s02e02 Rory Billing, First single employee billion dollar company?

00:00:05 Welcome to the Big Cheese Podcast. I am today's host. My name is Sean Heis. I'm here today with, 00:00:10 as usual, Brandon. Hello. Corbin, the Savant. We got Jacob, the CTO. 00:00:16 Idiots. And special. Idiot, Savant, for sure. And we also have a special guest today. 00:00:20 Rory Billing. And we are glad to have you on the show. 00:00:22 I'm excited. It's cool to be here. 00:00:26 So Rory is a, we will get into this. Rory is a bootstrapped founder. 00:00:34 We met, we were just talking about this, probably six months or a year ago, and just kind of been 00:00:41 on your journey. And have had some exciting times with that. And so we're excited to kind of hear 00:00:45 some of your perspective and get a little bit more information and hopefully provide some value to 00:00:51 all of our wonderful viewers. And there's, rumor has it that there's an idea that you have for 00:00:55 an AI startup that we can then- Yes. Yeah. Rory is going to be pitching an idea to us live. 00:00:59 we have not. We haven't heard any of them. We haven't heard any of them. No idea. 00:01:05 We're doing it live. We're doing it live. And you probably will. He will probably will. 00:01:09 No, but it's good. It's the first time I've hit your idea. I've gotten ripped to shreds. 00:01:16 So that kind of aligns with today's topic. So we spent some time thinking about a question that 00:01:20 we've had, which is, you know, with the advent of AI and all the abstraction layers on top of 00:01:26 technology today. Well, we see the first, you know, billion dollar unicorn come from a single 00:01:33 person. So Sam, Sam, I forget what interview he was doing where they talked about that. You're 00:01:38 going to see the first 10 billion or 10 person billion dollar company. And that then he predicts 00:01:46 you're going to see the single, a single individual who builds a billion dollar company leveraging AI. 00:01:50 So I then once I heard that, I'm like, all right, so so I went and I did the math. And that is what 00:01:58 was it? 27 million dollars a day that you have to make to do. Well, I would refute that. He said 00:02:09 valuations. So you just have to evaluate it. Brandon is a savat engineer. But when it comes to the 00:02:16 dollars and cents. That's not my school. No shit. So pretty much the way that it's turned out is 00:02:21 that Sean pretty much just says, listen, anybody ever wants to talk to you about money, just send 00:02:25 them to him, let him do the money. Because again, I'm an idiot. I'll do stuff for free just because 00:02:31 I enjoy the fucking show. Brandon's probably built multiple billion dollar companies in a weekend, 00:02:36 and they didn't even know it. This is like, someone else got the money. I'm bored with this. 00:02:41 I'm just going to put this on the shelf for 10 years. Don't do something else. It's horrible. 00:02:49 Yeah. So thinking about products and what's happening with AI, we got some research. 00:02:57 We've got a founder here. And we're going to talk about the concept of the AI product and 00:03:01 really what it takes coming from people that have built stuff without raising a lot of money. 00:03:06 Two, we've all built stuff without doing that. Or we've seen what's happening in the marketplace. 00:03:10 There's some really good people out there that you can follow or watch that have built tons of 00:03:17 products with limited team size, whether they were leveraging AI or not. So that's interesting. 00:03:20 But we're going to kick it off with some news because that's what we do every week. 00:03:25 And the first one, I think, is an interesting topic. It's my topic for this week, which is, 00:03:31 if you haven't seen this, it's pretty disturbing at some levels. But Microsoft, 00:03:38 they have their new product, their co-pilot plus, their co-pilot basically baking in AI into the 00:03:46 everyday computing experience on a PC. So you can do things inside of Microsoft applications, 00:03:52 inside the cooperating system. I mean, is this all of this happening within those new laptop or 00:03:57 the new Surface Pros that they announced, which has the new AI chip? Yeah. So part of this is, 00:04:00 it's the heart. It's the device play. They're trying to get that reverse cycle. 00:04:06 They need to compete with the Apple and the Apple Silicon chips, which have just destroyed 00:04:11 everybody. So now Microsoft announces that they've got these new Surface Pros that 00:04:17 the battery lives longer than the MacBook Air or MacBook M3 Air. Supposedly, the power is supposed 00:04:22 to get my base issue. It's still Windows. Like, I don't want to use Windows and that's whatever. 00:04:27 But I definitely think that the move they made there makes all the sense in the world and they 00:04:35 have to. Is NVIDIA involved, or not NVIDIA Intel? Is Intel involved at all with these new ones? 00:04:39 I don't know. It feels like they might be. But anyway, yeah. So this recall though. 00:04:44 So this is a kind of a crazy thing. I mean, you saw like two weeks ago, there was news like, 00:04:52 recalls the next major feature for AI. And basically what it's doing is, it's a feature that is 00:04:56 on your computer that's taking a screenshot of things that you're doing every three or four seconds. 00:05:03 And it's saving that to your computer. And first of all, you ever had like a cash like in your 00:05:08 code? And you're like, holy shit, my cash is for a gig. Think about it. You're going to wake up 00:05:13 day three of your new 100%. Your laptop and it's going to be, why do I have no space? Right. 00:05:20 I mean, so I'm not even sure what the use case of recalling everything that you've done in your 00:05:24 computer is. Well, I mean, we've talked a lot about like the winners are going to be the ones 00:05:29 who have a lot of context, right? So it's really going to be, I think, about keeping track of 00:05:34 context and having that like, hey, here's what you've done or do on a regular basis and being 00:05:39 able to process that data and then make predictions that are actually valuable to what are relevant 00:05:45 for you. Well, surely they're pruning them though. No, so the issue with this feature is that it's 00:05:53 just saved in plain text on your computer and SQLite database. Perfect. So some hacker got on 00:05:57 there and was like, holy shit, he opened up his computer and all the data is sitting there, 00:06:03 completely unencrypted. Now you can just, yeah, you can basically, if the thought is, it's so, 00:06:08 it's not necessarily persisting to the cloud. It's all, it's on your computer was backed up. 00:06:13 Right. It's on device. It's on your Microsoft has access to it. Sure. And that's my question. 00:06:19 And that's the variable, right? Because Microsoft is, again, they've got a huge advertising arm, 00:06:27 right? Like that's a big piece of what they've got. So that seems like that data might be the most 00:06:31 private data possible. Exactly. Stuff you're doing. I mean, back in the day, we said privacy 00:06:35 screens on our laptop. Do you really have one of those? Well, privacy browsing. Like, so privacy, 00:06:39 like when I go into privacy browsing, there's only one reason that I'm going into privacy browsing, 00:06:44 right? And it's going to be for the next seven, eight minutes, I'm going to be occupied. 00:06:48 And, and that's just capturing screenshots of all the weird stuff that I'm looking at. 00:06:52 Yeah. Or just like logging into your bank account, something a little bit more docile 00:07:00 brain than that. He's sick. Oh, my brain doesn't have anything in his bank. 00:07:06 We're working on that. I'm the asshole, right? Like, because I'm just being real. 00:07:11 So, I think it does like, I've talked to different enterprise. I was actually at a, 00:07:17 unless it's why I'm dressed like a schmuck today, but I was at enterprise, you know, 00:07:24 client today, which great client, but I'm doing some biz dev. And, you know, I ask him about AI. 00:07:31 And I'm like, and you know, you get a lot of, it's like blank stares, you know, and the, you know, 00:07:36 these companies are a little bit further behind. But like, they can't just put AI on computers 00:07:41 like that. They can't do that. And so like, they're, they're, they're obviously not adopting 00:07:46 that stuff. But for Microsoft, I guess to conclude, for Microsoft to put this out there and have no 00:07:51 controls or security in place, big miss, it's turned into a PR nightmare. Yeah. And you know, 00:07:55 obviously that's going to, that's going to hurt the reputation in the market from an AI perspective. 00:08:00 But so I will say though, that conceptually, it's kind of an interesting idea, right? 00:08:02 Yeah. Yeah. What is the use case? 00:08:07 So the use case would be that, hey, you know, what was that one website that I was looking at 00:08:11 yesterday that talked about this, right? And it would be like, oh, yeah, you did this. And it's 00:08:15 just, it's all there. So, you know, as each of those images are coming in, we're probably running 00:08:18 them through a vision model of some sort to extract kind of context. 00:08:21 Yeah. It's just it's using Azure AI to OCR the data. 00:08:27 Right. Yeah. It's why remarketing is so powerful, right? Because I went and looked at, you know, 00:08:30 deck stain today. And now I'm getting ads for a deck stain. And like, yeah, yeah, it's true. 00:08:34 I do like that. But the data that comes with that cookie in that profile, 00:08:39 that's a little bit more lightweight than a bunch of totally, yeah, your entire screen of 00:08:42 everything that's going on. I mean, it's not saving images. I guess it's just saving the OCR. 00:08:46 Yeah. But trouble understanding what you would do with all that crap. 00:08:48 And I think that's what you've got. Yeah. Yeah. 00:08:53 I was thinking, if I take a longer view of it, there could be some really interesting and unique 00:08:57 stuff that you could learn about yourself over a long period of time. So true enough, 00:09:01 it's collecting data for five or 10 years. And you look at even just from a professional perspective, 00:09:05 your career arc, you could do some really interesting analysis on the things that you've 00:09:09 learned and gained. And then you could almost build that into a mini resume of yourself to then 00:09:15 it could be like Spotify wrapped for you. Do you do you, sorry, sidetrack? Do you guys use DJ? 00:09:20 Yeah. Yeah. Yeah. Yeah. Does it continually just play the same songs over and over for you? 00:09:24 I tried one time and it was awful and I quit. I have like five songs that I can almost, 00:09:29 but I am boring and like to listen to the same five songs and tell. Yes, you do. And tell my ears 00:09:35 believe. But like, yeah, I mean, yeah. Totally. Yeah. So I've listened to Espresso by Sabrina 00:09:42 Carpenter 1000 times in the last month. So for me, like, I have, I have a crazy, 00:09:48 just insane depth of mute, not to, not to like to my own horn, but like, I like everything from 00:09:55 like cannibal corpse to, you know, Enya. And so like, I've got this wide range, but there is still 00:10:00 like five songs, hate, breed, as they play that thing over the typo negative, they play it all 00:10:05 the fucking time. They play a lot of the songs that you play a lot. But I'm wondering, like, is it 00:10:12 is it more of a, is it more of a cost saving measure? Like, was it by design? Because it seems 00:10:16 like everybody that I talked to has a very similar experience. And that they basically, 00:10:21 I mean, again, I've got like eight years of Spotify data and they play, they play the same 00:10:26 five or six different bands over and over. I have wondered about that with Netflix. Like, 00:10:32 do they keep the, they do the top 10 list because they want to keep those, those video files, 00:10:37 or those videos on the edge, stream those at a lower cost, right, to try to get so that they 00:10:41 don't have so much variance. Right. And that's the only reason I could explain why, like, 00:10:45 closer to getting to like a live TV. Yeah. Now I will say it's a very cool, like, so the voice 00:10:50 that they've got. So I actually watched a little video about the dude who's actually the DJ voice. 00:10:57 And they, they've cloned his voice. And so, but it, but that man, it's, it's on point. Like, 00:11:02 like that dude, like when you're listening to DJ and that dude talks, it is exactly the same. 00:11:06 So I'll give him props on that. Oh, also they, they probably just like, save it on your device 00:11:11 and play the ones that are only on your device. Oh, yeah. Yeah. Sure. My theory on some of the 00:11:14 bigger guys coming out with AI though, is that everybody knows they need a product. And so I 00:11:17 think they're rushing some of these things to market. Yeah. So that they get in the game and 00:11:20 they get in the news. And it's not there yet. Right. So I don't know if they're doing that on 00:11:24 purpose or it's honestly just not there yet, but they don't want to be six months behind everybody. 00:11:29 So boom, there it is. We got AI. I do listen to it. Like, like that's pretty much it. If I'm 00:11:34 going to launch, if I either I'm going to listen to a podcast, or I'm going to launch DJ from Spotify, 00:11:38 that's it. Those are the two things that I'm ultimately going to listen to. Also, like, 00:11:43 I thought it was going to be showing me music that was similar adjacent to stuff I already liked. 00:11:46 Right. Now I'm like, and it probably did at some point. I was probably like, Nope, 00:11:50 no, no, I don't like that. So basically, Spotify has recreated Pandora. 00:12:00 Yeah. Sometimes the AI features are less AI than the non AI features because like Spotify has 00:12:04 the chill mix and the reggae mix and all these different things. And like, that's saying what 00:12:08 genres you like. And those playlists are really good. Yeah, I agree. And then they build their 00:12:13 AI feature and it sucks. Right. Whereas like Apple's coming out with AI, but like you've always been 00:12:17 able to search for all your fish pictures on your phone, right? You've always been able to 00:12:20 search for your kid's name and all the pictures. You know what I mean? Yeah. Is that AI? That's 00:12:24 machine, you know, that's a really good feature. Right. And they've probably been working on it 00:12:28 for 10 years. Yeah. So those machine learning models kind of going back to when we were talking 00:12:34 with Zach, right? Those things are probably just very fine tuned, very just hyper focus. 00:12:37 And they're getting the word thing and they work and then you go and you have, you know, 00:12:43 general LLM and it's just like, Hey, let's listen to a song you listen to in 2022. And then it plays 00:12:47 the same fucking song that it played just like the last time. Like, what are you doing? 00:12:52 Anyway, sorry. Anyways, so we were talking about devices. We obviously have a huge refresh cycle 00:12:59 coming with with Microsoft and Apple. Everyone's WWDC. This week is next week. Is it this next week? 00:13:03 Yeah, it's next week. And then Nvidia has been Nvidia. Everyone was like talking shit about 00:13:07 Nvidia. Yeah. And now, and they blew up again, what's going on with Nvidia? 00:13:11 Yeah. So they just came out with their new next generation chip. They, and it's only been three 00:13:16 months since their last version of it. It's an AI chip. They're crushing the market. Like, 00:13:22 everybody is all the big players are just like dying to get their chips as fast as possible. 00:13:24 So who's buying their chips? Microsoft mostly. 00:13:29 All of the big guys, but they have, they don't share that information publicly, but 00:13:34 they have like someone who makes up like 15% of their revenue. And it's probably Microsoft, but 00:13:41 so they came out with a big chip arm is right there. And then it's in its intel, intel is the 00:13:47 lagging chip maker, but they have a couple of advantages. They produce their own. So they, 00:13:53 they manufacture, or they design and manufacture their own chips. And Biden just assigned that 00:14:00 $20 billion chips package. So they would be one to look for. Are they, are they, are they manufacturing 00:14:04 them here? I mean, I would say a good amount of it. I don't know the numbers. Because all like 00:14:10 everything that's coming out of Nvidia is going through TMC. Yeah. That's out of Taiwan. Yes. And 00:14:13 that's their competitive advantage right now, because they're just so much cheaper. Right. 00:14:18 Also, their designs are fantastic. And they're a three month cycle is insane for it. 00:14:22 Yeah, three months. Yeah. Well, and their new announcement was like, 00:14:26 Hey, it used to be a two year cycle, but you can expect a year cycle from now. And they attribute 00:14:31 some of that to actually designing with AI now. So they're able to like, you know, it's the same 00:14:35 thing with like, I don't know anything about it, but like drug design, where like you have 00:14:40 scientists go out and they try to figure out molecules manually and test them here and there. 00:14:45 Well, AI can like figure out all possibilities and simulate it to and then find the ones that 00:14:49 might be good, good like test cases. This same thing with, with the chip design is, 00:14:57 is use AI to actually design it. So, and then of course, Microsoft, Apple and, and Facebook, 00:15:01 they're all pretty trying to produce their own chips as well. So I think the coolest thing so far 00:15:07 out of AI has been like, because they're all racing to like be at the leader in this market, 00:15:13 the technologies that are coming out of it, whether or not Gen AI ever becomes like this 00:15:18 really, really powerful thing that everybody's using. The hardware that gets produced from it is 00:15:24 so sick. It's like my favorite deltron lyric, it's crisis precipitate change. But in this case, 00:15:30 it's opportunity. And so what happens, right? You, you invent shit that's never been done before 00:15:34 because you're chasing something. It's like the moon, like when we went to the moon. I was thinking 00:15:41 the same thing. Or, or the dot com bubble, right? So the dot com bubble, everybody's building, 00:15:46 but the, the major thing that came out of that is that we were laying pipe just left and fucking 00:15:53 right, right? And most engineers were not all, but, but at the end of it, when all those companies 00:15:59 were said gone, right, we still had fiber optic across the entire country. And now we have now the 00:16:03 infrastructure to be able to do everything we want to. So it's going to be a similar thing here, 00:16:05 I think, is that we're just going to have all of this hardware. There's going to be a lot of 00:16:10 people that are going to ultimately collapse and fucking fail. But you're going to have this 00:16:14 infrastructure now that's in place that's going to allow us like what, what happens in the next 20 00:16:19 years? That was where we're at. Yeah, that was my question on, on the chip specifically. And this 00:16:23 might show my ignorance of infrastructure. But I've read a lot about the amount of computing power 00:16:28 necessary, obviously, to support all the new stuff going on. As part of this with the chips, 00:16:32 that it actually reduces the amount of power necessary to create these models. And so when we 00:16:36 talk about long term alliance on the grid and all the instability there, I think that's what we're 00:16:41 talking about here is a lot of these advances might actually reduce our reliance on really faulty 00:16:45 infrastructure. That that was sort of the way that I was taking it. But I think that's really 00:16:49 intriguing because there's a lot of work that needs to be done there. And maybe somehow we're 00:16:53 going to get there sort of through a back door because of all the generative AI models that are 00:16:59 being created. So I very vividly remember. So we were working at a company called LMIV back in 00:17:06 20 to 2000s. And it was an MS startup that was MS Jefferson Pilot Chorus in Bonneville, 00:17:11 which is basically for huge radio conglomerates that we're all competing against Clear Channel. 00:17:14 And so they all came together and said, we're going to form a new company. 00:17:18 We're going to build a technology stack to power all of our websites, right? And everybody's going 00:17:23 to invest $1 per listener. And that was kind of where we started. And we ended up building this whole 00:17:29 free and ginormous infrastructure. But then the .com bubble crashes and that company collapses and 00:17:34 whatnot. But to this day, all of that technology that we built is still being leveraged by a lot 00:17:37 of these radio companies. So it's kind of that same, that same space. 00:17:42 Well, and I wanted to revisit Rory's point because Jacob and I were talking about earlier this week. 00:17:50 The amount of computing power and and power in general, it's being consumed by the AI stuff is 00:17:56 is more than the Bitcoin problem. Which is problem. So you're you're you're doing 00:18:01 another thing where you're grinding all this. You're grinding so hard, right? But 00:18:06 are you producing more value than you're extracting from right really the ecosystem? 00:18:11 In 2024 is the hottest year. I mean, on record again, well, it's on pace to be. 00:18:18 If you look at the day by day, it's like the graph doesn't look good. It's like mostly red, right? 00:18:22 Are you a climate? Are you a climate truth? Are you a climate truth? 00:18:29 I'm a believer in that. I'm a believer in we are able to measure the temperature on the globe. 00:18:32 And we can see that it keeps getting hotter. 00:18:33 Yeah, it's a momentous work. 00:18:36 The momentous work, Mercury has been here for a while. 00:18:38 It's a great state element. 00:18:45 I don't know. I so I have this one individual in my life, and I won't go into it. But we got 00:18:49 into this frigging heated debate to the point where my kids and my wife are like brand and just 00:18:55 walk away from this. But that yeah, the climate there's no problem with the climate. I'm like, 00:18:58 it's getting this is like five year four or five years ago. And I'm like, listen, we had the 00:19:03 record heat. And then next year, I guarantee while record heat and the year after that record 00:19:07 fucking heat, and sure enough, that's where we're at. And it's just like, I don't know what to tell 00:19:10 you. It's not like I'm a fan of this. It's not even 80 degrees. And it's June man. 00:19:17 Now, I will say that we've actually here in at least in Indiana, we've had a lot more rain. 00:19:22 We've had a lot more moisture happening within our environment than we have in that past. 00:19:31 So maybe it's a climate change. So you've been on a little product journey on your own 00:19:36 over the past week. And that was kind of what you wanted to bring to the table today. Talk to 00:19:41 us about what you're building. Yeah. So so my news today is less about a specific news thing. 00:19:46 So but what I did is I started building an app because I record so many of my meetings 00:19:50 that I was like, I'm just going to build an app that I could just drag one of those recordings 00:19:55 onto this app and have it automatically transcribe it as a business meeting. 00:19:58 And then it turned into, oh, I want to ask questions about it. I want to do all this stuff. 00:20:04 So I built this thing with a technology called Tori, T-A-U-R-I. And it's basically, 00:20:10 it's kind of rust in the back end. So if you're a nerd, rust, you know, it's like memory safe, 00:20:14 all this stuff. But then you're just building normal front end stuff. Anyway, it doesn't really 00:20:19 matter. But I ran into a bunch of problems with like dealing with the audio and all this. And I'm 00:20:22 like, you know, I think this might be better to try to build this with like a native 00:20:29 or a Mac OS application using Swift. I don't do Swift. I've never programmed in Swift. 00:20:35 I've never built a full application with Xcode. I've only ever used Xcode as a way of like 00:20:40 delivering my apps that I would build through, you know, different like cross platform things. 00:20:45 So I'm like, I'm going to do it. And I jump in and I set up a GPT 4.0. 00:20:49 God, I hate that name. And basically said, okay, here's what we're going to build, right? And then 00:20:54 I'm like, generate me the files every time. So anytime that there's a change to a file, 00:20:59 I want you to output the entire content of the file. And I just start copying and pasting, 00:21:02 copying and creating new files, copy and paste, copy and paste, copy and paste. 00:21:09 Sure enough, I get to a point where I have a full application that is that I can drag and drop a 00:21:15 media file on it automatically starts transcribing it. And then it's a matter of now, the next step 00:21:20 I have to do is I can actually take that subscription or that transcription and start doing AI on it, 00:21:26 which I've got the AI piece done. But nevertheless, I've done all of this without ever writing a 00:21:30 single line of code. I'm just copying and paste. So every time that there's some new thing, 00:21:35 I copy the entire thing, I paste it into the file and it's working. And so what happened 00:21:40 during this experience, though, is that the moment that I start writing a line of code, 00:21:46 now that the AI is out of sync with it, right? So it's it's now it's like, I've got to tell it, 00:21:50 hey, I've modified this. So here's my new code. So anytime you output it, make sure you've got it 00:21:55 in there. But what it made me realize is that this whole idea that AI is going to replace 00:21:59 developers is not going to happen, in my opinion. What's going to happen is you're going to have a 00:22:04 bunch of people who aren't developers building things that now developers historically built 00:22:08 they're going to release it and they're going to be like, look, I built this thing. And then it's 00:22:12 going to break. And then they're not going to know how to fix it. Because that was the one thing is 00:22:16 every time something would go wrong, I have to go in and I have to tell it, Hey, here's the errors 00:22:20 that happened. Most of the time it figured out, but sometimes it didn't. And I'd have to go look at 00:22:25 the code and go, Oh, it seems like you're binding this. Maybe you shouldn't be as like, Oh, you're 00:22:30 right. I shouldn't be. Yeah. And then it goes and it fixes it. So it's almost like what's going to 00:22:35 happen is that this AI is going to empower a bunch and non technical people to start building stuff. 00:22:39 Some of those things are going to become popular. And then you're going to need developers to come 00:22:46 in to actually do it, like to deal with all of the edge cases, which the AI just isn't good at. 00:22:51 But it does get you to a point where I mean, again, I can like every time I'd copy and I'd 00:22:56 paste it and I'd see that no errors happened that I could then compile that and run it. I was like, 00:23:02 Holy shit, this is amazing. And so that's, that just absolutely blew my mind at how well it worked. 00:23:06 But again, the farther down that path you get in building this thing, 00:23:10 then all of a sudden it starts to lose its mind, right? It starts to hallucinate. So, 00:23:15 because I went on like a six hour chat conversation with the context window for a chat 00:23:21 single chat with chat GPT to build something on a six hour conversation. And that's outputting 00:23:26 code is a lot of basically. Do you use co-pilot? I don't use co-pilot. Yeah. So my 00:23:31 next step, I guess. Well, so then, of course, what I did is as I was going through it, I'm like, 00:23:37 you know what? Actually, all I really need is all of the source files in a system prompt, 00:23:41 and then my question, and then just answer it. I don't need like, I don't need the entire chat 00:23:47 log. So then I took the transcription app, I cloned it, and basically, oh, no, 00:23:53 son of a bitch. I'm going to have to hold this, and I apologize, everybody who's watching this 00:24:00 right now. Oh, okay, there we go. So then, so I end up building a version of it where you can 00:24:06 just select a folder. It takes any text files, basically adds that to your system prompt, 00:24:10 and then you can ask a question, and then it just answers it, and it actually works really well. 00:24:17 Until I realized that my whole process was taking any file in that folder that's text and turning it 00:24:24 into so that your node module's folder, it ends up turning into like the most ridiculously huge 00:24:29 prompt you've ever seen. But nevertheless, once I kind of figured some of that stuff out, 00:24:35 it was like, oh, this, this, this seems like a huge opportunity for a lot of people to be able to 00:24:41 build some really cool shit that have no idea how to code at all. Yeah, and for me, I don't really 00:24:46 care about the, can you code? Can you not code? Are you going to lose your job? I think for me, 00:24:52 it's like, you learned a new programming language really quickly. Yeah, you didn't go through a 00:24:57 tutorial. Nope. And so you probably are incrementally learning this, since you have an engineering 00:25:02 background, you're more, you're more prepared to be able to do something like this. I think it's 00:25:10 truly, it's truly awesome to be able to, that that technology exists, because we look at, 00:25:17 look at the, the landscape of AI products, and think about what is them actually the most useful 00:25:24 thing that AI is doing right now, like in your opinion, like me? Yeah. I mean, I, I agree, 00:25:28 I think it's allowing you to, to do skills that you don't have more quickly and cheaply than you 00:25:33 would otherwise be able to do it as a, as a, you know, founder without a lot of resources, I can get 00:25:39 maybe half or 60% or 70% of the way to accomplish something. And it's good enough. Right. But where 00:25:43 I could never have done that, never have done that before. Yeah. You, you actually inspired me 00:25:49 with that story. I have an old program that I wrote like 10 years ago, and it's on an older 00:25:55 version of all this other stuff. And I wanted a Docker container that encapsulated all the older 00:25:59 versions of PHP and all the other bullshit. But I was like, I'm not that familiar with Docker, 00:26:06 but let's let chat GPT help me out. The last day I built this really, in looking at it now, 00:26:11 I know Docker a lot better today than I did two days ago. It's working. It's running. Yep. 00:26:15 And now I had this repeatable thing that I was not otherwise you're going to stack overflow, 00:26:18 you're going to the internet Google and it's like, okay, that's kind of what I'm asking, 00:26:23 but then I have to like make this mental model of how do I translate that answer to what I'm trying 00:26:28 to do. And you end up like in this hell, it's like, yeah, you're just sitting there like you're, 00:26:34 I don't know enough to use that. So, so that's a, that's another good point, which is most people 00:26:39 reach for, you know, the tools that are in their chest. But like in certain examples, 00:26:41 do you really need to build a web app or are you just building a web app because you know, 00:26:45 JavaScript, HTML and CSS. Right. Right. Historically, yeah, that's what we reach for, 00:26:49 because that's the tools we know. Yeah. But but in that in a certain case, 00:26:54 maybe swift is right. Maybe, maybe you should be building something in Python if you're doing, 00:26:59 if you're handling data or rust or whatever. So one of those things that came up during this 00:27:06 experience is that so in the, in the tari version of it, I used whisper, which is open AI's voice 00:27:13 to text trans model. And as I'm going through the chat GBT was like, well, you can actually, 00:27:22 now that you're in the swift world, you can actually just you can actually use apple speech 00:27:29 for the translation and not have to have the model or not, not don't need whisper or FFmpeg 00:27:33 to convert the files. We can use this stuff over here. And it does it all. Now the problem 00:27:39 that I ran into though is that the Apple voice was literally like 300% you just told you to hey, 00:27:45 you just hear in here's the code. And I'm going to, I'm going to build you a, you know, speech, 00:27:51 speech recognizer dot swift file. And here's all your code and just copy and I pasted it in there. 00:27:58 And it worked great. Yeah, but it was slow. So one piece of advice I'd give anyone who's 00:28:04 doesn't code at all is sit down, find like a 10 year old kid. Okay, sit down with them on 00:28:12 chat GBT, right? Who being weird? Sorry. My son's 10. I did this with him and his best friend. 00:28:18 We sat down and I said, we're going to build Mario. Okay. And we, and we just said we started with 00:28:22 chat GBT. And we said, we're going to build Mario. I want to build it as a web app. We're going to 00:28:28 build a Mario game. And we iterated over this code over about a 30 minute to an hour period. 00:28:34 And we got to the point where we had a, because I know web development, I knew how to take the date, 00:28:38 you know, take what it was doing and put it over in like a little project, right? But within about 00:28:45 an hour, we were jumping over obstacles. We were running into things that could hurt you. And you 00:28:53 were no longer able to move. And I learned more about animation than I had ever even tried to learn. 00:28:59 Yeah. Because it's like, I want to try. And that's how kids are now. That's the next generation 00:29:03 engineer. But I want to do this. Well, they don't give a shit how it's built. 00:29:07 The advice I always give people learning to code, and that's what I know, but it goes 00:29:13 any, anything like piano or whatever you're trying to learn, go find an example of something that 00:29:17 already exists, and then try to build that thing. You know, piano, it's like I'm learning 00:29:22 Jingle Bells right now as a song. I'm like getting pretty good at it. You learn a piano? 00:29:27 Yeah, I suck at it. But I'm using a, no, I'm just gonna, I'm not using chat. 00:29:32 I'll give you my, my eighth grade, or I was eight years old, my piano teacher, she's still alive. 00:29:37 She was phenomenal. Yeah. Yeah. And it's really like, find an example that you already kind of know. 00:29:43 And in chat, GPT allows you to do that and enables you to learn those things. And we use it for, 00:29:46 like, business and other things like that. I'm like trying to figure out, all right, 00:29:52 a sales strategy to go after this or whatever. It gives me at least the 60, 70, 80% solution, 00:29:57 where I can fill in the gaps and go learn more and helps guide me and get me faster, 00:30:01 which we've talked about that a lot. Yeah. And I mean, I think what we're looking at with AI 00:30:09 products is like about 10 to 15% of the market right now is probably just based off rough 00:30:15 research is related to coding assistance. So the big part of AI is, for some reason, 00:30:22 it seems that one of the best ways to utilize AI is learning how to code. 00:30:25 Yeah. Which is kind of funny. I think it makes sense though, 00:30:31 because one of the toughest things for me as a non-technical founder is to trust that devs are 00:30:35 doing what I'm asking them to do. Not that they're at doing it. I know they're trying to do it, 00:30:39 but that the quality of the code or the quality of the product is what I'm expecting. 00:30:41 Yeah. And then I'm not going to get two years down the road, and it's going to start breaking, 00:30:44 and I'm going to have to go refresh and rebuild. That's really, yeah. 00:30:50 So for me, it kind of gets to some of the things that I want to talk about when it comes to AI 00:30:54 and the ideal pitch you guys later. But yeah, it's very challenging for me to know. All I know 00:30:58 is you told me it's going to take 90 days and does it work or not in 90 days? I can't read your 00:31:02 code and have an idea of what you're writing, whether it's good or not. And I might get 90 days 00:31:06 and now I'm screwed, and now I've got to go find someone else, because I didn't know you weren't, 00:31:12 it wasn't working in the 90 days leading up. So I think there's a whole bunch of value to that, 00:31:18 both on non-technical founders being able to maybe create something quickly that can sort of test 00:31:22 a theory without having to involve a dev, but I'm also potentially using AI to then 00:31:27 read code and let someone use-- No, that's what I was trying to apply that to a use case. 00:31:37 And I think if I was giving advice, I would be very keen. And we've seen this, 00:31:42 we worked with a very specific startup where we walked in day one and the company we were 00:31:45 taking over for that we were really competing with, hadn't written a line of code in a month, 00:31:49 and they were paying. And I was like, just look at the fucking GitHub history, dude. 00:31:51 Right. They haven't committed anything since April 1st. 00:32:00 And so as a, but to your point, there's a product idea or there's a concept of, 00:32:05 if you're a non-technical co-founder, you should be evaluating not just the user testing. 00:32:08 Right. No. The quality of the code. The quality of the code. 00:32:12 Yeah, to be able to go in and be like-- That's very subjective. 00:32:17 Here's my repo. Tell me how deep of shit am I in right now? 00:32:22 I mean, you could use AI to literally go, yeah, your developer, here's the developers working on 00:32:27 your code. Here's how much they've cut. Here's their average amount of commits that could probably 00:32:32 qualify to how many hours they probably put into it. Here's the quality of each person's work. 00:32:38 I like it from like the data input side or the input side or where, no, but like when you're 00:32:45 talking about validating an idea. Right. So one of the problems that I see a lot is like, 00:32:49 there's people are like, I want to build something, but I'm not quite sure what I want to build yet. 00:32:54 And I don't have an infinite amount of money to build it. Right. And it's this whole song 00:32:59 and dance of like, what to build, when to build, how to build. There's so many questions to answer. 00:33:05 Well, if you knew, if you were six months down the road, then you could tell past me or someone 00:33:09 else to be like, this is what you should have built. Right. And we're all trying to figure out 00:33:13 what to build before we build it. But you can do that. Wait, can I stop? Can I pitch you guys my 00:33:22 idea? Yeah. Hold on. Hold on. I want to set it up. So Rory's got an AI product. We don't know what 00:33:27 it is. It's just an idea. It's an idea. But what they are. Rory's gonna pitch it to us. 00:33:33 And we're gonna, we're gonna, we're gonna sit and see, and see, we're gonna figure it out. 00:33:37 Let's go. Okay, I have never pitched this before. This is half brain. So we're gonna go and you 00:33:42 were kind of getting there honestly. So that's why it's sort of stuck to you. So I have an app, 00:33:47 the fan's place. It's a consumer application. People use it and play on it every single day. 00:33:53 Before you just brush that off, sure. I'd love a little bit, just a little bit more detail on the 00:33:58 application. Yeah, tell us a little bit about the fan base. Yeah, the company's called the fan's 00:34:04 place. And I've always called it the non-alcoholic beer sports betting. Right. So the idea is, 00:34:08 instead of betting money and losing, you play for free. You're in tokens. The tokens are worth 00:34:13 money. If you go to a local business, you know, we're mostly here in Indianapolis, you snap a 00:34:16 picture of a receipt, you turn it in the app, we give you a couple of bucks back, you know, 00:34:20 off that receipt. So, you know, instead of betting in DraftKings or FanDuel, you play our app. There's 00:34:24 no risk. And then we get you to go spend money at one of our partners in the community, you know, 00:34:29 you're supporting a local bar owner, you're supporting a local business owner. That's the, 00:34:35 the general loop. And so, you know, I think with anything consumer, in my background largely has been 00:34:41 B2B product in sales, mostly for startups. And in B2B, you can, you can get someone to commit to 00:34:46 paying you X amount of dollars to build something. And so, you know, it may not work, but at least 00:34:49 you're going to get that check. And then you can iterate on that. And you typically have a pretty 00:34:54 decent amount of length of time to then kind of go fix and iterate if you set it up the right way. 00:34:58 And B2C, you can do as many surveys you want. And until that, that consumer actually goes in 00:35:02 your application and buy something or does something, they honestly lie to you all the time. You have 00:35:06 no idea what's actually going to come back on the opposite end of putting something out there. So, 00:35:09 you spend a bunch of money, you spend a bunch of time, you're six months down the road, you put it 00:35:14 out in the wild, and nobody uses it or it does not go anywhere near what you thought it was, 00:35:18 and it was going to go. So my idea, and I honestly don't know if this is feasible, 00:35:24 which is why I'm really, really curious, could you create a company that allows you to kind of 00:35:30 custom create AI users that mimic a population, whether an entire population or a subsegment of 00:35:36 a population that is able to give you feedback on whether those users will want to use your product, 00:35:41 how frequently they will want to use your product, and then even how they might use your product, 00:35:46 because you can kind of mirror the segment that you're trying to sell to quickly enough. So, 00:35:52 I put my app out there, and I actually get a whole bunch of QA, not QA is not the right word, 00:35:56 I get a whole bunch of user feedback really quickly from a community of AI people 00:36:02 all using my app really, you know, in different ways based on their profiles that should mimic. 00:36:11 It's an AI user group, that basically you give access to the product, and then they try to go 00:36:19 and understand how it fits within their own demographic. Exactly. They're mentality. 00:36:26 So tech feedback, or tech consideration number one would be platform, so it'd be if it's, 00:36:31 you'd have to build specific technologies that AI would be able to interact with. 00:36:35 The easiest one always is web, because it's the most accessible technology out there. 00:36:41 So if you had a web-based product, it would be much easier to interact with, but you'd have to 00:36:45 build, so you'd have to build probably, I mean, you're assuming some of this is digital, 00:36:49 you're assuming there's some level of digital, or is it a pre-product, like you don't even have 00:36:53 anything yet. I think that going back to kind of what we're talking about, maybe you get a 00:36:57 simple version of the product out there, so you get a relatively simple version of it out there, 00:37:00 you send it out to this community of AI users, they give you really quick feedback, you get six 00:37:05 months of feedback in 12 hours. You iterate on it, you get a little bit more, you get a little bit 00:37:09 more, then you sort of build the full version, knowing at least I'm pretty confident some people 00:37:12 are going to use this, and I'm going to get something out of it instead of really not knowing 00:37:16 anything at all, sort of all the way through that journey. I think that would be where the value is. 00:37:20 So there's only one way to solve, there's a couple ways to solve it. Problem one is data science, 00:37:26 like literally doing a qualitative analysis on the market and the products that are in the market. 00:37:32 So nothing specific to your products at all, literally external factors, which isn't really 00:37:40 AI. It could be AI using an LLM, but the actual concept of an AI having an opinion about something 00:37:48 that is racking my brain right now. You would almost need to have some sort of a vision model 00:37:53 to be able to take and say, "Okay, here's the app," and then somehow this vision model 00:37:57 goes and extracts a bunch of the different screenshots of the app. That could be a thing. 00:38:05 Yeah, and then those models are like I'm a middle-aged woman who has very little technical stuff, 00:38:12 what's your analysis of this screenshot, and then it can come back and basically be telling you, 00:38:16 "I don't understand what these things mean," or what that could potentially be. 00:38:22 And I think it could help maybe not on what's going to differentiate your product, but on the 00:38:27 things that you might have a bias towards that don't work in the market already. So I use it all 00:38:30 the time for a bias check where I'm like, "Oh, I think this is a good idea," and they're like, 00:38:36 "Yeah, typically, we've read every word on the internet," and actually that's not what most 00:38:41 people are doing in this arena. Let's ask chat GPT what it thinks about the fan's place. 00:38:43 Oh, do it. Do it. Do it. 00:38:58 Okay, so I would like some feedback on my product. Think of it as the non-alcoholic. 00:39:01 Done. Thank you. 00:39:04 Fear of sports betting. 00:39:10 I do love that name. That's funny enough, so a buddy of mine built something that was kind of 00:39:16 similar. Yeah, something similar where it was this concept of being able to bet on your sports, 00:39:21 but it wasn't actually betting, right? Or betting towards the other things. He ended up, 00:39:24 I think he ended up shutting it down, but I was like, as soon as I saw yours, I'm like, 00:39:30 "You guys should totally connect because he has a React," and I don't know what your text is built on. 00:39:31 React Native. 00:39:33 So he's a React Native developer. 00:39:34 Yeah, it's going to be great. Okay, yeah. 00:39:41 So Zach, if you're paying attention. Okay, by the way, he needs to have a staffing agency. 00:39:41 Yeah. 00:39:45 You know the fucker? He really refers more to developers and jobs than anyone. 00:39:48 Oh, yeah, by the way, he's like, "Oh, you should have charged them all." 00:39:49 Dude, yeah, you should have. 00:39:53 Yeah, X-ray, if you're listening, you should pay me my 20% for finding your product. 00:39:56 I would like some feedback on my product. Think of it as the non-alcoholic fear of sports betting 00:40:00 you play for free and win tokens instead of money that can be used to local companies, 00:40:02 bars, restaurants. Yeah, for cashback. 00:40:05 Yeah. Well, should we ask for something more specific? 00:40:08 Hi, I'd like some feedback. Do you think that this would be a-- 00:40:15 Do you think sports fans would use this product frequently or would play this product frequently? 00:40:18 Does it have support for UFC fights? 00:40:20 Not yet, but we're getting there very, very soon. 00:40:22 Why don't you're there? 00:40:26 So, because here's the thing that people don't think about. There's a lot of folks who are team 00:40:30 sports, and then there's a lot of individuals like myself. I don't care about team sports. 00:40:36 I care about one-on-one fighting MMA boxing. Anything that's a one-on-one, 00:40:40 that's what it interests me. So, everybody who doesn't like sports loves MMA. 00:40:46 But I think it really is it. I think it's the difference between a team sport and a one-on-one 00:40:50 sport, right? I really think that that's the variability for me like-- Do you like wrestling? 00:40:53 I know why I wrestled. Yeah, yeah. Yeah, but do you think wrestling? 00:40:56 And then a little bit, not as much. So, I do a lot of Muay Thai, 00:41:04 Brazilian Jiu Jitsu, and kickboxing. But, yeah, UFC and MMA, any time that two people are going 00:41:07 at it, is-- No shit. We learned that. 00:41:11 And my specialty is my two dudes rolling around on the floor. 00:41:15 Let's send it in. I don't expect-- I mean, expectations here, I think it's going to give 00:41:18 a bunch of general bullshit. Yeah, I'll be curious as to what it says. 00:41:22 But here's a tangible example too while it's running. Yeah, yeah, go ahead. 00:41:26 To your question about UFC, when I started the company with my brother, we kind of felt like, 00:41:31 OK, this is going to-- the casual sports fan is going to want to use this product because it's 00:41:34 no risk, why wouldn't you? Even if you know nothing about all these sports, you can earn little bits 00:41:39 of money, and they actually are worth something in the community. And then, we thought that people 00:41:43 also really like sports a lot are also going to want to play, but maybe that was sort of secondary. 00:41:48 The primary motivator is something for free, and the secondary motivator is I love sports. 00:41:52 We found over time is people who come into the app and actually play frequently are people who 00:41:56 really like certain teams or sports, and they see contests. So, the primary motivator 00:42:00 has to be like IndyCar. We have a lot of good IndyCar users here because we're in Indianapolis. 00:42:06 People really like IndyCar here. Obviously, the 500 is a landmark event, but those users 00:42:10 stick around and play more frequently, and then they see there's other stuff and they might kind 00:42:15 of get into some other things. So, there is some truth to the fact that they'll sort of play anything, 00:42:20 but it had to start with, I love IndyCar and I want to play IndyCar contest. And I'm curious if this 00:42:26 product could have told me that by mimicking the sports universe and realized, yeah, I actually 00:42:31 really like sports. So, I'm back here all the time. I don't like sports, but I do like clipping coupons, 00:42:36 and I don't really want to spend this much time to earn a couple bucks. I'd rather go try to find 00:42:40 the one deal over here that gets me $5 immediately. That's where I'm curious. That kind of user 00:42:45 behavior is really hard to know or to ascertain immediately. And if you knew that sooner, 00:42:49 you could build the product a little bit better. You could also market better from day one and 00:42:52 spend those dollars. So, that's where the kind of the product idea came from. 00:42:58 Yeah, you got our minds going because that's a totally different question asking. 00:43:04 However, we've got some interesting stuff here from, I mean, it wrote a lot. 00:43:08 Is this 4-0? So, here's another thing. Oh, yeah, I'm on 4-0. 00:43:13 Okay, so here's another thing I realized about 4-0 is that thing wants to burn through tokens. 00:43:17 Like, even when I told it, hey, don't worry about output in the entire script, it always is like, 00:43:21 no, here's the full script. Here's 300 lines. It likes to burn tokens. 00:43:24 Broke up its answer to our question, which I'll repeat. 00:43:27 So, hi, I would like feedback on my product. Think of it as a non-alcoholic beer. 00:43:30 For sports betting, you play for free and win tokens set of money that can be used at local 00:43:35 companies, bars, restaurants, etc. Do you think sports fans would use this product frequently? 00:43:37 I'm kind of nervous here because I've gotten a lot of news from human beings. 00:43:43 So, your product concept of a non-alcoholic beer equivalent in sports betting world is 00:43:49 intriguing and innovative. Here are some points to consider. So, it broke up its response into 00:43:54 three sections, the positive, the challenges, and the feedback loop and the marketing strategies. 00:43:58 So, and I think these are really good answers. 00:44:03 So, you appeal to a wide audience, non-gamblers, people who enjoy the excitement of sports betting 00:44:09 are averse to gambling. Younger audiences don't have disposable income, but still want to engage 00:44:14 in sports betting and are all not allowed to because they're too young. Responsible betting, 00:44:17 this helps promote responsible betting and community market is a safer alternative. 00:44:22 Things that you've already thought of as a human being. For the most part, the kid thing is kind 00:44:27 of interesting. Local business engagement, win-win, local business benefit, community building, 00:44:31 strength and community ties with connecting sportsmen with local businesses, regulatory 00:44:34 advantage, legal ease, operating a non-monetary betting system might help you avoid some 00:44:40 stringent regulations. Frequent use, frequent use, frequent sports events provide regular 00:44:44 opportunities. Monetary value, how the fuck does it do this stuff? 00:44:50 It's magical, almost, right? And it's literally just what they do this from from some neural 00:44:54 network there. Word over word, the next word, and the next word, and the next word, 00:44:59 and out comes this emergent property that seems like it's intelligent. It's magical. 00:45:05 It really is. I know I've done a million of these, but like, do you want to live like this? 00:45:09 Every time. Every time. It's just like, this is crazy. 00:45:16 It really is. Okay. Challenges, monetary value, perception, value tokens, ensuring perceived 00:45:20 value of tokens is high. What do you think about that? 00:45:28 I think that makes sense on the face of it, but I don't think it's too hard to educate a user 00:45:32 quickly about why you would want the tokens. All you got to do is get them into the reward 00:45:35 loop immediately and show them what they can do very quickly. 00:45:36 With those tokens. 00:45:41 Like, what is the time to when they start getting tokens to where that turns into like a buck or 00:45:46 whatever to? Yeah. I mean, I think the easiest thing is you put an offer out there for 50 tokens 00:45:51 on signup, and it's worth $5. And then if we get the monetization strategy right on the business 00:45:55 side, and I send someone in with $5, I'm going to end up making money on that because the business 00:46:00 does a rev share off of the additional revenue. Most people you talk to will happily give someone 00:46:05 $5 to walk in off the street and purchase something. So what you're doing is just making an instant 00:46:08 right away they get some info. 00:46:11 Yeah, and then it's like a whole virtual gift carding system for small businesses, in a sense. 00:46:13 Redemption process? 00:46:18 Very tricky. Very difficult. That's a really good one. 00:46:21 But also an advantage to use in some ways. 00:46:22 How are we drinking? 00:46:24 If it's more tricky, the redemption process, unless you're cost a good soul. 00:46:28 What is the redemption? 00:46:32 Right now I just ask people to snap a picture of a receipt, and then we could use receipt scanning 00:46:36 technology to automate. At the moment, I literally look at the receipts. 00:46:37 Yeah, I mean, I don't really care about that one. 00:46:39 They can totally use an attractive use of that. 00:46:41 Redemption process will be crucial. 00:46:45 If users find it cumbersome, they might lose interest. 00:46:52 User acquisition, attracting users, cryo-effective marketing, attention, blah, blah, 00:46:55 that's all. I thought a lot of these answers were really good, but that was just general. 00:46:59 Competition, other free-to-play models. There are already numerous fantasy sports, 00:47:02 and other free-to-play sports betting platforms. 00:47:07 Yeah, so how are you positioning yourself against the other ones that are out there? 00:47:10 Yeah, so honestly, no one does it quite like this. 00:47:13 I mean, almost all the other free-to-play is either a sponsored contest in a draft 00:47:18 Kings from Bud Light or Miller Corps or whoever, and it's 50,000 people competing for $5,000. 00:47:23 Or you might find some free-to-play thing, a sort of an entry to get you to sign up for 00:47:25 something and then continue to market. 00:47:30 Yeah, the brand wants to market to you. The reason for you to play honestly hasn't worked on a big 00:47:33 scale is because no one's figured out how to get enough rewards in the rewards pool 00:47:38 for a user to be incentivized to continue to play. What people have done in the past is 00:47:41 try to get enough clicks and views that they can advertise, and that a certain percent of the 00:47:45 advertisement dollars come back into the pools. It never really works. 00:47:49 I think the way to do it, obviously, is to get businesses that have the money. 00:47:54 You've got to find somebody who wants that consumer to basically pay into the pot over time, 00:47:59 and then you get this redemption cycle that actually works and is enough. 00:48:02 That's why that's essentially how we differentiate. There's real money there. 00:48:05 You just have to take an action to be able to get the value. 00:48:07 And when did you launch this? 00:48:13 Let's see. We started talking about it April 2020. We launched it officially. 00:48:15 We had a beta customer and stuff. We launched it officially in August of '22. 00:48:16 Okay, cool. 00:48:17 Yeah. 00:48:20 The only other things that it has is you should collect feedback from users, 00:48:24 which we are still trying to figure out how AI would do that. I just don't have an answer. 00:48:27 Have you figured around yet? 00:48:31 We could totally do that. Is that a rule of efficiency? 00:48:34 That's because that's the product. That's the other product. 00:48:37 Okay, this is the last thing. This is pretty good. 00:48:41 You should use local sports figures to spread the word. 00:48:42 Yeah, absolutely. 00:48:44 Well, Pat McAfee. 00:48:45 Pat McAfee, yeah. 00:48:46 If you're listening to this podcast, give me a call. 00:48:54 Yeah. Hey, Pat, anybody who has access to Pat, this would be a perfect relationship with you 00:48:59 guys. Like there's like seriously, so anybody that's listening that has a relationship to Pat, 00:49:04 that would be the ideal guy, 100%, no question. 00:49:08 So overall, your product has a unique niche that could resonate well with sports fans who 00:49:11 enjoy the thrill of betting, but prefer a non-monetary system. 00:49:15 The success of the product will depend significantly on effective marketing, the perceived value of 00:49:17 tokens, and the ease of redeeming those tokens. 00:49:19 Yeah. 00:49:22 How much of this is an Indianapolis play? 00:49:25 At the moment, almost all of it. 00:49:28 You get to get so much of the payout is ultimately tied to... 00:49:32 It is. We found out quickly the more places you can go redeem the better. 00:49:35 And also, back to your question about how we compete. 00:49:37 It's actually probably less about competing against free to play. 00:49:41 And it's more about competing against just all the other things you can do around sports, 00:49:45 especially going back to realizing that the person has to be a sports fan first. 00:49:48 And then the free... You know, you can win stuff for free as a secondary motivator. 00:49:50 Well, if you're a sports fan, you can play draft games, you can play fandile, 00:49:53 you can play free fantasy sports, you can play a whole bunch of stuff. 00:49:57 So what we found is anytime you hop in any of those apps for the most part, 00:50:00 everybody's seen the same thing, whether you're in California or Indiana, 00:50:01 it's not localized. 00:50:05 And so when you hop into our app, it's a lot of Pacers, contests, and Colts, and IndyCar... 00:50:06 That is very unique. 00:50:08 That is the... Yeah, that's how it's unique. 00:50:12 So that's why you need the right teams and the right region. 00:50:14 And then... But it all works together, right? 00:50:17 I'm trying to get you to grab your friends and go have a beer and support your local team 00:50:20 at your local bar and play our local contests all the time. 00:50:22 So the interesting thing about the tech, without going too far... 00:50:23 I mean, it's sort of a big problem. 00:50:25 We have the localized fantasy sports product, 00:50:26 which is very difficult. 00:50:30 We're in a ton of contests on different teams and different cities at the same time. 00:50:33 And then we also have to build a rewards network. 00:50:34 And those two things have to work together. 00:50:35 That's it. 00:50:36 Let's go ahead. 00:50:38 Well, I was going to ask, do any of these bars... 00:50:40 So there's a brewery down the street for me, 00:50:43 and they have adopted some soccer team overseas. 00:50:46 But are any of these bars identifying themselves as, 00:50:48 "Where are the Pacers bar? Where are the Colts bar?" 00:50:50 I know those are obvious, 00:50:52 because most bars are going to be playing those sports. 00:50:56 Is there an opportunity there that you've seen to be like, 00:50:57 "Hey, this... What is the Chatham?" 00:50:58 Or no, yeah. 00:51:01 Chatham is like the soccer bar, right? 00:51:02 Yeah, so we're actually partnered with Union Jack. 00:51:05 They do indie 11 watch parties. 00:51:08 They do a lot of Premier League. 00:51:09 It's just in there the other day. 00:51:13 And the Borussia Dortmund supporters crew was there 00:51:14 for the Champions League final. 00:51:15 Unfortunately, they lost. 00:51:18 There is definitely a market for leagues and teams 00:51:20 wanting to host watch parties at places 00:51:22 and sort of try to match those two things up. 00:51:26 Yeah, we can capture some of that attention without a doubt. 00:51:30 But it's kind of like anything else. 00:51:32 We just need... We need more eyeballs than that in the app 00:51:33 to create the community. 00:51:36 So we'll use Pearl Street. 00:51:38 So Pearl Street is... 00:51:40 They've got to play state downtown, right? 00:51:42 And they've got to play really close to my house. 00:51:43 Oh, yeah. 00:51:44 Yeah, off of Olio. 00:51:46 It's in Geist and Olio. 00:51:48 And so I'll go there and I'll hang out with him 00:51:50 and I'll talk with him and his wife a lot and just love. 00:51:53 I love the restaurant business just as... 00:51:54 I would never want to be in it, 00:51:57 but I love to hear the plight that they have to go through. 00:51:59 But so like for somebody like them, 00:52:01 when they hear that there's a product like this, 00:52:05 what could they do to kind of like become part 00:52:07 and to participate with what you've built? 00:52:10 Yeah, I mean, we've spent a lot of time and effort 00:52:13 making it as light a lift on the actual bar restaurant owner. 00:52:16 As possible, you probably know from your conversations, 00:52:19 most of these folks don't have a support staff 00:52:20 to do any sort of marketing. 00:52:23 So they've got their hands tied just running a kitchen 00:52:24 and making sure people show up on time to work. 00:52:29 And so there's not really a lot of work on the staff, 00:52:33 but we have sort of a two-tier model 00:52:34 where we have a lot of places in the app 00:52:37 that you can play and you can redeem for a call 00:52:39 at 25 tokens as a dollar. 00:52:41 And then if someone wants to be really a part 00:52:44 of the upgraded network, then 25 tokens, 00:52:45 or sorry, 10 tokens equals a dollar. 00:52:47 So you increase your redemption value, 00:52:48 you go into those places, 00:52:49 they become much more attractive in the app. 00:52:51 And we do see people go in there more often. 00:52:54 So that's how like a bar like or a place like Pearl Street 00:52:57 could say, I want to be sort of the second tier 00:52:59 or the premium tier. 00:53:01 More people are going to go there than to the other spots. 00:53:04 And all we got to do really is just change it in the app 00:53:05 and start driving traffic. 00:53:10

• So if I'm Pearl Street, how would I contact you 00:53:12 to say, I want to be part of this? 00:53:14
• I just call you up. 00:53:17
• Rory info@thefansplaced.com. 00:53:19 Rory@thefansplaced.com. 00:53:21 We got a contact thing on our website, 00:53:24 but yeah, we're in the market walking into places 00:53:26 and stuff too, so yeah. 00:53:28
• So I've been thinking about what Rory was talking about. 00:53:33 And I think that there is not a, 00:53:34 I don't think there's a clear, 00:53:38 I don't think there's a product market fit answer 00:53:41 to that question right now. 00:53:43 But I think that it breaks the question, which is, 00:53:51 okay, in order for AI to provide feedback on an actual thing, 00:53:53 that is valuable. 00:53:56 That AI has to have an opinion. 00:54:03 And that opinion is based off of some objective prompt. 00:54:04 It's going to be whatever we tell it. 00:54:06
• It's based off right. 00:54:10 But the question I think becomes is, 00:54:14 it's almost like to have an opinion, 00:54:16 you have to have some sort of emotional experience, 00:54:18 because you're taking all the information 00:54:20 that's been coming into your whole life. 00:54:25 And you don't say everything is on an even playing field. 00:54:31 All the information is taken in with some level of subjectivity 00:54:32 based off of how you perceived it, 00:54:34 and your perception is based off of your emotions 00:54:35 and your experience. 00:54:38 And AI doesn't have emotions and experience. 00:54:40 It just has information. 00:54:42 And it just has the next token. 00:54:46
• That's why it would be useful for like Duolingo, 00:54:47 spend a ton of money to try to figure out 00:54:51 when the best time to send a reminder to do your lessons. 00:54:53 And what they found out after a bunch of AI, 00:54:55 a bunch of training was about the same time 00:54:57 that the person did the lesson the day before, right? 00:55:01 That's when they were available. 00:55:02 That's not when they did yesterday. 00:55:07 But my point was like AI or large language models 00:55:08 have some of that information baked in, 00:55:11 like the notifications from the way, 00:55:13 the way Instagram batches notifications 00:55:15 or the way this works or whatever. 00:55:16 These are just things. 00:55:18 That's why there's off libraries. 00:55:20 And there's these conventions to do things. 00:55:23 It would be able to give that information pretty easily of like, 00:55:25 oh, here's the direction I would go 00:55:28 if you're trying to implement notifications 00:55:30 or if you're trying to do this or reminders or whatever. 00:55:33 But then to the emotional, 00:55:34 like if someone's sitting there but in a bar 00:55:37 and they do that regularly, 00:55:40 would they be willing to go? 00:55:43
• But does that make the AI actually more valuable? 00:55:45 Is the information we'd got back right there 00:55:47 actually more valuable because it's less emotional? 00:55:49
• Well, I think it's more valuable 00:55:53 if it's for the purpose of making unbiased decisions 00:55:55 and the direction of the app. 00:55:58 But when it comes to what humans are uniquely good at, 00:56:00 which is like finding the value proposition, 00:56:03 what's unique, like you can to build the same thing. 00:56:04 Yeah, it's going to be more valuable. 00:56:07 But to make something that's differentiated, 00:56:11 you need to sit there and watch a dude or chick at a bar 00:56:13 and will they play the thing, you know? 00:56:15
• I think we're actually going back to honestly 00:56:15 the very first question, 00:56:18 which is can we get to a billion dollar company 00:56:22 with a single person and I, on the one hand, 00:56:25 my gut reaction to that was no. 00:56:26 As soon as I thought that was ridiculous. 00:56:28 But then I realized, I remembered, 00:56:29 Instagram was bought for a billion dollars 00:56:31 by Facebook in 2012 with 13 employees. 00:56:35
• Wait, wait, wait, wait, wait. 00:56:36 It was only 13 employees and they got bought. 00:56:38
• 13 employees, holy shit. 00:56:39
• Yeah, so I thought, okay, 00:56:41 so maybe there's this crazy case out there 00:56:43 or someone and you're talking about, 00:56:44 this is valuation, right? 00:56:46 So someone sees some crazy idea. 00:56:48 If there's any froth in the VC market right now, 00:56:50 it's in AI products. 00:56:52 Someone creates something and it catches on like wildfire 00:56:54 and someone says, with a lot of money, 00:56:56 he says, I gotta get this before anybody else 00:56:58 gets their hands on it, I'll pay a billion dollars for it. 00:56:59 I think that's possible. 00:57:01 But when I think about all the things 00:57:02 that I have had to learn, 00:57:06 I don't think you could do this in the B2B world 00:57:08 with one person in AI. 00:57:09 I don't think that's possible. 00:57:12 In the B2C world, I think it's theoretically possible, 00:57:14 but I think without the emotional component, 00:57:18 I don't know how you can learn enough 00:57:21 with just one person and a bunch of AI helpers 00:57:24 to get to that point with any sort of consistency. 00:57:26 Again, the moonshot may be a lottery ticket. 00:57:29
• It might be a luck shot that you hit. 00:57:30
• Exactly, but I don't think we're gonna see 00:57:33 this wave of companies happening with one person 00:57:34 because I think you need enough people 00:57:36 to take in information and process it 00:57:38 at a human sort of emotional level 00:57:41 to make the right decisions, find that product market fit. 00:57:43 And then once you have that, 00:57:46 then I think you can get some AI helpers 00:57:48 to really help your business explode. 00:57:51 But I think it's just very difficult in there. 00:57:53
• So if you had a really, really awesome 00:57:55 differentiated product market fit idea, 00:57:57 everything's ready to go, 00:57:59 then you're saying one person could take it 00:58:00 once you got to that point 00:58:02 and maybe elevate it to that level, 00:58:05 but it's all the work, all the trial and error. 00:58:07
• I think so, and I think Brandon, 00:58:08 what you're saying earlier about your code 00:58:10 and realizing maybe this is actually 00:58:11 going to require more developers, 00:58:14 I think there's some technical challenges too to do in this. 00:58:15 So it's probably not quite that simple, 00:58:17 but I do think the human element 00:58:20 is finding that product market fit and working through that. 00:58:22 And then once you get there, 00:58:24 yeah, the mechanics of how do I scale this thing 00:58:28 is really where I think the AI help can come into play. 00:58:30
• So my dad, and this was years ago, 00:58:33 my dad said, I think we're going to see 00:58:35 a millionaire eight-year-old 00:58:38 that's gonna come out from all of this. 00:58:41 And sure enough, like five years later, 00:58:45 there's this one kid who does YouTube reviews on toys. 00:58:49 He's a millionaire and he makes like eight years old. 00:58:50
• Those economies have all changed. 00:58:53 I mean, there's money coming out of places 00:58:54 that never existed before, 00:58:57 but I'm still stuck on this detail about AI, 00:59:02 which is really people talked about general intelligence. 00:59:06 And I mean, that was a very coherent thought. 00:59:11 If I could have spent a day with your product, 00:59:15 talking to people, and probably come out with a similar response, 00:59:16 and you would have thought, 00:59:17 you know what, you spent a lot of time on that. 00:59:19 You did a really good job. 00:59:19
• Exactly, here. 00:59:21
• That responded to me in within five seconds 00:59:22 while you were talking. 00:59:23
• Yeah. 00:59:25
• In between that trip to go to the bathroom. 00:59:27
• Yeah, so what does AGI actually look like 00:59:29 if this is where we're at right now? 00:59:34
• I think that it comes down to execution. 00:59:39 I think that the information might be closer than we think it is. 00:59:47 But the AI is gonna maybe figure out what is strategy? 00:59:48 What should I be doing? 00:59:52 Not just being asked questions going back and forth, 00:59:56 but thinking about, hey, you should do this. 00:59:58 I'm thinking, you know what I mean? 00:59:59 This is what you should do right now. 01:00:00
• Yeah. 01:00:02
• Not like, oh, you know, you-- 01:00:03
• A reactionary thing to me. 01:00:05
• It's all reactionary right now. 01:00:07
• This way it could come back and be like, hey. 01:00:08
• And not in a way of like, 01:00:10 hey, the average American needs to work out 01:00:12 three times a week and your heart rate is, 01:00:14 you know, not that kind of stuff. 01:00:16 Something truly unique that's like, you know what? 01:00:19 If you take a right here, 01:00:22 pick up the pizza from Arnie's on the way back home. 01:00:24 You might be able to get, you know, 01:00:27 the kids to the game, you know what I mean? 01:00:31 On time for the first time this week, you know what I mean? 01:00:33 Like literally thinking strategically 01:00:36 based off the information it's given. 01:00:40 Not necessarily just like like a heavy handed Siri. 01:00:44
• We have it at a certain level for mapping, right? 01:00:47 Like Google can tell you, hey, there's a wreck 01:00:50 and we have data that says that it's slow on that road. 01:00:51 So take that road instead. 01:00:55 That's a simple application of what you're saying. 01:00:57 But like, that would where it gets really exciting. 01:01:00 Right now I see this as like the guiding light of, 01:01:03 I don't know enough, help me fill in my knowledge gaps. 01:01:05 Now I can go talk to a professional 01:01:06 and they can give me the next steps. 01:01:08 And I can feel very confident 01:01:09 what you were talking about earlier. 01:01:11 How do you know the developer did the thing 01:01:12 that they said they did? 01:01:14 Well, if you can go and ask a few questions 01:01:16 and get a general concept of what they're trying 01:01:19 to accomplish and how they're going to accomplish it, 01:01:22 then you don't fear or not understand 01:01:24 what it takes to do what you've just asked them. 01:01:26 Maybe you don't ask them to do that now, you know? 01:01:29 So like that's where it's like, that's where it's at now. 01:01:30 But that'd be really cool. 01:01:34 It'd be cool if I could tell you a turn right here, you know? 01:01:37 Yeah, and then also it also comes down to the prompt. 01:01:38 Like chat GPT's got a slight, hey, 01:01:41 be slightly jovial and positive, you know, 01:01:42 there's some underlying problem. 01:01:46 And that's the bias thing that you don't miss 01:01:48 is that's what to your point, like the person, 01:01:50 the person might lie to you. 01:01:52 So you don't get the answer from the human or the AI. 01:01:53 We should, you know it. 01:01:54 Yeah, exactly. 01:01:57 If there's anything that's I think a death sentence 01:01:59 and a lot of startups, it's just not, 01:02:03 people just not giving you super honest, true honest feedback. 01:02:07 And I've definitely had someone along the way and I love it. 01:02:08 I mean, it hurts sometimes in the moment, 01:02:11 especially if, you know, the percent of that feedback, 01:02:14 that's tough, but that's what helps you iterate faster 01:02:15 than anything else. 01:02:16 Let me ask you real quick. 01:02:19 So one of the things that I've found is that the feedback, 01:02:21 the product feedback that I would get, 01:02:24 that is the most guttural punch, right? 01:02:27 The most painful ends up being the most accurate. 01:02:28 Absolutely. 01:02:31 Yes, the one that I'm like, "Fuck you, you're so wrong!" 01:02:34 I go then home and I start thinking about like that. 01:02:36 Son of a bitch was 100% right. 01:02:40 That, yeah, anyone's that I have like a very visceral reaction to, 01:02:42 tend to be the most valuable feedback that I've ever gotten. 01:02:46 Yeah, also similarly as tangent, but in sales, 01:02:47 the people who are actually really honest with you 01:02:49 about your product when you're trying to sell them to, 01:02:52 and even to the point of sort of difficult with you, 01:02:54 oftentimes end up being your best customers on the other side. 01:02:56 100%, yeah, totally. 01:03:00 Yeah, I do think that the honesty is really key and helpful. 01:03:01 It sounds like kind of what you're getting to though, 01:03:05 is there's a way for AI to sort of fill out your idea card, 01:03:07 so to speak, or maybe elevate the ideas you have 01:03:10 that you want to try to get to that next level. 01:03:12 Perhaps it can help you shorten the feedback loop 01:03:14 by getting product to market more quickly. 01:03:17 What it's not going to do is provide 01:03:21 the actual kind of human interaction with your products. 01:03:23 I know it can't do that, but that was sort of my question is, 01:03:24 do we think it's possible? 01:03:27 I mean, I think that there's, I'm remembering one a few places, 01:03:29 but yeah, I think that you summarized it well. 01:03:34 And, you know, based off research, 01:03:37 one of the things that AI is not necessarily getting to 01:03:39 is going out into the world and executing things. 01:03:42 They're talking to us, they're responding to us, 01:03:45 but they're not to the point where they're doing- 01:03:46 They're not pro-actists, yeah. 01:03:49 They're like rabbit-tried, you know, they try to- 01:03:53 There's auto-GPT, which is sort of, you know, 01:03:55 a thing that iterates and can execute a little bit, 01:03:56 but it's not good enough yet. 01:03:58 But to Brynd and Brynd and I were talking about it yesterday, 01:03:59 it's such a lift. 01:04:01 And so we did a little bit of research, 01:04:03 and basically we went through all of the- 01:04:09 We went through all of the most popular AI product categories. 01:04:14 So the number one most popular AI product category right now, 01:04:17 or the number one most diluted product category 01:04:18 is image generation. 01:04:19 Okay. 01:04:19 Okay. 01:04:21 Second content generation. 01:04:22 Yep. 01:04:25 So, and four out of the top five is creativity-related things, 01:04:30 just creating stuff, chatting, writing, SEO content. 01:04:34 But then we cross-referenced it with the most popular, 01:04:38 but didn't have the least amount of products. 01:04:39 Yep. 01:04:39 Okay. 01:04:43 Wait, sorry, explain that to me. 01:04:47 So let's say that there's 500 image generation AI products 01:04:48 in the registry, right? 01:04:49 Right. 01:04:52 It's also- It's very popular. 01:04:56 But there's products that are highly shared, 01:04:57 but they're not diluted. 01:05:00 So they only have five or six products on the market 01:05:01 in their category. 01:05:03 But people, everybody wants those products. 01:05:03 Right, right. 01:05:07 And then three of the top five categories 01:05:09 had to do with virtual girlfriends. 01:05:12 Okay. 01:05:17 But the other categories auto-GPT. 01:05:17 Right. 01:05:17 Yeah. 01:05:21 And that is where you're not just getting stuff from 01:05:22 from AI. 01:05:24 It's going out and doing stuff for you. 01:05:25 Yeah. 01:05:28 And I think that's really the next step in AI is- 01:05:30 And it's such- It's hard because you're- 01:05:33 It's, hey, go to the internet. 01:05:34 Go use this app. 01:05:36 You can't even leave this- 01:05:37 You're not a human. 01:05:38 You can't walk. 01:05:38 You don't have legs. 01:05:40 So like you can't do that, right? 01:05:41 You don't have legs. 01:05:45 I mean, I can't wait for the first big brand to say it wasn't us. 01:05:46 It was already us. 01:05:46 It was sorry. 01:05:47 Yeah. 01:05:48 It wasn't me. 01:05:48 It wasn't me. 01:05:48 Yeah. 01:05:49 All right. 01:05:49 I did that. 01:05:50 Yeah. 01:05:50 I'm not seeing it. 01:05:53 I mean, and I truly believe that that- 01:05:55 So last year was all about chat GPT. 01:05:56 I think any year from now, 01:05:58 it's going to be all about either some robot, 01:06:00 somebody built that actually works, 01:06:03 or some sort of auto GPT that's doing everything for you. 01:06:06 But wasn't there that one bought auto GPT bought 01:06:07 that like was selling? 01:06:08 Yeah. 01:06:10 So, no, the one that was selling feet picks 01:06:10 and made like- 01:06:11 Oh, Jesus. 01:06:13 Change brand or something a month. 01:06:14 I mean, and they like had- 01:06:15 It was like a- 01:06:17 It was like an only fans or something to count those all- 01:06:18 That's actually brilliant. 01:06:18 Yeah. 01:06:19 It was so, really brilliant. 01:06:22 Dude, are there any interest in dudes' feet picks? 01:06:22 I know. 01:06:23 I'm just asking you. 01:06:24 I'm just asking you. 01:06:24 Yeah. 01:06:25 Hypothetically. 01:06:26 That's your virtual girlfriend. 01:06:26 I've been having all- 01:06:29 I'll ask open Hermes. 01:06:31 Have you ever seen that one video where the- 01:06:33 Everyone, the friends are all sitting on the couch 01:06:34 and all of a sudden, 01:06:35 the friends at next to the other one's a girlfriend, 01:06:37 the flash goes off on their camera? 01:06:37 Oh, yeah. 01:06:39 And they're like, what the hell, man? 01:06:40 That's right. 01:06:41 See the photo of her feet? 01:06:42 Yes, right. 01:06:44 The whole feet thing just blows my mind. 01:06:45 Like, I don't get it. 01:06:46 Now, I kind of- 01:06:48 I'm starting to get it a little bit. 01:06:48 Don't try to get it. 01:06:49 When you were talking about your seven, 01:06:50 eight minutes in incognito, 01:06:51 you weren't talking about feet? 01:06:52 No, I wasn't. 01:06:52 Okay. 01:06:54 But however, that being said, 01:06:56 I will say everyone once in a while, 01:06:58 my wife will be sitting there with her feet up, 01:06:59 and I'll be like, 01:07:01 I could do some weird shit to those. 01:07:04 Like, I can see like, is this an age thing? 01:07:05 As I get older, I'm thinking like- 01:07:07 This will be edited out. 01:07:10 No, no, no, no, no, no, cut, clap, cut. 01:07:13 But that said, hey, that was awesome combo. 01:07:15 We went longer than usually, 01:07:17 which is usually a good thing. 01:07:19 So, yeah, I think we- 01:07:23 So, real quick, how can people learn about fan's place? 01:07:25 Yeah, thank you. 01:07:26 The fansplace.com, 01:07:30 and then we've got apps in both the iOS and Android stores, 01:07:32 especially if you're in Indianapolis. 01:07:34 Lots of games to play, 01:07:36 lots of places to go and redeem. 01:07:38 We have people redeeming stuff all the time. 01:07:40 It's pretty easy to get five or 10 bucks really quickly, 01:07:41 and then go get a free beer and- 01:07:46 And if you're a restaurant or an establishment here in town, 01:07:48 you can contact- 01:07:49 Yeah, either Rory at the fanspace.com, 01:07:51 or there is a contact form on our website. 01:07:52 Check it out. 01:07:52 Check it out. 01:07:53 That'd be great. 01:07:53 That'd be great. 01:07:54 It's been great- 01:07:55 It's been fun. 01:07:56 Watching your journey and- 01:07:57 Thank you. 01:07:58 Best of luck going forward. 01:07:59 Thank you. 01:08:00 All right, guys. 01:08:02 This has been the Big Cheese Podcast. 00:00:00 [silence]

Prompt: Include Chapters in the Show Notes.