How I AI - How a VC and tech founder used AI to launch a brick-and-mortar business in their spare time | Andrew Mason (CEO of Descript) & Nabeel Hyatt (Partner at Spark Capital)

Starting point is 00:00:00 There's this kind of renaissance going on with board games that I've just gotten clued into in the last couple of years. Both of us had independently thought, wouldn't it be cool if there was like a membership-based social club, like a physical space that you could go to and just show up and play board games, either bringing your own friends like you might do at a board game cafe if you've ever seen one of those, but also you could play with other members. If you're starting your board game navigation and maybe if you've ever been inside of a board game cafe, you've had this feeling, you walk in and you've made. you played Catan and maybe your friends had you play Wingspan and now there's just like a wall of games and you have absolutely no idea where to start. Well, this has been solved. In libraries, you have a card catalog library system where all of the French history books are next to each other and so and so forth. This is a kind of thing that would be like literally impossible without AI, but we basically have a bunch of design prompts in Claude back in those projects that help

Starting point is 00:00:52 to categorize into kind of a dewey decimal system of board games. I've seen how AI can help you code at an extremely proficient level. You can PM, but you are showing me how you can take your slightly nerdy, little bit niche hobby to an extreme level and even turn it into a business. It's truly the most ridiculous idea. If he had proposed this and we didn't have AI, I would have gone to the mat. Like, this is such a crazy thing that nobody's going to want and it's going to be such a waste of time. But because we do, we can indulge in this kind of stuff and it's really fun. And maybe for every five of these, one of them will be the best thing ever. Welcome back to How IAI.

Starting point is 00:01:35 I'm Clarevow, product leader and AI obsessive, here on a mission to help you build better with these new tools. Today I have Andrew Mason, CEO of Descript and Nabil Hyatt, VC at Spark Capital. Sure, they're a founder and a VC by day, but by night they used AI to take their hobby and turn it into an in-person business. Let's get to it. This episode is brought to you by Lovable. If you've ever had an idea for an app but didn't know where to start, Lovable is for you. Lovable lets you build working apps and websites by simply chatting with AI. Then you can customize it, add automations, and deploy it to a live domain.

Starting point is 00:02:12 It's perfect for marketers spinning up tools, product managers prototyping new ideas, or founders launching their next business. Unlike no-code tools, Loveable isn't about static pages. It builds full apps with real functionality. And it's fast. What used to take weeks, months, or even years, you can now do over the weekend. So if you've been sitting on an idea, now's the time to bring it to life. Get started for free at lovable. That's lovable.dev.

Starting point is 00:02:44 Nabil and Andrew, thank you for being here. What I love about what we're going to talk about is I've seen how AI can help you code and an extremely proficient level, you can PM, like 10 PMs, but you are showing me how you can take your slightly nerdy, a little bit niche hobby to an extreme level

Starting point is 00:03:07 and even turn it into a business. So, Andrew, can you tell us a little bit about what you built with AI? Nabil and I, the way we got to know each other is I'm the founder of an AI video startup called Descript. Nabil is a, partner at the venture firm Spark. He invested in Descript. And it turns out that we actually live

Starting point is 00:03:31 a couple blocks from each other. And it turns out that both of us have gotten into board games over the last couple years. Do you know anything about board games? I have a six and eight year old and we're just entering board game era in our house. It's a long road ahead. Yeah. Yeah. But when you say that, do you mean like monopoly and stuff like that? We run the gamut. We do. We just got Scrabble because we're spelling now. So that's good. We got like the dungeons. Well, dungeon mayhem. You know, we're playing like a lot of dungeon mayhem. So we're stepping into like the D&D world. And then no, we have not gotten into these complex board games. Although my friends are trying to get me into this like bird call game, like this birding birding board game. Wing span. It's definitely wing span. There's a there's a small cohort of parents in Bridal Heights into wingspan.

Starting point is 00:04:17 Yeah. So that's a good example. There's this kind of like renaissance going on with board game. that I've just gotten clued into in the last couple of years. I kind of always wrote it off as this incredibly nerdy thing, as you put it. But it's actually a really fun, social way to come together, an excuse to get off your screens and do something. And Nabil and I started getting together to have like game nights and play together and just found it like hard to pull off

Starting point is 00:04:49 with everything that was going on. And one day we were, sitting around just like talking about this and it turned out that both of us had independently thought wouldn't be cool if there was like a membership based social club like a physical space that you could go to and just show up and and play board games either bringing your own friends like you might do at a board game cafe if you've ever seen one of those but also it could you could play with other members and you could offer a kind of service to help people organize games with other members, so you could just be like, I want to show up and play this kind of a game on Friday.

Starting point is 00:05:26 Can you see if other members are interested in doing it? So we were like, oh, this would be so cool. And the next time a retail space opens up in the little shopping area in between us in Berkeley, we're going to have to rent it, even though it's totally irresponsible. And neither of any time for something like this. And sure enough, that's what happened. So we rented this space when, Nabil? This was probably four months ago, five months ago, something like that.

Starting point is 00:05:54 Yeah. And we've been on a crazy adventure trying to figure out how to be small business owners on top of two well over full-time jobs that were also managing and families and the rest of that stuff. See, I don't think the listeners or the watchers were prepared for this episode to be what I built with AI, retail location, physical board gaming cafe. So let's get into what, because this is a very in-person. service-oriented human physical business that you're building, very different, Andrew, than I think, your business, your AI business that you run now. How did you use AI in the mix of standing up this new opportunity? I think the first thing and the reason I think we feel good talking about it here is that like there's just no way this business would have existed without AI at about

Starting point is 00:06:43 100 different levels, which will become very obvious very quickly. So first of all is just neither of us are small business owners, neither of us have ever run retail locations. This is like one of the millions of things that millions of people just like randomly say over dinner or talking amongst their spouses or whatever. We're just like, wouldn't it be great if? And AI, and in particular in this case, Claude is very good at being like, that would be great. Sure, you can do that. And then, you know, this is ostensibly an exercise of like, you know, pulling a thread until the like duvet covers all over the you know, room and you end up with a retail location. And so the first bit is just we use Claude to ask simple questions like, I want to open up a place to play board games with my friends

Starting point is 00:07:30 and Berkeley. These are my goals. And instead of being prescriptive, you're trying to just say what you want. And then it giving us lots of information. So this is a couple of weeks into the project is now months ago, an example of just it's giving us how to lay out the space properly, what the budget should be, you know, how to think about pricing, financial projections, the like myriad of 15,000 different things that you'll need to do in order to get this thing open. And you said this very simply. You just said sort of what's the business plan for this? Or was this a complex prompt? Like, how did you get to this level of detail on your business plan? I'd say there were three things that we at core used AI for throughout this project.

Starting point is 00:08:12 The first one is, sensibly AI as a co-pilot business partner. It's making the business plan. The second one was AI as kind of like a manual worker, which we'll get to, which is just an unbelievable amount of just like grunt work in getting this open, not least of which is something like trying to categorize hundreds of board games through a unique Dewey Decimal System we invented and then trying to individually label them, which we'll get to in a minute. And then lastly, is trying to use AI to make new kinds of experiences. We built this AI. concierge service where you can actually just like text a number and say, I feel like playing a certain type of game with these guys of people on a Friday afternoon. Can you find them and figure it out? And all those coordination problems that end up being kind of impossible to articulate until you have an AI kind of making sense of the back end. So in this first phase, yeah, we don't know what we're doing. So a lot of the process I'd say was a cycle of more simple prompts. These are not massive PRDs. We don't even know what it is we want yet. And so I would say the cycle was much more like we make a document. And if I had to time of externalize it now, I'd say the pattern was instead of it being a chat, we're talking to the AI, it talks back to you and you talk back.

Starting point is 00:09:23 A lot of it was document generation. And then that document generation would become context for the next set of conversations. And so you're slowly working through like a little bit of a mission statement and a little bit of a business plan and so on and so forth. And then you slowly move that back in. And so each time you're talking back into a Claude project or whatever we happen to be using, it's gaining context more and more and more as you are moving to the project. Can we look at that business plan again? Because I want to go through a couple of the specific use cases. And what I think is interesting about them is you both are busy people.

Starting point is 00:09:57 And so I can imagine all of these problems, your big brains, they're all solvable problems in some sense. Or some of them are solvable problems. But I bet the speed at which you could solve them was kind of like. tremendously helpful to you and your time-constrained world. And so I'd love to talk through like a couple. I think there's some like space layout, use cases. Maybe we can start with that just as an example. Well, the thing that won't come through here because this is a single document is,

Starting point is 00:10:23 is if I were to back up into Claude, so this is just literally one document, maybe a month in. If I back up into Claude now, like what you'll see is this is part of a project that has gotten to a somewhat insane level of types of. questions. So you mentioned one thing, which is how do we think about the space. But, you know, if I just scroll loosely through here, you can start to get a sense at the number of different things that come up in just trying to open up a small business. For instance, we were reclaiming redwood from a forest. How much redwood should we need? We were trying to write a letter of intent to a landlord. I have no idea. I have never written a letter of intent to a landlord to

Starting point is 00:11:09 Lisa space. How do we do that? I could keep going through these, but inside of here are things like timelines and to do lists, Berkeley permitting. That's a fun thing to try and figure out, right? A draft of a PowerPoint deck that we wanted to send to the landlord to look like we're a legit business. Again, something we could have done. We've obviously both built a lot of PowerPoint decks in our life. But this was for a landlord. Like, I can pitch for a VC or a founder or like, oh, you know, or trying to get a project done. I don't know what a landlord wants to see in Berkeley. Anyway, so a lot of it was this. And then the important thing, of course, is that as these contexts, these docks are building up,

Starting point is 00:11:46 we're just dropping them in. And you can start to see over here on the right-hand side, things like, this is the board game cafe overview I showed you a second ago, a competitive analysis of what the market looks like, pricing, figuring out pricing. And then we started to get into things like player personas. So like, okay, what kind of people are going to come to this place? And that, of course, leads into revenue calculations and all the rest of that stuff, which I can show you in a second. Yeah.

Starting point is 00:12:14 So it looks like your general workflow was use cloud projects, set up a project for the entire business. Just this is where everything goes and you have, I don't know, 100 plus chats in here. And then those chats were creating these documents and these sort of artifacts. And then you were loading those artifacts into the project as context for future chats. And it seems like you're solving a variety of problems. One is like real estate in Berkeley, SOS. Then it's, you know, just calculations, whether those are physical area calculations, financial calculations. And then communications, presentations, content sort of generation.

Starting point is 00:12:53 I wonder if you would show us one of the chats. And again, I'm like really attracted to this floor plan one because you and, you know, you, Andrew and I. You should pick a different one with floor plan one because it was bad. It was really bad at floor plan. I mean, maybe we would to see how bad it was at floor plan. That could be helpful. Or was there something that it was good at that, you know, you felt like would have taken you a really long time to figure out as a novice in this kind of business.

Starting point is 00:13:20 Yeah, you're right. You're right, Andrew. The floor plans, it's good to talk about what's good and bad. So as we tested Claude and in general lots of models, which we bounced back and forth between here and windsurf and cursor and every single AI product we could possibly use to make this thing, I would say you start to get good at, you know, finding the edges of it. So obviously, LLMs have no spatial memory.

Starting point is 00:13:42 So a lot of the floor pan things it did were pretty poor. But a lot of like the persona work, what kind of people would come to a place like this? How often would they come? It was really good at a lot of business plan things. It was very good at, ostensibly the kinds of things you can imagine people have written about on the internet and it can do research on. And so it's inside of its context window. So this is an example of a person. persona identification system that we worked on with Claude over the course of a couple of weeks.

Starting point is 00:14:09 And to explain it very quickly, like, this is the top axis, X to Y, is basically how gregarious you are about games. Are you devoted to playing only one game? I just want to play chess every single day. And on the far right is like, I want to try games all of the time. I want the new game every week and so on and so forth. And then down the Y axis is ostensibly how introverted versus extroverted you are. right? So the top right is somebody who literally just wants to play chess with their friends. Like, I don't want to play any other games. I don't want to do anything else. And in the bottom left is somebody who is, wants to really use this to meet new people. And so we ended up

Starting point is 00:14:50 with this kind of, you know, three by three matrix of all the different types of personas. And then Claude did a really good job of kind of dividing up and then saying, well, in general, what percentage of those people come to a business, like come to a business? Like, a business like this. And then importantly, what kind of events would you build that would handle the myriad of people that are going there? Because part of our experience, when we go to board game shops or card shops, is that we kind of feel like they're mostly tuned for a very specific kind of individual, which is like they're going to have, if you like playing chess or you like playing Magic the Gathering, then you come to Magic the Gathering night and you play with

Starting point is 00:15:28 other Magic the Gathering people. And it's very, and that's, I've done that. I've played Pokemon with my kids for years, like it's wonderful. But it wasn't the kind of full panellope of people we felt were now playing board games in this world where it started to become much more mainstream and much more of like a broad social activity. And so we're trying to cover for that. So the other thing that I see here, which I think is there's two things that I see are really useful. One is just coming up with a way to categorize your users or your customers by some axi on which you can prioritize some part of your product and your product is these events or ways to attract people in. You know, the other thing I think is really interesting is if I were to come to a business like this that I was unfamiliar

Starting point is 00:16:08 with having lived in the warm waters of enterprise software, I would say like, I know you need a building and I know I'm going to put games in it, but how do I turn this into a business? And what I see here is it's giving you frameworks for, well, you're going to need events and you're going to need private parties or you need open hours, you're going to need closed hours. And how much of how much did you brainstorm the actual business model and shape of the business? using AI or did you all have a pretty strong point of view of how it would look? Even before we signed the lease, we used AI quite a bit to figure out if this was even a business that could be made viable.

Starting point is 00:16:48 Like, I don't know if you can tell, we're not trying to turn this into a multi-billion dollar unicorn. But we are hoping that it's just not a money pit. Yeah. And we didn't know where to start in answering. those questions. So I think what was useful about the floor planning exercise that we did was like, how many people can we fit in here? And I would say that the way that this is gone for us, it's less that there's like particular like specific tactics or tips or tricks. It's more of like having

Starting point is 00:17:21 developed this mindset. Like I remember when Amazon was transitioning from being a bookseller to an everything store. And you could buy like paper towels on Amazon. And for you. And for you. years after they started selling paper towels on Amazon, I would still go to Costco to get my paper towels. So it's like I thought that was really cool, but I hadn't rewired my brain yet, right? And I think that's where a lot of us are with AI, where we're like, this is cool. I'm a convert, but I'm not rewired. And there's something that needs to happen where you remember that AI exists as you are contemplating a problem in front of you and you remember to engage it. And that's actually like the biggest thing. And as soon as you get there, a lot of this just takes care of itself.

Starting point is 00:18:10 This, like, right now, we're getting ready to do a vibe coding hackathon at Descript. And I've gotten pretty good at vibe coding the Descript app. And someone on my team was asking me, like, do you have, what are the tips and tricks? Can you like record something where you share something? And I couldn't think of anything. It's, it's not really that there's specific ways of whispering to the to the agents or whatever as much as it is just remembering that this stuff exists and having it front of mind. Yeah, I agree with you. I think so much of AI adoption is about changing muscle memory about where you go to solve problems. And it seems like from the beginning of establishing this business, you said the place that we go to build this

Starting point is 00:18:55 business is this Claude project. So the tool number one I'm going to reach for is this AI partner that we have. And I think that pays off dividends as you put more and more. It's going to be really interesting in two years or three years when you have all this context of not only the beginning of the business, but how it's performed over time, what challenges you've had. You know, that's the institutional memory of your business that can only make it better, I think, as a partner moving forward. Yeah, I think a lot of the early part of this project was really just getting in the habit of almost like reminding each other as we just ask some dumb question. Like have you asked Claude yet? Like and it's, you know, we both live in AI every day in our day jobs.

Starting point is 00:19:42 And yet you'd still find that you needed to get reminded. Like it was, it was amazing to me how quickly you had to, you know, I would go to look up information very obviously thinking about it as Google and saying like, oh, how, you know, what are some local board game shops? That's an easy thing that we have this Google kind of search mentality. But treating it as a muse instead of as an Oracle was kind of the shift. Every time I'm going to go think about the thing, I should go drop into this area and use this as the partner to think about this thing. That was the big shift. I'll just say I'm conscious as I hear Nabil and I talk about this, that it might sound like we're just like outsourcing thought to this AI and generating like the retail. equivalent of slop. But it's, it really like couldn't be further from the case.

Starting point is 00:20:33 We were just talking about this the other day that it feels like that energy, that creative flow that you get in the best parts of creating something new. It's like you're just mainlining that because your feedback cycle is so fast. You can have this thing go out and do a bunch of grunt work. And then we're sitting around in kind of creative ideation. analyzing what it's come back with and building off of it. It's really fun. Speaking of grunt work, for folks that are watching that are not in the Bay Area, the thing that I think is most brave about this is you all went into Bay Area commercial real estate,

Starting point is 00:21:13 which is a place that totally terrifies me. And I'm curious if you could just give us a few minutes on how you navigated this real estate transaction with confidence. And I would call it a complicated real estate market and how you got to confidence you were getting the legal, the regulatory, the local advice that was accurate, you know, using some of these tools. Well, let's keep in mind that there were still lawyers involved. They're always still lawyers involved. At some point, you know, there's there's real estate agents, there's lawyers. But what was really helpful, obviously, is you now hear all these anecdotes of people who, for instance, are still going to a doctor but are taking their health diagnosis

Starting point is 00:21:55 and dropping it into LLMs to get a second opinion. I would say the cycle felt a lot like that. I think it was about building a level of confidence that felt like you could lean into the experience because you weren't going to get fleeced. And so we still relied on experts all the way through. We still, but I would just sit at night and say, like, I don't even know.

Starting point is 00:22:16 We're trying to describe a concept for an experience of this membership club with retail in the front and we're going to do special events in the back. And what is the best? permit for this in Berkeley that won't take like two years for me to get done. And that's where you get a little bit of confidence as you're navigating through things. Andrew would come back and he'd be looking through Berkeley weird like local ordinance websites and copying the pages and dropping them and clawed and getting feedback on them and things like that. And so it really wasn't, it wasn't, again,

Starting point is 00:22:47 like AI was doing all the work. It was just that AI, in that instance, I would say AI was giving us the confidence, whereas maybe five years ago would you've been like, like, you would have had the, I would have had this fatal expression you had, which is just like, this seems scary, it seems Byzantine, like the inertia can't even get started. And so I'm just going to go do the other thing. And this will still just be a thing I complain about. And when I fast forward that, it's just the thing that makes me excited is not just that this exists. It's just like, literally this wouldn't have existed without AI. Two, as Andrew mentioned, there's so much grunt work in getting something like this open, that, again, it couldn't have happened as a side gig.

Starting point is 00:23:26 It just couldn't have it. It would have had to have been a full-time job. And then there will be this new, interesting, hopefully break-even place that people actually like in the local community. And if you fast forward that, like, obviously we are in AI and we have some coding backgrounds and we're very highly technical, we're dropping into things like cursor and windurf, which is not what the average local business person is doing today. But if you fast forward three or four years, you can imagine the multiplicative effect as these tools get broader and broader and broader.

Starting point is 00:23:56 And for communities, that means you get interesting businesses and places to go and new people to meet and ways to connect. That could not otherwise exist. So I'm very much an optimist that the more we can build, the more amazing things we're all going to get to experience. This episode is brought to you by Persona, the B2B Identity Platform helping product, fraud, and trust and safety teams. teams protect what they're building in an AI-first world. In 2024, bot traffic officially surpassed human activity online. And with AI agents projected to drive nearly 90% of all traffic by the end of the decade, it's clear that most of the internet won't be human for much longer.

Starting point is 00:24:37 That's why trust and safety matters more than ever. Whether you're building a next-gen AI product or launching a new digital platform, Persona helps ensure it's real humans, not bots or bad actors, accessing your tools. With Persona's building blocks, you can verify users, fight fraud, and meet compliance requirements, all through identity flows tailored to your product and risk needs. You may have already seen Persona in action if you've verified your LinkedIn profile or signed up for an Etsy account. It powers identity for the internet's most trusted platforms, and now it can power yours too. Visit with persona.com slash how IAI to learn more.

Starting point is 00:25:18 You know, you mentioned Cursor. I just can't get out of a how I AI podcast without somebody talking about vibe coding. So, Andrew, I'm going to kick it to you to talk about how you actually built an AI concierge for this gaming place. And what is it called? I forget. We did mention it. Tabletop Library. Tabletop Library.

Starting point is 00:25:42 In Berkeley. In Berkeley. In Berkeley. We forgot to promote it at the top of the show. But how did you build this AI concierge so people could actually access this and have a great experience? Yeah. So one of the ideas that I was quite excited about with this was something that could help organize on-demand game sessions with other cool people that took away a lot of like the social weirdness of setting these things up and the social risk in the way that, you know, nerds in Silicon Valley are prone to worry about. And so the mechanism was going to be that you would

Starting point is 00:26:26 have, you would have this way of saying, I want to play a game at some specific time, see if you can find a group of people, form a quorum, and then if you can form that quorum, you can reveal us to each other, and we'll book it. Otherwise, nothing actually happens. And we started out by, thinking at the very beginning, like, okay, we'll just find a SaaS software package that gets as close to this as we possibly can. Let's look at like table management stuff or other brick and mortar reservation systems. Then we started looking at like co-working space management systems. So there was like the just standard book a table straightforward with your friends. And so the idea was that you would be able to use this software to either just book a table with your friends or try to do what we're calling a LFG looking for gamers mechanism where you could say, like, go out and find this group for me and then book a table.

Starting point is 00:27:30 And we were playing with all this SaaS software and it's like it often is where nothing's quite right. Like you want something so much more custom. And then at some point we realized like all we're talking about building here, is a relatively simple set of tables, a relatively simple database, and then you could just have like a chatbot living on top of that. Why don't we just build it that way? Getting back to that mindset thing, it wasn't at all obvious to us when we started this. We spent a month going down this. In retrospect, I'll show it to you, it seems like a no-brainer, but it wasn't obvious to us at the time. So I'm going to give you a quick demo of what we built, and then I'm going to show you

Starting point is 00:28:15 a little bit of how it works, because the one thing about it is we're still building it, and it's kind of slow. So let's say I want to organize a game this weekend, and I can go like, hey, can you set up a game for me any time this weekend, maybe a deck building game. So now that goes off to the agent, and I'll show you how we built all that now, because it takes a second. and then we'll come back and see its actual reply. There's two parts to this system. One is the database itself, which we built in Airtable. And the other is this workflow from a company called N8N.

Starting point is 00:28:54 And let me start by walking you through the basic things of the database. So there's a table where we have all of our users. What's really interesting here is things like availability. So right, it's going to eventually go and try, to recruit other members. Normally in a database, you would store that as a bunch of related tables, right? But here you can just have free text. Free text, this is where you put your gaming preferences in.

Starting point is 00:29:22 So it's just like a free text list of the games that the agent will update as stuff goes. So structuring these databases becomes so much simpler. Then we have like the tables that are available in the, in the, in the, in the, in the, in the, in the shop, we have games that we stock, and then we have actually the reservations once they get made. I would like to pause here and just call out you're using AirTable instead of another kind of, I'm sure there's lots of databases that you could have done on the back end if you wanted to. Could you just tell us why Airtable and why you chose to do something like this for this particular project? Yeah. So what do you think should have been in the consideration set before I

Starting point is 00:30:07 answer that. Oh, I was just wondering, I mean, I would have just spun up a Postgres database. And, you know, if you have a cursor, I would have just gotten straight to just an empty database and a schema. And that's where my developer mind went. So I'm curious why you started with a tool like air table. I would not have gone there. I would have been more like should we be using Notion or some other kind of something with a heavy GUI? Because the nice thing about something like this is we can also build. human interfaces for interacting with this directly. So if we have somebody that wants to call up and book a table or something like that, we can build an interface for somebody to just go directly into it or manage events or something. So it just makes it much easier. And it means that people who are going to be, you don't have to be an engineer or build a, build a front end for every command that you want to do. It's a good, it's a good prompting though, Claire, because this is, again, this was probably four months ago. It's very possible that four months from now, The right answer is you can one shot a web interface to all of this, and then there's no, no, that it looks exactly the way you want it to look like.

Starting point is 00:31:14 In this world, Airtable, one, I don't have that confidence in Windsorff, certainly not four months ago, as we're getting to Claude 4 and beyond maybe so. And second of all, this just has all of the sorting mechanisms and, like, I'm filtering mechanisms and a million other things that come along with the human use of typing in 100 games or doing other work in those workflows. It's just helpful. Well, and one call out for folks out there thinking about building something. like this is in past lives, I would have thought, oh, my God, I never, I don't want to work with the Airtable API. I don't want to, but so many of these things come preintegrated now with these workflow builders. So much of what you use with these AI tools can figure out the integration points for you. So you can pick the thing that's easiest for you to use. And when you need to

Starting point is 00:31:58 integrate it, I think it's very accessible. So it makes, it makes sense to me for for a business like this. I was just curious if that was your. I mean, a ridiculous amount of these cells, to be clear, came from that Claude project we talked about earlier, and then were MCP connected into Airtable, and it was just injecting data directly in. Can you talk about that flow really quickly? So you told us how you sort of set up these tables for how many physical tables you have, what games you have, who is interested in what kinds of games. How did you actually populate, populate this data through MCP? I'm really curious about that. Yeah, let me give me an example of that one. So, like, this is, we have a small retail section in the front, and we had this,

Starting point is 00:32:37 idea of like what if instead of just putting things in alphabetical order or by like a traditional category we merchandised and curated these these categories you can see some of them here like brain burners push or a silent strategists almost like the way your favorite like independent bookstore merchandises their books in the front right yeah and but there's this like puzzle of how do you how do you pick the games and how do you make them mutually exclusive so that you're not merchandising the same game in multiple categories. And that is the kind of thing that probably people have the idea for all the time when they're opening something.

Starting point is 00:33:18 But it just ends up on the cutting room floor because of the amount of time it would take to do it. Or you do it at the beginning. And then you're like, it's just not worth it. And you cut it. And it just makes it trivial easier to do it. So it's like games about cats, for example. You can do something like that. You just give a brief description of what your category is, and then it'll go off and come up with the games automatically.

Starting point is 00:33:43 So it's nice in Airtable that they have these like built-in AI fields and stuff like that. Got it. So you're using built-in AI categorization by Airtable to create a category and pull in relevant that looks like records or games that might fit. Right. So that's a way that we use the existing tools. Nabil, what were some of the things that you? you populated with MCP? Is there something, a tab I should switch to here? Well, probably the easiest one and the most wonderfully nerdy is if you hit games on that tab, and then go click to in library. So one of the things we wanted to do, given this is called Tabletop Library, is to think about this place. If you're starting your board game navigation, and maybe if you've ever been inside of a board game cafe, you've had this feeling, you walk in

Starting point is 00:34:29 and you've maybe played Catan and maybe your friends had you play Wingspan. And now there's just like a wall of games and you have absolutely no idea where to start. Well, this has been solved. In libraries, you have a card catalog library system where all of the French history books are next to each other and so and so forth. And so this is a kind of thing that would be like literally impossible without AI, but we basically have a bunch of design prompts in Claude back in those projects that help to categorize into kind of a dewy decimal system of board games, all of the games in existence, and then looks at each game and applies them. So if you look at here, this is a list of games that are in there, And if you look at the bottom, you'll see the TLCS code, like 420.5 on this game Nemesis lockdown.

Starting point is 00:35:11 Right. And that means it is a category name in cooperative. That's the 400 section is cooperative. The 200, the 20 section, 420 is adventure co-op games in adventure. And then the dot 5 is the weight of the game, one being the lightest weight and 5 being the heaviest weight, as in complexity of the game. And so now you can imagine, if you're interested in, playing an adventure co-op game with your friends, you walk over to the 420 section and they're all nicely grouped together. And then those colors, those yellows and reds are how hard that game is to

Starting point is 00:35:45 play and so on and so forth. Impossible, literally impossible without AI. We are really taking people on an adventure here. It's truly the most ridiculous idea that if he had proposed this and we didn't have AI, like I would have gone to the mat, like, this is such a crazy thing that nobody's going to want and is going to be such a waste of time. But because we do, we can indulge in this kind of stuff and it's really fun. And maybe, you know, for every, there's five of these, one of them will be the best thing ever. And maybe this is it. Okay. So you've shown us, you've shown us your air table. It's hooked up to this workflow, kind of like a gentic system on the back end. What does that expose to

Starting point is 00:36:30 end users? So let me show you this now, right? So it looks like it already set this up when I tried this demo earlier and it found that. So it's already set up a game for me this weekend with Slay the Spire, Great Dick Building game. I've scheduled it for this weekend and created a request.

Starting point is 00:36:47 And so now what's happened is it's gone out on its own and it's messaging other people who like these kinds of games and it's trying to build enough people. to schedule a game. And this whole flow we're using an 8N4. So basically what happens is you send a text to this phone number. We hooked up Twilio. It comes in and just does a little bit of

Starting point is 00:37:12 basic formatting. Then it looks up my phone number in the air table, members table here, and it finds out who I am. And it figures out if I'm a user or not, if I'm not a user, it has a whole flow, and this is the agent. You can see there's these long prompts that basically do all of the routing that we've learned that says what to do if it's not a user. I'm not going to focus on that one. I'll focus on the if they are a user. So if they are a user, we have this central agent, and these are all the tools that it has

Starting point is 00:37:45 access to. So these are all Airtable MCP tools. So you have something that can just, like, get a list of records. something that can create a new record and the instructions for how and when to use that are all in this big prompt that we've put together which you see gets quite long and it's just been like iterated on over time and it's crazy because like you'll see there's parts in here where it's just saying you know use the use the create record tool to to do X, Y, or Z, and you just reference it in plain text, and it knows how to go and find that. Yeah, you're not calling a function or anything like that. You're just literally writing down what you think it should go do, and then it'll figure out how to find the function and use it. And so this, this agent, available to your customers via text, so super simple interface, but is routing

Starting point is 00:38:48 to your source of truth data for your business. And then it's engaging with other customers saying, I got somebody who wants to play this game this weekend for an hour, an hour and a half. If you're in, let us know and schedule it. And it brings those people together. And as somebody who has put together children's birthday parties, just even that flow seems very, very useful. And I can imagine. We've got to turn this into a startup.

Starting point is 00:39:16 Third startup. That's when you've joined my club is when you're doing three. But what I imagine as a business owner, though. Like, it's cool as a customer. That's nice. It's nice experience. It's fun as a builder. But as a business owner, you are probably then getting much closer to, you know, your gentle, break-even goal of higher utilization, more bookings, more people with lower sort of person-to-person kind of grunt work that has to happen to make that happen. I'll just say, like, I guess so. We never thought of it that way, although maybe we should have. But it's just the nature of this enterprise. that that's not the calculus that's typically going through things. I will say, though, that it's not about replacing a human job. There's no, like, a human couldn't do this type of work at this level where you're going out

Starting point is 00:40:10 and contacting three members to see if they can play the game. And when one of them says, no, you contact another. It's like maybe you could do something like that. but for organizing a board game night, it just doesn't make sense. It would be too slow. So this is just, like Nabil was saying earlier, an example of the kind of experience

Starting point is 00:40:35 that just couldn't otherwise exist. Yeah, I would say a lot of our experiences started with a situation where if you had a, in that first group of things, it's, you know, the business advisor, the co-pilot, which you're right, if we had people that had done this work before and you surrounded yourselves with a bunch of advisors,

Starting point is 00:40:54 those are human tasks that get you a couple of business plans and we built up Python capacity planning tool and a bunch of other stuff that, like way outside of even the scope of what we can talk about here that we just built. The second phase of this is it's just enabling experiences that you just never would have done with humans, period.

Starting point is 00:41:12 It's not replacing anything. It's we just wouldn't have offered these services. You would not have categorized by hand 200 plus 300 plus board games into a category system, you would not have set up a bespoke way that I can natural language, say a board game category, not even a game, over text, and say, like, I want to play an adventure game this weekend. Who likes adventure games that can play between six and eight o'clock and then have that just like manifest itself into a space and then a table get booked?

Starting point is 00:41:41 Very unlikely any of these products would even exist. And frankly, the whole project wouldn't have existed without AI in the loop. And then before we wrap on this section, I just have to call out, looking at everything you've shown, it doesn't seem like you wrote much code to make this happen either. It seems like you wrote a lot of prompts and you hooked a lot of tools together. But I'm not seeing you pop open cursor in a repo. This all looks like kind of no code stuff that folks could use in their day to day, even if you're not a software engineer. There's a little bit of code in the N8N stuff. So I think with N8N, you still need to be a little bit technical. with the way that things exist now, but that's just like, you know, a matter of time

Starting point is 00:42:24 before this becomes more and more something that anyone can do. We all, we have repos for like the website that we built and Nabil built a whole like ratings system for members

Starting point is 00:42:40 to rate their games and get recommendations and stuff like that. So there's that too. It's just, we figured that was a little more boring. Yeah, we also have a Python application where we took,

Starting point is 00:42:48 those personas that we talked about earlier. We fed them with an event programming and then a pricing plan and then that goes into a Python application that uses Pulp and CBC to basically like build out a like integer linear programming application to figure out our capacity planning to basically figure out whether we can break even dynamic revenue business model planning. So you can set up plans, map it to personas and then it just says like, oh, if this is the utilization rate of this place, it'll max it 250 members and then we can loop that back in. So there is some heavy programming that goes on at some portions of this.

Starting point is 00:43:28 But Claire, the vast majority of it starts from prompts and artifacts and projects and then document creation and that loop back and forth between producing a new document, adding it into the context window of like, this is a new canonical truth of this project, and then iterating and brainstorming from there. Okay, let's wrap. This has been so fun with lightning questions. One no AI and one AI question. So the non-AI question is, what's your favorite game at Tabletop Library?

Starting point is 00:43:57 The next one is my answer. Like, part of what I love here, and I know that's not a real answer, Claire, but part of what I love here is I actually really love exploring games. I think of it like a movie opening, and we have a culture at one point where you'd really care about the movie opening and you'd think about it. You think about the director releasing their next thing. And that's kind of where I've gotten to with games. And hopefully others can get that excited too. Andrew, do you have a favorite? Nabil is such a big shot in the game world that he couldn't actually name one.

Starting point is 00:44:26 He just alienate too many stakeholders. I like this game called Sky Team. It's a two-player game where you play as the co-pilots landing a plane. And it's really fun to play with my wife or my kids. So that's both a recommendation and an answer to your question. Okay. And I saw a little of this in your prompt. So Andrew, I'm going to ask you my rap question first, which is when AI is not listening, how to get to listen.

Starting point is 00:44:57 And I will say I saw a lot of like all caps, caution, exclamation, exclamation in your prompt. So do you have any tricks that you fall back on when you want AI to really listen to you? I saw a really good video that I think Y Combinator put out recently. And they were like, when you get one change that I did as a result of that is when you start getting stuck in a rat hole and you can't get it to do the thing that you wanted instead of just like keep on hammering at it, just revert back and try something different. Nabil, do you have a trick? You have a fallback. I think I've certainly gone through all of the stages of grief and anger and all caps to try and get the model to do what I want. My more recent one is to just try and add context.

Starting point is 00:45:41 I think my most common thing now is to try and throw tokens at the problem. So a lot of times what I will do is copy the problem I have, throw it into another model, and then talk back and forth for 30 seconds to try and stretch out and elucidate the problem and add a lot more detail and then copy that prompt back either as it. It might be at that point as big as a PRD or something really detailed or so and so forth, but just basically like it's confused and I need to give it a lot more direction. And so let me then partner with a different AI to figure out how to tell what's really supposed to be doing. Okay. This has been such an adventure. I can't wait to bring my kids over and see what other

Starting point is 00:46:17 cooler games we should get into. Nabil, where can we find you and who should reach out to you? Any founder building a wonderful AI product should reach out. That's where I spend all my time nowadays. You can reach me, Nabil and A, E-E-E-L on Twitter, wherever it's called nowadays or my first name at Spark Capital.com. Okay, Andrew, and what about you? Where can we find you and how can we help you? First of all, if you happen to live in Berkeley or North Oakland, you should sign up for this thing or check it out. It's go to tabletoplibrary.com and sign up for... Thanks for doing the business job, Andrew. That's very much appreciated, sir.

Starting point is 00:46:56 But also I have to say, yeah, I'm the CEO of Descript. And the way this whole thing started was my head of marketing was savilely like, oh, there's this great new AI show. You should go on it. And I, like, sabotaging my press, like normally, like I normally do was just like, I'll go on it. But I want to talk about this other thing that I'm doing that I think will be more interesting. But Descript is doing a lot of cool AI stuff too. We just launched the first video editing agent. So like cursor for video.

Starting point is 00:47:30 And I'll say, like, here's my like founder tip. This thing, this whole side project has been like the best. thing in the world for rewiring my brain around AI. When you have your ongoing gig, you're kind of stuck in your existing patterns and people you have to rely on starting something from scratch that you actually care about where you couldn't do it without AI. It's the best. I completely agree. I have done the exact same thing. It has been transformative in how I lead in a bigger company. And I will do your marketing person a favor. And no, this was not paid, but we happily use Descript to edit this podcast.

Starting point is 00:48:11 So we have been happy, and it's how we get the great quality that you're seeing today. So Andrew Nebiel, this has been so fun. Thank you for sharing all your tips. Good luck with the many businesses. And have a great day. Thank you. Thanks so much. Thanks so much for watching.

Starting point is 00:48:27 If you enjoyed this show, please like and subscribe here on YouTube or even better, leave us a comment with your thoughts. You can also find this podcast on Apple Podcasts, Spotify, or your favorite podcast app. Please consider leaving us a rating and review, which will help others find the show. You can see all our episodes and learn more about the show at how-I-A-I-I-I-I-I-Pod.com.

Starting point is 00:48:51 See you next time.

How I AI - How a VC and tech founder used AI to launch a brick-and-mortar business in their spare time | Andrew Mason (CEO of Descript) & Nabeel Hyatt (Partner at Spark Capital)

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.