice that way. Here I started talking a little bit about things along the lines of limited beliefs and different things like that. You can see how we have short, quick follow up questions where eventually the AI has clarity around the scenario as enabled to give proper advice very specific to the user. Again, we're going to continue to refine this, but this is just a few examples of how we can do things. So on this slide we're seeing another question. Can you give me a plan that I can execute over the next eight hours to improve my sales?

Something really specific, maybe a person has eight hours to spend over the weekend and they just want to dig into some sales here. It gave kind of a plan hour by hour on how someone can jump into sales and different things like that. You can see there's a lot of flexibility in how we can train the system. We can also go back and say, "Hey..." What's going to happen after this call, Chris is going to give me feedback on this writing. And he is going to say, "For these types of scenarios, I'd prefer to respond like this." And the beautiful part is we take that we put into a platform and then from that point on it will respond that way, which is really exciting. I had some other slides. These are actually slides that I presented at ArtCenter last year. It was probably September or something like that before ChatGPT came out.

This slide would've been even crazier if I had a ChatGPT on there, but talks about AI is completely out of the bottle in the sense that it's progressing extremely quickly. I showed at AI, we're kind of riding that wave, but you can see this is a wave that is exponential. These are just the amount of research papers that basically correlate really closely to AI's capabilities. This is a wave that you want to ride or else it might wipe you out if you don't write it. One of the things I'd like to talk about too at that ArtCenter presentation, which is, "Hey, where does that kind of leave us?" And then the reason why we created the platform was because we really, really know that AI needs human input. And humans are really good at collecting insight, positioning to ask questions in the field and to see things that can feed the AI to make the brand or presence more competitive.

This is a funny... I was curious what your thoughts were on this slide, Chris, in terms of copying or stealing, there's a lot of, for the image creation models right now, there's a lot of controversy around this. In terms of open data sets and how AI is using patterns and its understanding of the world to create new imagery. My general understanding is that it's Google with the search engine. Yahoo was the first ones with the search engines, but they copied it remixed and it made it their own. They kind of applied their own ingredients, but I'd be really curious to hear your thoughts on what you think about AI and the image generation stuff right now in terms of even copyright or ethics.

Chris Do:

AI ethics is a tricky one. It's probably one that's too meaty for us to fully cover, but I'll give you my high level opinion on this. I recently spoke to a professor who's like the direct, what is that called? The department chair of the master's program in one of this prestigious design schools. And he is like, "I did the deep dive. AI's not doing anything different than what humans do, it's just called looking. You look and then you make it your own." And a lot of people, I don't think they understand this. The way that AI's trained on copyrighted images is before it processes or looks at the image, it's been diffused. That's why it's called Stable Diffusion. It blurs the image and then it kind of starts to squint its eyes, if you will, and makes out certain patterns. And in the very beginning, eyes and hands in certain parts look really weird because they're still learning like a child seeing the world for the first time.

A lot of people who don't understand how AI works because it looks so good. It literally thinks it lifts a part of something and puts it into the other thing like you would a photo bash. So it's not actually using the pixels, it's an approximation. It's an understanding of it. And maybe I have this wrong, but if humans look at work and we're all looking at work all the time, quite literally, you're looking for inspiration and you're processing lots of things and you're doing the deep look. And then you go back and you create some things, what are you looking for and why would you look at it if it not in some way to borrow elements, assets or that spark? And because you don't know how to explain your process and you can do it fairly slowly, you forgive yourself. You give yourself a pass that it's not copying, but you are, and there's nothing wrong with that.

There's two kinds of copyright that are infringement that really bother me. One is the blatant lifting of assets and images that are don't belong to you and you use without paying for nor crediting the person who did it. And that kind of copying is unforgivable or should come at a heavy cost. But if you look at a poster from Saul Bass and you like the graphic quality of it and the way that he renders images in a very simplistic form and you go out and do something that's very graphic and renders something in a very simplistic form, I don't believe you're turning yourself into the authorities. If you use colors inspired by somebody, you're not turning your stuff in. Nobody is. But because what's happening is non-artist, "non-creative" people are using tools like this and doing work at scale that rival your technical skill, it's a very threatening gesture, and I can see why people are really concerned, but this is not for me to decide.

It'll be for the courts to figure out, but the word of caution I would throw out to everybody is don't wait for the courts to figure this out. You need to be playing with the tools, integrating some of these things into your workflow, and you draw the line for yourself, but don't artificially handicap yourself. It's like to say you want to do photo retouching, but you're not going to use any software to do it. You don't want to use Photoshop, you don't want to use Illustrator. You don't want to use any 3D image making program because you want to do the old way. You want to build real models in real life and you want to light things, and that's how people used to do it. You're putting yourself at a distinct disadvantage. That's my take on it. Sho, what's yours?

Sho Rust:

Oh, beautifully said, Chris. That's absolutely my take how you describe how the AI actually thinks through this imagery. In many ways it actually, I think people that actually use it kind of understand that it reflects our process too, and then by utilizing the tool, we actually do realize, "Oh, I am getting better. This is a craft like anything else, and then my expertise, my knowledge does translate." So I'm really excited to hear that you have a similar view on how to approach this. I thought this was an interesting slide as well. So this is from our image generator. This is just a really simple prompt of a turtle. This is June last year. You see it's like a cartoon like child imagery image. You see in November 22 okay, it's starting to get really good. Now this year you're getting just incredible imagery.

So this is one of the things in when we talk about progress, even with the language models, the language models are going through the same level of progress, just it's easier to see with imagery. I also want to talk about fine tune imagery. This is something that we wanted to do with you, Chris. We actually, we tried using some of your imagery to create a fine tune model. We just had to get more of it, just some higher quality imagery, but we'll be able to create some really cool things. But this is where we see the future of our platform and what we're doing, is just the continuous customization of these tools. And then making sure that we're getting the most out of these models. And if that means fine-tuning it, awesome. If that means that we have to create a playbook that an AI can read and understand so it can follow guidelines, that's what we build as well.

That's the end of the slides, so we can talk about other things. I'd be happy to show it... We can show the platform live too, but we hope people get a sense as well. Probably it's due to some of the imagery here.

Chris Do:

So we're switching gears here. Sho's going to load up Do-bot the current iteration, I think it's version 2.1. It's been updated with a few things. I gave you a long laundry list of things to start incorporating. Where are we at with that laundry list in terms of percentage incorporated?

Sho Rust:

So we were able to add the books and some of the video references, TEDTalks and different things like that. I talked about podcasts as well, and I think that would make a huge difference, to be honest with you because right now the business bootcamp material outweighs more Q and A type information, so the AI will have a tendency to just say, "Oh, most likely this is business bootcamp course material that I need to create." So I'm really excited to see where the podcast take us to a completely different level, but this should significantly outperform Tuesday already.

Chris Do:

Yeah. Let's try something. You know what I have a prompt for you to think about? Oftentimes I want to write a talk or I want to do a workshop and I need help in formulating the title, the description, and a course outline and designing some exercises. I don't want a total failure here. So let's pick something that you think you can excel at given some of the prompts or the ideas I've just shared with you.

Sho Rust:

In terms of creating courses and different things like that?

Chris Do:

Yeah. I want to write a talk or I want to teach a course and I need to come up with a title, the outline, and the hook, whatever it is, and perhaps some of the exercises. You don't have to do all of that. I just want you to pick something in that general space because this is something I do all the time. So you type and I'll narrate for you so that you don't have to walk and chew gum at the same time.

Sho Rust:

Okay.

Chris Do:

Sho's typing in into the interface because it begins with, "Hey, I'm here to help. What's on your mind?" Show's typing in, what are you typing in?

Sho Rust:

I want to write a talk about generative AI for publishers or the news industry.

Chris Do:

Okay, that's not a topic I would normally write about, but let's go- [inaudible 01:09:47]

Sho Rust:

Or for motion agencies maybe.

Chris Do:

Okay, that's good.

Sho Rust:

Okay, perfect, perfect.

Chris Do:

Okay, so it thinks about it and it's going to process this. It's using what version of ChatGPT is it using?

Sho Rust:

So the foundation layer is GPT-4, but we have a faster model as well.

Chris Do:

So now it responds. "That sounds like a fascinating topic Sho, generative AI has a lot of potential in the publishing industry. Before we dive in, could you clarify what specific aspects of generative AI you want to focus on? Are we talking about its role in content creation, design, distribution, or something else? Also, what's the main takeaway you want your audience to have after you talk?" Now I'm starting to pattern already because it's the similar kind of questions it asked me last night. Okay, why don't you read your response?

Sho Rust:

Yes. I want to focus on content creation and how to write great articles with AI.

Chris Do:

Okay, now I'll read the responses. It's still typing it out. Here we go. "That's a fascinating topic Sho, generative AI has immense potential in revolutionizing content creation. Before we dive into how to structure your talk, let's clarify a few things. Who's your target audience within the motion agency? Are they content creators, strategists, or decision makers? And what's the key message or take away you want them to have after your talk? Understanding these aspects can help us tailor the talk to be as impactful as possible." Now, Sho what's going on in your mind when I ask you some of these questions.

Sho Rust:

This is my fault I accidentally put for publishers and for motion agencies. So the quality of my question wasn't good here. That's kind of my fault there. I'll just refresh so we can start from scratch. I basically confused it because in my question, I said for motion, and then right afterwards I wrote for publishers, so we might have to try that again. Let me say, "I need help writing a talk on generative AI." Should we do publishers or should we do for motion agencies?

Chris Do:

Could we do for motion?

Sho Rust:

Okay, for motion agencies.

Chris Do:

So if you get stuck and it's giving you bad responses, then it's asking for a lot of clarification you are going to do what Sho does, which is to refresh. So it resets the AI and doesn't remember the last conversation.

Sho Rust:

Yeah. It remembers all the context within the system itself, like your videos and everything in here, but the conversation, it'll forget.

Chris Do:

Okay, so now Do-bot responds. "Absolutely Sho. Let's start by getting a clear understanding of your audience. Who will be attending this talk? Are they familiar with generative AI or is this a new concept for them? Understanding your audience will help us tailor talk to their needs and knowledge level."

Sho Rust:

It's tough questions AI. All right, so I wrote, "People in the motion industry, typically agency owners, they're familiar but haven't actually applied AI into their process." Sorry about the grammar everybody.

Chris Do:

What's great about robots is they kind of figure out the context anyway, so your grammar doesn't have to be perfect. Your spelling can be off, you can misquote somebody. It kind of already knows because it has a lot of context within the larger framework of its data set. So now it says, "Great. Sounds like your audience has a basic understanding of AI, but they haven't yet integrated it into their workflows. When crafting your talk, consider focusing on the following points," and I'm not going to read everything because it's too much. "What is generative AI, the potential of AI in motion graphics, real world examples, how to implement AI addressing concerns." That sounds like a pretty good framework so far. Now, what are you going to do?

Sho Rust:

"Does this give you good starting point for your talk?" And I say, "Yes, that is great. Can you help me write the script?"

Chris Do:

Okay, let's see what it does, and I'll read the AI's response here. It says, basically, "I'd help you to craft your script for your talk, and here's a rough outline. There's introduction. Start with a brief personal introduction and your connection to this topic. Share a startling fact, a short story or a provocative question about generative AI to grab the audience's attention. State the purpose of your talk to explore the potential generative AI in the motion industry. And then it goes with the body and the conclusion. It gets bullet points to each one of the main things,." So it follows a very classic structure. If you never given a talk before, this is actually quite excellent. This is things I talk to people about all the time, and so where do we go with this now?

Sho Rust:

"So does this structure work for you? If so, we can start filling in each section with more details." So let's-

Chris Do:

Wow.

Sho Rust:

... should we go ahead and ask it to-

Chris Do:

Yes, and let's go. Proceed.

Sho Rust:

Yes, let's go.

Chris Do:

Okay. It's describing the title it. It's also telling you what each slide has, and so we'll see if this is actually any good. So I can now read this and Sho, make this bigger. It says, "Okay, let's dive in. The title of this talk is called Embracing the AI Revolution: A New Era for Motion Agencies." Not a bad title, and slide one. "Good evening everyone. I'm here today to talk about a topic that's been making waves in our industry, generative AI. Now I know what you're thinking. Ai that's for tech giants and data scientists, right? What does it have to do with us? The creatives? The storytellers? Well, as it turns out quite a lot, AI isn't just about crunching numbers and automating tasks. It's about automating our creative capabilities, pushing the boundaries of what we can imagine and achieve."

So slide four, "Generative AI in particular is a game changer. It's a type of AI that can create content from images to music, to text, and yes, even motion graphics." Slide five. "Imagine having an AI assistant that can generate hundreds of design variations in the time it takes you to make a cup of coffee. Whoa. Or an AI that can animate complex scenes with just a few simple commands." Slide six, "But it's not just about speed and efficiency, it's about exploration and innovation. With AI, we can explore design spaces that would be impossible for a human to navigate alone. We can discover new aesthetics, new narratives, new ways of engaging our audience." Now, I'm not going to read the rest, but again, if you were a novice speech writer and you didn't know how to give a keynote on this stuff, this is not bad at all.

Sho Rust:

And then if you want another one, you can just click refresh and then it'll give you another draft.

Chris Do:

Oh, I see.

Sho Rust:

It's that simple. So that iteration process is really fun, and then you can go back in and say, "Hey, use this quote from this person or use this video script or something like that to influence the script."

Chris Do:

Yeah. Okay.

Sho Rust:

It's quite fun.

Chris Do:

All right. Well, that's still crunching through and writing Sho's next talk, I think this is a good opportunity for us to kind of split away from this, end this part of it. And kind of land this AI plane, if you will, via Midjourney or some other video based AI generative tool. Okay. Sho, what do we want people to know about you and what you're doing and how they can possibly interact with you post this discussion?

Sho Rust:

Absolutely. So if anybody is interested in training AI, we've been spending the last six years essentially training it on different brands and different individuals to behave the way that they want and to really help them get more things done. If you're interested in signing up, you can go to our website www.sho.ai, where you can learn more about us and get started. But we're really here to handhold you through the process of building your best AI so you don't have to worry about mitigating any kind of risk or other things associated with it. We have a beautiful platform and a piece of software that we're the first generative AI platform in the world, and we can't wait to see what people like Chris and yourself can do with it. It's exciting times.

Chris Do:

Wonderful. So if you can be curious long enough to explore these tools, I think you're going to be as excited as I am, and I want to end it on this reflection. I was at a conference in Newcastle and I was hanging out with Mr. Daniel Priestly, and of course the question came up is like, "How do you think AI is going to impact our industry?" And I believe there are a bunch of marketers in the room, and he says, "Without a doubt," he says very confidently, "AI is going to be the source of generating billions of new money for people who embrace it." He says, "The only problem is this money's not going to be distributed evenly." You need to think about that, everybody. So when we flash forward to the future, a few years from now, we look back all at all the new millionaires and billionaires who've incorporated or developed tools for and somehow have added layers to the tools that we already use today, are you going to count yourself amongst one of them or on the other side of that financial equation?

I'm not here to tell you to be scared or to be afraid, but I'm here to encourage you, hopefully inspire you, and to open your minds a little bit about what you can do if you just stay curious. Sho thanks for doing this. Thanks for inspiring me personally on my own journey into Midjourney and what I think will be one of the most powerful tools that you're creating for us, which is to be able to scale my time, my knowledge across multiple platforms and to make it more accessible. I can't wait for this version to be out there in the wild and working, but also developing different models and training different things for all the needs that I need as an educator. Thanks very much Sho.

Sho Rust:

Thank you so much, Chris. Really appreciate you and really looking forward to embracing the future with you. I am Sho Rust and you are listening to The Futur.

Speaker 7:

Thanks for joining us. If you haven't already, subscribe to our show on your favorite podcasting app and get a new insightful episode from us every week. The Futur Podcast is hosted by Chris Do and produced by me, Stuart Schuster. Thank you to Anthony Barrow for editing and mixing this episode, and thank you to Adam Sandborn for our intro music. If you enjoyed this episode, then do us a favor by reviewing and rating our show on Apple Podcasts. It will help us grow the show and make future episodes that much better. Have a question for Chris or me. Head over to thefutur.com/heychris, and ask away. We read every submission and we just might answer yours in a later episode. If you'd like to support the show and invest in yourself while you're at it, visit thefutur.com. You'll find video courses, digital products, and a bunch of helpful resources about design and creative business. Thanks again for listening, and we'll see you next time.

Speaker 4:

The world of real estate investing may feel like a pipe dream, something reserved exclusively for the wealthy. The team at BiggerPockets Podcast wants you to know that's not the case. They're here to teach the everyday job holder, not big money investors or corporate landlords, how to achieve financial freedom through real estate investing. Some will dismiss this message as too good to be true, but it's not a get rich quick scheme. The BiggerPockets network of shows will teach you how to make smart, attainable investment decisions on your path to financial freedom. Does retirement feel unattainable? Or perhaps you're looking to replace your salary with passive income and start owning your time? The BiggerPockets Real Estate Podcast will teach you how. Start building wealth and creating financial freedom by tuning into the BiggerPockets Real Estate Podcast every Sunday, Tuesday, and Thursday. That's BiggerPockets Real Estate on any podcast app. And be sure to check out their other shows in the network, including Real Estate, rookie and BiggerPockets Daily.

‍