How AI Can Supercharge Your MVP Development | Aram Melkoumov

So leveraging AI to speed up how you build and test your MVP. But remember that a clear problem statement and understanding of your user is irreplaceable. So there's always tools out there like Idea Browser, ChatGPT to find ideas. Now you could definitely use these tools, but I'm always a big fan of people who actually try to scratch their own itch. So if you have a certain frustration or pain, it could be a small thing, but if you have it, there's a very high chance that other people would have it. So maybe try to focus on your first MVP that you want to build out towards something around that, okay? Don't get like, don't get overwhelmed with all these different ideas everybody has, everybody's talking about. Just focus on what you think matters to you because if you align closely with the problem, that would help a lot in terms of like your desire to actually solve for it.

Hey everyone, how's it going? My name's Ian Isaacs. I'm the head of growth and partnerships at the Founder Institute. Thank you so much for being here today. Today, really excited to have everyone. We have a really, really interesting event planned out. We have yet again Aram Melkumov from Crowdlinker. He's participated in a number of different events over at FI. He's come back to talk about all things MVPs. I think this is his 5th event with us. Awesome mentor, awesome person to chat with. Today he's doing something a little bit different. He's going to go through an MVP build using AI and some specific tools. Hopefully this workshop is super valuable for you folks. I'm really looking forward to it. We have about an hour here, so I'm going to kick it over to Ram shortly, but just want to go over some very basic housekeeping. First off, as you can see, this platform is called Airmeet. If you're using a mobile phone, not the best experience, try to log in through a laptop or a tablet so at least you can see everything. On the right-hand side, there's a chat box with the, with the team and everyone. Just like, definitely put it in where you're messaging in from today. This is a live event. I see Montreal, Karachi, Abu Dhabi, Portugal, Israel, Switzerland, um, Colombia, Morocco, DC, Richmond, Luxembourg, Amsterdam, Sweden, a lot of cool places, um, Romania, Brazil. So We had about 1,100 people sign up for this event. I don't know how many people, I'm expecting around 250, we're already at 195, so I'm pretty sure we're going to hit that number. Really looking forward to having everyone. The last thing here on the left-hand side of the screen, you'll also see a Q&A button. Use that to feed us questions, we'll use that towards the end if there's time, we'll take care of that. Then right towards the end, we want to leave about 30 minutes for networking. There'll be a lounge that we call the networking lounge. You know, feel free to jump on any table. Aram, if you have time, you can jump on as well. We didn't have a table in your name because we didn't know if you're going to stick around since it's so late on your side. But yeah, if you want to, feel free to stick around. All right, I think we're good. Yeah, and last but not least, at the bottom there's emojis. So, um, if you think whatever Aram is saying is relevant, thumbs up, wow, heart emoji, whatever you think is, uh, it's the most relevant. Okay, cool. Maybe I will kick it over to you, Aram. What I'm going to do is I'm also going to publish this poll and see if I can put this on stage. Why can't I put this on stage? Okay, weird. All right, never mind. On the left-hand side of the screen, you'll also see a poll button.

I put it up there.

Usually I can put this on stage, but It's fine. Maybe around while I try to figure this out, do you wanna give an intro to yourself?

Cool.

Yeah. Hi everybody. Nice to meet you all. I'm very excited to be here again. As Ian said, this is my 4th or 5th time doing a webinar with Founder Institute. So I'm calling in from Barcelona in Spain. I'm the CEO and co-founder of a company called Crowlinkr. We've been around for 12 years. We're one of the leading digital product studios in Canada and we have shipped close to 100 different products to market. And this event is very timely just because AI is talk of the town these days and more and more tooling and workflows are being created that helps founders to be able to ship faster and smarter. So that's what I'm talking about today.

Cool. I can't figure out the poll thing. We still We have 6 votes in. So I guess the question here, folks, is maybe let's try to do this a bit differently. If you've used AI tools in the past or you haven't, can you just drop it in the general chat on the left? So like, you know, just let us know if you use any AI tools, which tools you've used and stuff. You know, hopefully this gives you some indication of how people have been using. Okay, I see some people using Replit, Cursor, Lovable, and Groq, and all this other stuff. Okay, cool. So I think we have a decently well-placed audience, but, you know, I'll kick it over to you to run this now.

All right, cool, cool, awesome.

Yeah, um, you put, um, is my presentation on stage now? I can't tell.

I will put it back, my bad.

Awesome.

Cool. So I'm thinking everybody can see my screen now. Okay, cool. So there's lots to unpack in today's webinar, so I'm going to try to breeze through some of like the educational awareness type of slides first, and then I'm going to get into my vibe coding session that I want to do with you guys today and tell you what my workflow is. And hopefully you guys find it beneficial and then we'll end off with some Q&A. So I'm gonna, I'm gonna show you how a one-person team can ship and scale an MVP. I'm going to be trying to get to the point of showing you like a functioning prototype, which then I'll share into the chat and people can play around with after, just to kind of see how, how quickly we could get this out. So stick around, I'll be tackling the live demo towards the second half, okay? So the goal of today's agenda, my hope is that you're gonna leave this webinar learning how to ship something this week with your idea, your business, whatever it is. All right, just a quick refresher. I like always starting off with this 'cause I've done a prior one or two webinars on MVPs, but I like always to bring it up here. So MVP, is the smallest product that still delivers some real value. Its job is learning, it's not perfection. So whatever it is that you're afraid of shipping, which is embarrassingly bad, that's okay. You know, that's how a lot of MVPs start off with in the market. So as many of all know, and I see many of you guys all using, you know, different types of tools, so that's great. I love seeing that. But ever since AI really came into the fold, you know, it started off taking some workflows into consideration around how to build product and helping developers and founders be able to ship faster and smarter. So quick fact, GenAI tools cut dev time now between 30 to 60% and routinely code grinding by 20 to 40%. So on our end, with a lot of the products that we are building, we're saving about 30% on everything that we develop. That means more experiments per dollar and a shorter road to product market fit. And I did a prior webinar on product market fit, so go and check it out. It's on YouTube. So now the key thing is thinking days to prototype instead of months. So times have really changed and this gives solo founders real superpowers in this new kind of agentic world. I want to talk about something called Lean AI Startups. Something new and there's an influencer space who's really talking about a lot. His name is Henry Shi. But basically a lot of founders right now who are building out lean AI startups are starting to hit $1 million revenue per employee and they're building out their design, they're designing out their org and stack for that efficiency. So I'm going to show you this leaderboard. Henry Shi was the co-founder of a company called super.com. He grew that company up to $150 million ARR with very lean principles, and he verifies each of the companies on this list, which I'll show you by looking at their Stripe data in order to confirm like the verification of what people are claiming when they submit their companies. And there's many, many 10-person teams on there who are doing $10 million ARR. and some are solo founders doing $1 million a year. Lovable, which I think many of you guys are using, is a good example of one of the companies on that list. They hit $10 million AR in just 2 months with 15 people. So the target mindset here is how far can you go with fewer humans and smarter tools? So here is the leaderboard. You guys want You'll get this out later, but I'll post it in the chat here so you can all check it out. If you look at this board, you see a couple of things here, like total revenue of the companies that are on this leaderboard. Average revenue per employee is $2 million and average team size is 20. So you could see some pretty interesting ones here. I'm just going to focus on a couple. There's like Bolt here, there's Lovable on here. You could see how much revenue they have. How many employees they have, you know, and whether they're profitable or not. So that's pretty impressive. There's one I wanted to show you, which I think was very interesting, is this company called Cal AI. It's basically a calorie tracker being powered by AI. Their annual revenue is $12 million a year, and they have 4 employees, so they have a revenue per employee of $3 million, and they're profitable, and they had zero funding.

So there you go, it goes to show what's happening now in the space.

There's a link here to a Forbes article if you want to check it out. It's basically a 17-year-old kid who made this company. It's an AI-powered nutrition app, and they're going against the likes of, um, uh, MyFitnessPal, uh, Lose It, and many others. And they've done very, very well. And here's the And here's the website. So this was all built using different kind of AI tools. So food for thought for everybody saying or complaining maybe that things can't be done. This is kind of proof. So check out the leaderboard. I recommend going through some of the companies there, doing a deeper dive into them and seeing what, what they did, what, what they did online, how they got some traction and things like that. Some of these companies are actually building in public. All right, so the first step with any type of build of a company, you got to start with an idea. So I'm going to take Greg Eisenberg here. I don't know if you guys know him. I highly recommend following him. He's a pretty big influencer online. He's built a few successful companies, but he actually launched something recently called Idea Browser. It basically delivers one AI research startup idea a day for free. You could sign up and get access to the whole database. I think it's like $200 a year or month, I can't remember. But it's basically a great spark if you're blank paging and you have no idea what to do. And they use various data signals, which I'll show you in terms of the different sources that they use. So they scrape and pull data from things like Reddit, Facebook, and many other kind of platforms. If you're still stuck there, obviously use ChatGPT. If you're not on a paid version like Plus, I recommend it because of their O3 model. It's game-changing. It's gotten to the point that O3 has basically replaced BAs, business analysts at my firm. So that's one less headcount for me that I have to kind of think about because I'm using, and the team and I are using O3 model to do a lot of stuff. So anyway, you can use ChatGPT for your idea. Machine generation, you could ask it like, you know, in this industry, what are common pain points and things like that, and use an iterative kind of loop approach to kind of get to that point. Obviously, you could also look at Reddit forums, look for things like searching within Reddit, like, I wish there was an app for. So in that situation, look for upvotes, that basically shows you a demand signal. And Keep in mind, Idea Browser already scrapes a lot of this data and feeds it into their own system. So you kind of benefit from that. Google Keyword Planner, you know, look for searches with poor results, which have instant pain gaps that you could take a look at. Obviously, follow trends and validate fast. I put together a link here of, where is it, 50 tips. So I put this together for you. I'll share it with everybody, uh, once I share this deck, but it's basically 15 different tips for you to go and find your startup idea. All right, um, moving on. Okay, so once you have an idea, um, what I like to do is, uh, use ChatGPT's O3 model to basically get it to conduct deep market research. So there's a function there called enable Deep Research. It basically takes anywhere from 10 to 20 minutes at times to go and actually do a full crawl of whatever it can find online. And we're going to actually try to do that today and see how far we get. But basically, it, it learns on what's happening, you know, through different sources online. Once you kind of get that data that it extracts, Then you could obviously fine-tune, you know, the results, but that would help ChatGPT understand a very clear problem understanding of an opportunity and the market. So once you have that, then you need to prompt it to then focus on how to solve that problem. And what you're going to be asking for it is for it to create a product requirements document or a functional requirements document. Now you have to be very precise here. You have to kind of tell it that you're going to be looking for complete user stories, including acceptance criteria for all the functionality that it is that you want to build. That's very important. And then you fine-tune the PRD until you're happy with like how it's broken down. And so I'll show you some examples of that when we get to that. And then you're going to move into the UI generation. But it's just a very important thing is that when you're using ChatGPT, make sure you provide a clear role, goal, and context in order for it to produce clear formatted user stories, which will then go into the UI generation tools. All right, next. So fast prototyping tools. So a lot of people, I think, are using different ones. I'm a big fan for prototyping of using vo.dev and lovable.dev. So the reason why I like VO the most is if you look at the company that created it, it was called, it's called Vercel, and they actually host and run a lot of front-end applications for, I think, hundreds of thousands, if not a million businesses online. So they had very good access to lots of data to train their model on in order to create the best solution to generate your front-end prototype interfaces for. So what we do then is we would describe what we want to do in plain English. So gone are the days where you have to create these fancy 5,000-word magic prompts. Now with a lot of these tools, you could just speak in plain English, or better yet, what we're going to do is upload the PRD that you got from ChatGPT as a PDF and let vo.dev then instantly generate that production-ready frontend interface for you. So the good thing is that it's actually pretty decent code that you could actually build off of. Now, Lovable takes things a step further where it generates the entire application. So frontend UI, backend logic, database, including Supabase Auth, Stripe Pay, OpenAI endpoints, all out of the box. On our end, the reason why we prefer V0 over Lovable is that V0 doesn't lock you up into a no-code box. If you look at tools like Replit, Bolt, and Lovable, and some of the other ones out there, they're basically trying to lock you in to using their specific solution to build everything and deploy it because that's how they kind of lock you in. If you're a non-technical founder, definitely look at using Lovable or Bolt or Replit as one of those other ones. But on our end, our workflow is a bit different because what we do is or our ideal workflow is that we create a two-way sync between a two-week— sorry, a two-way Git sync with our view and your IDE, which is where the developers work in. So we use Cursor, Windsurf. We're using Windsurf now more. And now the beauty about this is that the code stays synced in GitHub. So when you make changes in your IDE via prompts or changes that you want to do, that then gets pushed and gets reflected into your Vue instance, which is where you then can continuously then optimize your frontend experience. So yeah, this workflow basically allows you to ship things very, very quickly. All right, the coding side. I'm not gonna spend a crazy amount of time on this. Basically, in this specific time that I have with you guys today, I'm going to try to get to the point where I can create a functioning, working prototype. But we use a few different tools. So Cursor is basically a VS Code with superpowers. It has inline chat, refactor, testing capability. You can prompt it to refactor this or write tests, and it kind of does it. Windsurf takes it a bit step further. It's kind of the first agentic IDE. So it could do multi-file auto-fixes by you working and, you know, collaborating with that AI agent. And then Cloud Code, which is pretty new, then excels at being able to compile, test, self-heal loops. So these agents that you can interface with can now fix TypeScript tests alone. Now, together between all these different tools, our dev team saves I want to say, you know, 10 to 15 hours off every sprint. If you really want to level up your game when it comes to user experience, because obviously the more data that you have or the more references that you could get access to, the better. But if you really want to kind of upscale your design creation process, then I would recommend this flow. It, it might be slightly unethical, but it's a great way to avoid costly reinventing the wheel aspects on design. Okay, so I don't know if you guys have ever heard of something called Mobbin, but it's basically one of the largest databases out there online, which has captured every single user flow, onboarding experience, buyer journey for the most successful and well-known apps that are on the market. So what I would do is I would go on Mobbin, I would scrape or mine very proven flows from top apps. So for example, like Uber's booking or onboarding flow, I would go in, I would capture that. Then there is a one-click copy to Figma via plugin. And then now you have those designs, you know, in Figma as editable wireframes. The next step is then using Windsurf Cascade where you can connect via Figma token and then prompt within Windsurf to rebuild the entire flow of what you have of your ridesharing, say, MVP using these new designs that it has access to. And Windsurf outputs like React Native code with routing and forms. So the result of this specific cheat code, so to speak, is that you get world-class UX patterns without hiring any UI team. And it can just adapt the visuals and copy for your niche. So I'm going to talk about our business workflow. So we always start with the problem, not the solution. So think of AI as being the engine, not the destination. Then you laser focus the scope. So really focus on that one core job that your product want to do and make sure you're providing that value and build only that first. Obviously build with AI now that we have these tools, but you measure the success of whatever it is doing with proper user feedback. So continuous, you know, user feedback cycles are very important here and then learn from it and adjust and, you know, go back into this kind of like cycle loop that you're constantly iterating on your product. And then lastly, always have a human in the loop. It's important you use AI for speed, generate that first draft, get suggestions, but use you or your team's expertise to basically refine and make final decisions. Don't just give all the keys to the castle to AI to go figure out for you. It's not there yet, let me put it that way. Maybe soon, but for the time being, just, you know, keep your, keep your hands on the wheel. And then basically repeat until traction. And then once you get that traction, then scale. Now, you could try to scale with a lot of these AI workflows that I'm mentioning here. But what we're seeing is that to handle very sophisticated production-ready code, it's not there yet. A lot of the vibe coding tools out there are great to getting you to prototype, POC, MVP. But then once you get to that point, you need to start thinking critically in terms of, okay, like, how extendable is this? Where can I go from here? As things to keep in mind. All right, so I'm gonna tell you about our specific build workflow that we have. So what we do is that we typically build 70% of whatever we try to conceptualize or prototype within VO. Now, this is important. We don't wanna jump to code immediately, we try to complete all the core screens, page routing, and all that within Vio. Once the base is solid, we then sync the Vio project to GitHub and then open the repo inside of Windsurf. And this is where we switch from visual to then code prompting. And the reason why I do this is because since Vio handles 70%, Windsurf's help comes in and fix and optimize the remaining 20%. So here it's like where it could handle error handling, fallback logic, all that kind of stuff within the IDE. Now, you know, important to know here is like you can use Windsurf and I have a bonus content here that I got from online of somebody doing this all within Windsurf and it's captured and he's not a coder. But when it comes to more complex product builds, like you're going to start hitting some walls and that's where you're going to maybe need to bring on a developer to help you. But if you're a non-technical founder, I think using the agents within Windsurf, you could get far enough that you could actually have an MVP, but you might get challenged eventually. Anyway, going back to this. So once we make any kind of the fixes or changes that we need to do in Windsurf, then we push all the changes that we did back into GitHub, right? That's our main repo. Now the beauty about this is that then VEO auto-syncs. So it sees the new changes that we pushed. And once we're fixing, done fixing and refining everything in Windsurf, we then can see an updated code within VEO. Now the beauty of this is that then everything lives back inside of a visual editor again. Now Windsurf is great for logic, like coding and things like that, but not for UI fine-tuning. That's where VO comes very handy again, because then you could go back into VO and resize any of the buttons, add logic to any components. You could tweak the layout and adjust any kind of styling visually within VO. And obviously, if you need to kind of go back, you could kind of keep this loop going.

All right, cool.

So good. I got— I'm going to try to see if I could get this done in the next 20 minutes. So I'm gonna start the live demo now of what we're gonna be vibe coding. So I'm gonna aim to see how far we could get in like 15, 20 minutes. The timing, I'm gonna be doing this live. So a couple disclaimers I wanna point out is that the process matters more than perfection and live demos can be unpredictable. So, you know, fingers crossed if this doesn't work, you know, I'm sorry, but I'm gonna try to get it to work to see what we could actually accomplish in 20 minutes. The goal isn't to create the perfect solution on the spot, but it's really to demonstrate the process of leveraging AI for rapid prototyping. Okay, so what I'm going to do is I picked an idea. Okay, so let me jump to this. So the idea that I picked is funny enough, the one available on ideabrowser.com currently, which is Greg Eisenberg's website. So as a free user, you get access to a free idea of the day, as I mentioned. And so the one that I'm going to be doing is building out an AI immigration assistant. I looked through it, I thought it was pretty good. I think it's something we could maybe come up with in the time that we have, but I'm going to be doing a bit of a spin on this because this is very much more of a B2C product. In the way it's positioned. And I'm a big fan of B2B because it's easier to find the businesses, it's easier to interact with them, it's easier to sell to them. So I'm gonna try to tweak this in the time that we have towards a product focused from a, to a B2B perspective, okay? So hopefully Greg doesn't hate me for this, but I'm gonna be stealing his idea and seeing what we could vibe code in the time that we have. Alright, so I have a copy of this as well, which I have saved here as a link that you guys can access to because after today this is not going to be available to anybody to view anymore. So I saved a PDF copy of it here just for reference. Alright, so here's what we're going to do. We're going to open up ChatGPT. I have my ChatGPT open. I have my O3 model enabled. I'm going to turn on deep search. What I'm going to do is I'm going to go up here and I'm going to import all this content. I'm going to import it and then I'm going to give it some instructions.

God, I keep losing this tab.

And then I'm going to incorporate the why now stuff.

Oh boy. I'm going to bring in the proofs and signals. Market cap. Execution plan.

All right, so just to explain this a bit more, so they— Idea Browser specifically, trying to make it simpler for everybody— they, they build out this solution based on pulling a lot of data from different systems. So they rank this based on opportunity, problem, feasibility, The why now, you know, they talk about business fit, execution difficulty, go-to-market, and they're kind of the way they categorize it in the community signal. So you could see here they pulled data off of 5 subreddits, 5 Facebook groups, 13 YouTube channels, and some other stuff. So I can't see anything because I'm not paying for it. So this is as far as we could go from a free perspective. All right. So let me give this some more information.

So, so I'm going to provide you with an idea around immigration AI assistant to help workers with their immigration process to come to the USA. Um, below is, uh, market research that was done. Yeah. Uh, now let's give it some instructions. This current idea is based on a B2C approach. As you can see, I would like to change the scope of the offer to be focused around selling this type of product under, under a B2B offer by targeting immigration firms, law firms, etc., to help them better manage their candidates who are going through the immigration process. You are, you are to review the information and conduct any online market research, better understand the problem and opportunity for B2B.

All right, I think that's enough.

And there's a lot of relevant stuff here, but it's fine just to see what it produces. Lastly, you are to ask me any clarifying questions before starting.

So I like to do this because it sometimes asks pretty good ideas. So let's see what it comes up with.

All right, so let's see.

So what specific service do you envision the assistant?

So yeah, client intake automation, asking them personal information required for immigration, providing documents. To review, case tracking, payment handling, etc.

Do you have any target firms and size?

So yeah, let's do boutique immigration firms and multi-practice law firms, uh, US-based. Would you prefer the research to focus on competitive solutions or retargeting B2B immigration workflow? Yes. Any current pricing? Yeah, per— let's do per firm, make it simpler.

All right, so it's off to it. So this, this process can take 10 minutes. So while this is done, I mean, while this research is being conducted, um, then I, uh, will continue into the rest of the, the deck, and then I'll come back to this, and then we'll jump back and forth. It's just this might take a bit, but I'm doing this live, so I hope you guys understand. I want to be as transparent as possible so that everybody can see the exact steps I take and learn along the way. Okay, so basically the next step after that, we're going to then create a PRD. So I have a link to a blank PRD in my Google Docs that I'm going to have. And once the research is done, I'm going to then ask it to basically create a PRD. There I'm going to provide some additional instructions, but I want to go and do the research first so that it can get trained and understanding of the problem. Then we're going to push this out to the— in order to create that initial prototype. And there we're going to try to iterate, tweak the UI, and then I'm going to publish it to an actual staging environment. And then once that's done, then I'll share it with you guys. You can play around with it. All right, so while that's working, so I want to talk about some common traps. This is important. So leveraging AI to speed up how you build and test your MVP. But remember that a clear problem statement and understanding of your user is irreplaceable. So there's always tools out there like Idea Browser, ChatGPT to find ideas. Now you could definitely use these tools, but I'm always a big fan of people who actually try to scratch their own itch. So if you have a certain frustration or pain, it could be a small thing, but if you have it, there's a very high chance that other people would have it. So maybe try to focus on your first MVP that you want to build out towards something around that. Okay, don't get like, don't get overwhelmed with all these different ideas everybody has, everybody's talking about. Just focus on what you think matters to you, because if you align closely with the problem, that would help a lot in terms of like your desire to actually solve for it. The next one is avoiding scope creep. Now you're going to see when I'm going to ask it to generate the PRD without a ton of instruction, it's going to give me like the bells and whistles of probably everything it could fit into an MVP or a product. Now you really want to try to stick to your core features that address the problem. So I'm hoping that when ChatGPT comes back to me with its research findings, it'll see that a lot of boutique immigration firms are probably doing everything manually, which I'm pretty sure is the case because I've dealt with a few helping immigration workers come to Canada to work for us. So like, it's a bit of a struggle and I know that they're doing it manually. So like, I'm hoping there's a market here and that we could sell to them. The next thing is obviously avoid building for an imaginary user. I mean, that, that user could be you, like try to build it for yourself. And I'm going to give you guys a nice bonus piece of content here of a guy who went and did something very similar. And he captured it all on video. And I think he did a fantastic job of showing it. So I will give that, guys, to you at the end as a piece of content. But get out the prototype, get it out to as many people as you can. So like, if any of you guys on this call by chance happen to be working in immigration, Perfect. You are going to be my first target customer. You're going to be the people who would see this product and tell me yay or nay on what should be improved, because like, I'm not an immigration expert, so I'm relying on AI to do it. So that's where feedback generation is going to be super, super critical, and you're going to want to show what you have. Now, the beauty about this whole, you know, opportunity, guys, is that You're able now to take something from concept to a prototype in like in an hour, and then you could just go and present it out to the people that you have. Now, how you find those people and where you source them, that's something you could take a look at this other webinar I did around finding product market fit and all that kind of stuff. But there's a big thing going on where people are building in public, they're publicly showing exactly what they're doing every day. Super helpful to provide that visibility. The, the guy who I'm to— I'm going to show you the piece of content around— Alex Finn created something called CreatorBuddy. Now he has 300,000 followers, so when he was talking about building this product, he got already a lot of like early validation of like, yeah, this is cool, something I would use. And he got to like, I think, 100— he claims he got to 100K ARR in day one. Now obviously when he posted the product out to his following base on Twitter, which he has 300,000 of, Um, he already had a distribution channel. So think about your distribution channel. Think about where these people are. You could use tools like Instantly, um, which now has great new features where it has a copilot to be able to source your ICPs, to, uh, enrich their contact details, to create the email campaigns and to launch the email campaigns. And I think I'm paying less than $100 a month for that type of tool. So you could use that tool to hit up your target audience, whoever they are, and get them a link to the staging environment that we're gonna have today and be like, hey, what do you think? You know, and hit up, in our case here, it's gonna be immigration or law firms, okay? Another thing I want to talk about is when using third-party AI platforms like GPT, be very careful about what data you feed to them. I've heard some horror stories of people giving a lot of proprietary or even PI data, like customer data, into these things and it creates leakage and things like that. Just be very, very careful. I already mentioned this, but don't get into this mindset that you build it and people will come, like figure out and spend time on your distribution channel. It's very important. A lot of people think build it and they come, but anyway, let's keep going. So key takeaways. We're really at an unprecedented time where a single founder can go from idea to a working prototype in a matter of days or weeks using these tools. So like, I have a very basic playbook that I showed you guys here today. Remember that AI just gives you like the tools, but you must kind of choose the destination, how you kind of get to the end. The key takeaway really is to empower yourself. Don't wait. And I'm, you know, I'm going to go against here what I'm going to say, but like I run a consulting firm, but don't wait to go and hire a company like mine. To build out your product or hiring your full dev team. Just go and do this yourself. You can do a lot now with a lean team, especially with, you know, helping with the help of AI, and then bring in the right experts when you need to. But try to keep it lean, you know, try to get to the point where you're going to get to a million per employee efficiency. That's like the end goal. A lot of people are thinking about, a lot of VCs that I'm speaking to are thinking about Is it possible to get to a billion per employee as like the next threshold, right? Like, can a solo founder become a billion a year AR company under one person? Possible. You know, a lot of agentic systems are coming into play where they can all work together. So who knows, it could be possible. So I did a quick calculation of the cost. So it costs $80 basically with the tools that I outlined to get started today. $20 a month ChatGPT Plus, $20 for VO Premium, which I'll show you, and then $20 for Vercel Pro and $20 for Cursor. So, you know, I think with $80 you could get started pretty with something pretty impressive. So pick an idea, your own or Idea Browser, start using ChatGPT, create that PRD, create a prototype and share it with people and iterate. All right, so this was the bonus content that I had here I want to show you. So there's a video here of a guy who spends 30 minutes, his name's Alex Finn, going step by step. He has like 6 steps that he goes through and he used, he ended up using Bolt instead of VEO because he had some issues, but he used Bolt to build out another app idea in the recording. And then he used Windsurf. So you could see how interfaces with Windsurf. He's not a coder, so he's just talking to it using plain English and he was able to get this app built very quickly. All right, let's see where we are at with our stuff. Okay, so it's still reading a lot of stuff. All right, so while it's doing that, you see, this is why doing live demos is challenging. Um, what I can do here, um, maybe good stuff so far.

Let me— why don't we do some questions while this is getting done?

Perfect, let's do it.

Um, so I'm just gonna go through this. So yeah, just for people, we are going to send out the recording. Ram, shoot out the links later on to us and we'll get it to everyone. So yeah, you'll get the PDFs and all this other stuff. All right, um, I'm gonna show this question. I don't know, can you see it as well, Abraham? Or you cannot?

Yeah, how friendly are these tools for founders with no technical background? Yeah, very friendly. I used to code early days on my first startup, but I don't anymore. And there are some prototypes I've been building without any newer technology understanding and things like that. So my workflow, that is the one I went through. I use ChatGPT to come up with the requirements. I feed into VO, I connect it to them to Windsurf, and then I just interface with it purely via like a prompt chat interface. And that way I could get my MVP kind of built and running.

Perfect. Thanks. Sorry, just a quick housekeeping. I know this, this event is for the next 15 minutes. It looks like we're going to go over, so whoever needs to hop off, feel free to hop off. We'll send you the recording because it looks like we're going to be here for a minute, but that's fine. So whoever can stay can stay. I'm sure Ram will finish the event until he's there. Okay, cool. I'm going to do another couple questions because it looks like this thing's working. Matthew from DC is asking, can you say more about a two-way Git sync? What does this workflow look like?

Yeah, sure.

So In Vio, if you can see my screen now, so this is vio.dev. So I'm signed into the account right now. And the way it works is you connect your GitHub. I'm hoping to do this with like an actual thing, but let me see if I could show you something else.

So maybe, yeah.

Let's take a look. Okay, so here's another project I'm working on. So you guys see my screen here. So we have a fork here of something we're working on. You see here the connect to branch. So what you do is you connect to GitHub. This one's currently broken, but then you select the branch that you want to create and then you basically commit the code that you have in here into your GitHub.

That's it.

And then you open up that repo then in Windsurf and then it will pull, transcode and, you know, make everything available on your local environment. And then once on your local environment, you can then set up the localhost or like the— to be able to interface with the app and then it could then using prompts you can iterate on and then push it back to GitHub.

Okay, um, Pawel from Poland is asking, do you still recommend doing some idea validation with potential customers before building a prototype, or is just doing deep research via GPT enough?

Um, so it's a good question. Um, I would say it depends, uh, First, I'm always a big fan of doing customer research. I think it's super, super important to be able to understand who you're building something for. But sometimes in some use cases, I've seen that people just want to see something. People just want to see something and they want to understand what it is that you're trying to convey. So that's something you need to kind of make a decision on. If it's a B2C app, I would recommend just trying to vibe code it. It yourself. If it's a B2B app, I would recommend maybe trying to connect with whoever the target audience is, do some discovery with them, and then learn in terms of what their pain points are. And then once you have those, then feed them into like a ChatGPT to create the PRD. And then once you have that, then feed into like a Vibe coding tool.

Cool. How are you with your deep research, Ajan?

Yeah, let's see.

Oh yeah, it's, it's going, uh, It's going full detail right now. 27 sources.

Yeah.

Okay.

You know, more, um, uh, Vishal from Mumbai is asking, how do you see live code tools like V0 and Cursor versus no-code tools like Bubble and Softr? It's a good one.

Yeah. So I think, uh, if you look at how things start off, like I think Bubble kind of came out first. I think they did a great job, uh, including Builder. But that company got into a lot of heat recently. But, um, I think the problem with those tools, to be honest, with a lot of customers that I've been working with over the last year, a lot of them are trying to get off of them. Like, a lot of customers trying to get off Bubble, they're trying to get off, and they're trying to move to more of a controlled code base which they own. Now, the benefit of that is obviously you're going to have a lot more flexibility and, and control over your codebase to do whatever you want. Those tools like Builder, they'll eventually hit a point where they're not scalable enough for your needs. And once you make that investment, it's going to be pretty hard to fork out all your code and, you know, move it somewhere else because they've locked you in. And this is exactly what my problem that I have is with things like Replit and Bolt is that they try to lock you in within the ecosystem and it makes it a bit harder to move things out, especially once you want to kind of continue off of your journey. I recommend that workflow which I mentioned, which is using Lovable and then using an IDE tool to kind of perfect out the code using prompts and then you know, perfecting the MVP that way.

Cool. I've heard the same things. What happens is, especially with Bubble, you start building, you start doing stuff, but then you've built everything on Bubble, and then to move off, like all of a sudden you're paying like $400 a month, and you're like, what's happened here? So just, yeah, be wary of that. I like this question, it's a bit technical, but Mohammed from Istanbul is asking, When third-party APIs are involved, like maps or payment services, how well does AI work for generating MVP given that those endpoints are shared token and all that type of stuff?

Yeah, for sure. So, um, a lot of tooling, uh, are a lot of tools like if you look at Canva, Figma, uh, are all making available an API endpoint. So as long as you can get the API endpoint from those tools and you could interface it with, say, Lovable or Windsurf, wherever you're going to be doing most of your requests through that API, then it doesn't become as much of a challenge anymore. So there's something becoming more widely available, like MCPs, where more and more companies are creating ways for you to interface with their products and solutions. So like, I would say it was a challenge before, but not anymore, because a lot of companies are even moving to like usage or credit-based tiers where you get charged for whatever things that you use.

Awesome.

From, from the API in terms of tokens.

So I don't know how to say this first name without butchering it, so I'm not. All the way from Honolulu, uh, this individual is asking, how do you take care of the backend stuff like auth, database security and stuff? So there's like, which product does that? Which tool? If you can comment on that.

Yeah, so I would definitely do that all within Windsurf. So you could ask Windsurf to set up the right relevant kind of database. You could ask it to implement your Stripe for monetization. Um, what was the other question there was in that?

And security.

Yeah, and security. So the security one is a tough one, to be honest, because in order to get like certain types of standards, you need to kind of do your SOC 2 and PCI compliance if you're going to be managing any kind of payment information. But yeah, data privacy is a big one, to be honest, in terms of like an issue in the AI space right now. But it just all comes down to where you're hosting your data at the end of the day. So if you're using Vercel to host your apps, then I think they have a pretty good, they security layer.

Okay, awesome. We have a lot of questions coming in and I'm trying to avoid the ones that you've already answered. Yeah, around security, non-technical.

Uh, yeah, sorry guys, it's taking forever. This is the longest one I've ever had, surprisingly.

Try refreshing the screen. What will happen? Will it just all disappear?

Let me see.

Sometimes that happens because I get a notification on my phone to see if it actually did finish.

No, it legit hasn't finished yet. All right, let's take a look. No, still going. Yeah.

Any other questions I could answer?

Yeah, let's burn through this. We're definitely not here in 8 minutes, folks. So, you know, if you had plans, you feel free to hop off and we'll send you the recording as well. But this will live on YouTube in a bit.

Oh, there we go. There we go.

Okay, I'm going to hop off now.

All right, cool. Let me take this back. All right. So it looks like it did analysis on B2B migration. If you guys are interested, I'm going to share this with you guys here.

So I'll post a link here.

So let's see, I'm going to try to skim through this.

No, no, no, no, no, no, no, no. Hey, okay, okay.

Pain points. So labor-intensive form filing. See, the nice thing about O3 is that it actually sources So issues with like AI before was like things around hallucinations, but I've been using this tool enough now where I could confidently say like I verified a lot of this stuff and it's getting pretty, pretty accurate. Uh, document overload, review, client communication, case visibility.

Okay, cool.

Multilingual complex intake, case tracking, deadlines, billing, invoicing complexity. Okay, interesting. Integration and workflow gaps.

Okay.

So then it goes into competitive landscape.

Great.

Okay, so that is very helpful. I love that because then it gives me a way for me to prompt for the PRD. Okay, so here what I like is that it shows me the solutions, what key features and focus they have. Okay, market demand. All right, there's a lot here, so what I'm going to do is just streamline this. I don't usually do this, but let's try.

Based on the pain points uncovered and the market gaps that exist, um, according to the analysis Please.

I like saying please because, you know, we're all going to succumb to AI overlords in the future.

So please create a product requirements document, PRD, to create an MVP solution of the core functionality that I can then share to the target audience being law firms and immigration boutique companies that specialize, that do this. The PRD should clearly state the functionality as well as having clear user stories broken out for each functionality, as well as acceptance criteria, which I will then feed into another system to generate the UI prototype. Ask me any questions you have.

All right, this is going to be a lot quicker, guys, because it doesn't need to go do a deep research. So, um, it's going to work on creating the PRD now for me, and then I'll actually show you how this looks like. So actually, the longest piece of this whole workflow is actually using the O3 model. So I should have timed this better, but let's see what it says. Okay. Okay.

One form scroll.

Which immigration form should the MVP auto-code?

Employment-based. Let's focus on employment-based. To start. That is PDF export, human upload accessible. Yes, PDF export and human upload are acceptable. No transmit to portals necessary.

So I'm just going from like a product perspective. I'm trying to keep this lean.

Okay.

Branding only.

Yeah. Versus white label custom domain. This would be a SaaS app. So no white labeling. For existing, which practice management CRM or document should we plan? Okay. Plan integrations for Clio and MyCase only.

I have no idea what these are, so I'm just kind of going along with it, okay?

Preferred Stripe payments, okay, so Stripe for payments.

Do target firms require SOC 2, Yeah, or is end-to-end encryption sufficient?

End-to-end encryption sufficient. Start. All right, client-facing language, only English.

Timeline or pilot size?

Are you aiming for a limited pilot? A limited pilot. All right.

All right, we're at time now. So let's see how fast we could go. I promised 50 minutes, guys, but I'm sorry, ChatGPT 03 model took a very long time to do this. But basically what we're going to be doing next is we're going to be creating a new chat. Let me start off with the prompt here.

I will provide a product requirements document for a B2B product targeting Immigration firms, you are to create a prototype based on what is in the PRD.

So simple, no fancy prompts, nothing super crazy. Okay, great. So I'm just gonna So let me go through this quickly. So purpose and vision, this is good. Document upload review, target users. So these are my ICPs. MVP scope and functional requirements. So broke them out between higher level kind of epics, so to speak, for us to consider, including key user stories and acceptance criteria. Then I did ask it to think about the user stories. So the non-functional requirements piece, like, it's not totally needed. You can kind of skip this stuff. It tries to overthink things for you, not very relevant. But the user stories and acceptance criteria, the detail, this is super important because this is what VO and other tools are going to be using to get a very good and clear idea in terms of what they actually need to design and build. So you could see here the different kind of epics, the success metrics. So form field auto-population accuracy has to be 90% accurate. And then there's like the stuff around the release pilot, out-of-scope stuff, assumptions.

Okay, good, good enough. Let's go.

So I literally copied and I'm gonna paste. Okay, so nothing fancy. Then I go File, Download PDF. Okay, I get the PDF, then I go back to VEO. Now this is a new project. I'm going to be using their VEO large model. I'm going to upload this from computer. As you can see, you could also do it from Figma. So it's going to, it's going to upload now. Let's just go. So let's see what happens.

It's thinking.

Anyway, while that's happening, I'll share this stuff out with you guys. We already went to the QA. I have some office hours here. Um, you guys can book a time with me. You could either click on this link that I'll share—

whoops—

you guys want, I'll put it here if you want to book any time with me, or you can scan the QR code. Um, but yeah, that's the end of this specific webinar. So let me go back into here and see where it's at.

All right.

So it's doing a lot of thinking today. You guys have any additional questions around what I just did there? I could answer. All right. So as you can see, there's been a bit of an update. So it's read the requirements. It's going to be focusing on the visual representation of the core features. The prototype will be a Next.js application. So that's what I like about Vio. We're big on Next.js, and so it actually builds on that. So it's going to be focusing on the main layout, dashboard, case management, billing, and settings. So it's working out on the core structure now. So it might take a couple minutes here as well. Normally I clean this up, like when I get this, I would typically clean it up or I would reprompt it, but just in the interest of time, I'm not going to be doing that. But You get the idea, right? You could go back into ChatGPT and reprompt basically what it is that you're trying to do so that it has a more clear instruction set at the end of the day.

Let's see, guys.

Okay, so I think what I really started this About half an hour ago. So it's going to take at least half an hour, if not more. Hey, all right.

Well, hey, hey, folks, in case you're wondering, why am I here and not Ehan? Ehan had to hop off real quick, but I will be here helping Aram close the event out. My name is Daniel. I'm a part of the Founder Institute marketing team. While we wait for this, do you want to maybe go through a couple of questions?

Yeah.

Awesome. I see here we have a question where somebody is asking, when do you decide this is the time for the help of a professional versus AI? As someone who's been helping people build applications, when do you make that decision?

Yeah, good question. I think the simple answer to that is when you're stuck. Like, uh, I've seen this a few times where people or founders attempt to do it. Like, they get so far as building out the PRD, they build out something in Lovable, um, they show it to us as like their prototype-ish that they want to do, but then they realize that there are limitations. So I think it's a good question. Let me, let me talk about that and expand that out. A lot of these tools are really good if you're starting off with a fresh slate, a clean idea. You want to prototype something, you want to get it out. Unfortunately, it doesn't work the other way around just yet. What I mean by that is—

Oh, shit, hold on. I got a runtime error.

In situations like this, I mean, I'll just deal with this. So I received a runtime error.

Please fix.

All right, so let me go back to the question. Um, so right now, like, the way I've been doing this workflow is like I create the designs, I push it to GitHub, and then in GitHub, um, I then push into Windsurf, right? So right now there's no way to do it the other way around for the time being, where you already have an existing design or existing app where you want to kind of push it through Windsurf into Vuo. So for you to be able to improve on the design, that is not available yet, but it soon, I think, will be. They're working on it. The other problem that exists currently with some of these tools is like admin panels. So say you build out an app, it has like the basic functionality for you to take payments. There's a database and all that kind of stuff, but like for you to actually manage it and manage the users, it's an extra layer of complexity. So like these tools just can't do that yet. I would recommend looking at things like Retool or Appsmith in order to basically do that for you, but that's where you need to create like an API to be, to be able to interface with these kind of like API-friendly middleware admin panels. All right, so looks like I got something.

Let's take a look.

All right, so I have—

hold on, let me go back to the dashboard.

Oh, it's buggy.

All right, collapse.

All right, so looks like I have just like a simple dashboard to start. Where I could see a workflow of client intake. LCA field, I don't know what that is. Petition ready, decision pending. This isn't working yet, but I could go into here, which is nice, where I could take a look at the specific case file for the person. I could send intake link, upload Generate forms. I could create an intake form.

You can upload documents for that specific file.

I could—

the billing is not working yet.

But in these situations, now that I have this working, I would actually just go back. I would actually just go back into VEO and start fixing everything. Like, for example, fix this. So in this situation, let me go back here. In the sidebar, uh, there is overlapping text, uh, for the name of the app.

Just change this so that it isn't broken. Make the name Legal Joy. I don't want to see what happens.

Uh, but yeah, so now that you have it here, you just continuously use the prompts to fix. So if things aren't working, like for example, you saw that it went to like kind of like an error page. That's because it hasn't generated the screens for that yet. So in that situation, you just tell it to generate the screens and it'll actually go and generate those screens that didn't properly get fixed. So it looks like, okay, so there's some sort of issue here with like the dashboard overlapping. But anyway, I'm going to just deploy this just so you guys see.

I'm going to deploy to production and I'm going to try to get you guys a link.

This is just going to be a link that will be shareable for you guys to be able to go through this yourself. While that's happening, any other questions I can answer?

I have one.

I saw something in the chat here, sorry, the Q&A. How should you think about scale when it comes to building something with what, you know, with the tool like Lovable or just white coding in general, right? Like if you know this is something that you have to scale eventually, should you even begin with an AI tool?

Definitely. I think the way you should think about anything when you do an MVP, it's throwaway. And I know that's probably harsh for a lot of people to hear, but An MVP's main job is for you to get something that you have in mind into a prototype state and then getting it in front of the customers, getting them to rip it apart, give you the feedback, and then going back and iterating on it. So very likely whatever you started with is very likely gonna change. And what I like to say is like, don't focus on the solution, focus on the problem. And that way when you're constantly thinking about the problem that you're trying to solve, then the whole solution's probably gonna be changing. From a scalability perspective, like, I think the reason why I like this specific toolkit or this workflow that I've been outlining here is it allows you to be able to continuously build on and scale from. Now, obviously there's going to come limitations where like you need to build out other portals, other tools, and that's where you can start more integrations with, with other kind of elements. But it should get you to the point where like you should be able to get a pilot customer if it's a B2B, or if it's B2C, you should be able to get a couple paying customers to this approach. But at least that way you have like some traction and some, and some ability to get something out of the market.

Awesome. Yeah, that makes perfect sense. And I think it's like you need proof of concept, right? So I think that's like a great to start.

Exactly, getting a proof of concept is super important. So what I'm just trying to do right now is that I'm actually trying to complete my workflow of this whole process. I'm trying to publish my app that I have that, that is here into a staging link so that I could then share it with you guys. So it's right now I have an integration set up with Vercel. Which is the company that's gonna be hosting this app, which is owned by Veo. But yeah, it's gonna take a couple, probably a couple minutes here, but I could answer. Oh, there we go. So looks like a v4 version of the app was published. Here is the link, guys.

Oops.

All right, there you go. Check it out. Rip it apart. Give me the feedback. Anyway, so I think that's a good way for me to kind of stop here in terms of like completing this kind of experience that I'm wanting to kind of show you. Again, I think this is a more complicated product that I decided to do here. Normally it's a lot smoother. It all comes down to the simplicity of the product you want to build. So this is a more complicated one. That's why there was a few hiccups along the way. Okay. Any other questions?

Awesome. So let me see. We do have a couple of questions. Maybe you could just kind of really quickly go over them. People are asking, how do you pick the right tool? Even in terms of if you're picking between ChatGPT and Claude, what are some of the criteria you should be thinking about?

I've tried out all of them. Here's my two cents on it. ChatGPT, I think, is a bit further along than Claude or Gemini or Perplexity. I like the way that it works in terms of its approach, its research. It's the most cost-effective one for me to get what I want. And then in terms of like the AI tools, like I like Vio. I think the designs that it creates for more complex solutions is better. In terms of the IDEs, I recommend Windsurf. If you're a non-technical person, I would recommend downloading the IDE, there's a free one.

It's this one if you guys want clear.

Here's a link to it. They do have a free one that you could get started on with, which gives you basically enough of a starting point. In the video that's done by Alex, he shows, I think, it skips, right? Because like nobody does these things really in 30 minutes because he does like pauses, but it basically shows how he ingested into Windsurf. He talks about it, doesn't actually show everything in clear detail, but I think it does a good job. But Windsurf is what I would recommend. It's pretty founder-friendly. The other ones like Replit, I won't recommend. I would recommend Bolt over Replit. And then Lovable, I think is moving fast. Every week they're releasing new features and so is V0. It was only, I think, a couple of weeks ago that VO did the integration with GitHub, which then makes it into a powerhouse with the whole WindSurf integration, which I mentioned. So like definitely something that you could start using and playing with today. Now, just want to be clear, like in these situations where like VO starts struggling with the code, that's maybe a good time to switch to WindSurf. Now I'm not going to do it just in the interest of time. It might take another half an hour, but I think I've gotten you guys all far enough to kind of show what is possible with this kind of like solution. And then the video which is in the deck gives you kind of another standing point to go off of in terms of like how to complete like the whole thing with windsurf. But there's so many videos on Twitter—

X, sorry.

Which people talk about how to do this as well, and they build in public. So like you have a lot to go off of.

Okay, so I have one last question before we close this out. Let's say you are somebody who wants to learn how to build something, right? But you don't necessarily have an idea, like what's a good starting point?

Um, I would just say start hacking it. Like, just start. Don't, don't think about trying to like overcomplicate it. Like, I didn't know this stuff. I mean, I've been building product for a while, but these tools are all kind of new to me at the same time. And I just started kind of hashing it out myself, figuring out what prompts are working, iterating on the prompts, testing out different tools like In my initial idea for this webinar, I wanted to basically take a PRD and do it across Lovable, Bolt, Replit, and VO. But like, that would just take too much time. But that's something that I recommend doing. Like, get up, get all 4 up and running and start playing around with them. And the cost is nominal. Like, some of them have free tools. Actually, pretty much all of them have free accounts with limited credits, but it should give you enough of a starting point. Starting point to get started.

Awesome. Yep, I think that's, uh, that's a pretty good answer, uh, from my experience in terms of like when we're trying to build out a tool, right? I think it's always, uh, good to go to what's familiar. And what we did was we started off with Bolt, and at the time, because it was like right at the start there was competition between who's better, and at that time Replit was kind of better, right? But I feel like just in terms of being a user, I found that familiarity to be way more helpful because at the end of the day, in terms of functionality, they're more or less, they're on par, to be honest, right?

They are. And to be honest, it's a very competitive space right now. If you look at GitHub Copilot, who was the first kind of like AI copilot for devs, that quickly became outdated with Cursor. And then after Cursor got popular, WindSurf came out and basically killed them. And then you basically have dev teams like mine moving from GitHub Copilot to Cursor to Windsurf, and then if Cursor releases something better, we might go back to Cursor. So it's like, it's a bunch of like jumping around. So like, figure out what works best for you. I do understand like investment in familiarity. Actually, one of the lead devs I know at a company, I can— I showed them Cursor, and they moved from GitHub to Cursor for their entire engineering team. It's like, I think 50 people. And now I told him about Windsurf, he's like, oh man, like, I can't move that fast. Like, I can't move now my whole engineering department now to Windsurf. But that's just the reality. Like, honestly, AI products are shipped out, you know, every hour. And it's a very, it's a lot of competition. And so I would recommend just like figuring out what you're comfortable with. And the beauty about these tools, especially now with like some of these new recent feature releases with like GitHub integration, is your code now is going to be living in a GitHub repo. So it's transferable. And what I mean by that is like what I mentioned earlier is that once these tools start creating the ability for you to import your existing GitHub codebase, or importing your entire full Figma design system that you have for your company, then that becomes a powerhouse because with that capability, you could basically have anybody on the team be able to go and ship code because that AI platform is gonna be very familiar with the code base, is gonna be very familiar with the design system 'cause that's all in Figma. And you could be able to go and add new features purely through prompts.

Awesome, awesome. And there you have it, folks. Thank you so much for joining us today. I hope you really, really enjoyed this. We had over 1,000 people who actually signed up, and I think at some point we had about 400 people live. So yeah, I don't know, Aaron, if you could see the hearts and all the emojis.

Yeah, I do. Thanks, guys. This is awesome.

Awesome. Uh, so folks, what we'll do is we'll send out a recording to you in the next 48 hours, so you don't need to worry about that. We are going to be sticking around for networking, so if you want to connect with founders from across the world, maybe other people who are passionate about building products using AI, please do stick around. We also have some really interesting boot camps happening next month, so we'll have somebody from the drop links in the chat. Do check them out. Aaron, before we close out, any last words?

Yeah, no, I hope you guys enjoyed this. I do apologize for going over and not being able to complete everything in the 50 minutes that I had set out for. But I mean, doing live demos is— it's here, it's hit or miss. That's why I said fingers crossed we'll try to get it done. But I wanted to be as authentic and real with everybody just to show you guys what a real workflow looks like. And not try to sugarcoat it or basically have files already done and everything kind of like ready, set to go and like actually show you how I think through it, what kind of prompts I use. I mean, obviously could have been better, but I'm just trying to keep it real.

Awesome. And thank you so much for taking out the time to actually do all of this. As you can see, everybody loved it. So folks, that's it from us here today. I hope you have an awesome day ahead. We'll see you in the next one.

Cool. Thank you, everybody. Bye.