The AI Daily Brief: Artificial Intelligence News and Analysis - How Vibecoding Unlocks AI feat. Riley Brown

Episode Date: April 15, 2025

Riley Brown joins AI Daily Brief to explain Vibecoding, an accessible way to build software with AI. Riley shares how he went from content creator to Vibecoder, the tools that made coding easy for anyone, and the types of apps people build. Riley also reveals how Vibecoding might evolve and his goal of making app development as simple and fun as Canva made graphic design. Get Ad Free AI Daily Brief: ⁠⁠⁠⁠⁠⁠⁠⁠https://patreon.com/AIDailyBrief⁠⁠⁠⁠⁠⁠⁠⁠Brought to you by:KPMG – Go to ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://kpmg.com/ai⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ to learn more about how KPMG can help you drive value with our AI solutions.Vanta - Simplify compliance - ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://vanta.com/nlw⁠⁠⁠⁠⁠Plumb - The Automation Platform for AI Experts - ⁠⁠⁠⁠⁠https://useplumb.com/nlw⁠⁠⁠⁠⁠The Agent Readiness Audit from Superintelligent - Go to ⁠⁠⁠⁠⁠https://besuper.ai/ ⁠⁠⁠⁠⁠to request your company's agent readiness score.The AI Daily Brief helps you understand the most important news and discussions in AI. Subscribe to the podcast version of The AI Daily Brief wherever you listen: https://pod.link/1680633614Subscribe to the newsletter: https://aidailybrief.beehiiv.com/Join our Discord: https://bit.ly/aibreakdown

Today on the AI Daily Brief, the vibe coder's vibe coder. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. To join the conversation, follow the Discord link in our show notes. The early lead contender for AI theme of the year 2025 might be vibe coding. Buied by tools like Bolt and Lovable and SoftGen and Agenic IDs like cursor and windsurf, more people than ever are actually putting code out into the world thanks to AI. The so-called vibe coding movement, a term coined by our Andre Carpathy, has taken root among solopreneurs and builders of all stripes, and one of the loudest

proponents and most interesting builders in this space is Riley Brown. Riley was and is the biggest AI TikToker, but his ambition has always been far beyond content and is about creation more broadly. He's just released a new app called vibe coding, which is an app that builds apps, and today we're going to talk about his journey as well as what he thinks the future of vibe coding really is. All right, Riley, the vibe coders vibecorder. Welcome to the AI Daily Brief. you doing, man. I'm doing great. I've been vibe coding. I love it. I know, I'm super excited for this conversation. I've been following you forever. Obviously, you were doing, you know, content in the AI space for a lot longer than, than you've been doing this. But I watched sort of every step of

this transition. Love it. Have been dabbling with this myself on the side. Obviously, spent a lot of time thinking about I, the weird perspective on vibe coding I have is thinking about how it's going to end up translating into the enterprise, which is still sort of very nascent territory. but it's one of my favorite topics is the short of it. So I'm thrilled to have you here. I wanted to start by just maybe having you introduce yourself to people who aren't familiar or even for people who are. I'd love to maybe get into kind of like how you've sort of transitioned from from where you started with the type of content that you were producing to where you find yourself now. Yeah, I'd always been interested in content. And I'd always realized just the power

of making content on the internet. And I'd also came across AI in, you know, in 2022 when I first tried mid-jurney. And obviously, I was making content for about two years before I found vibe coding. And so the origin of vibe coding for me was Claude Artifacts. That was the first thing in May of 2024. If you remember correctly, it was the first time where an AI company put in a code renderer into the platform. So previously, if you were non-technical, yes, you could get AI to generate code, but it's like, where do you put it? You know, it, you know, what does code do? Like, how do you even use code? But with Claude, you could say, create a cool website and make it a landing page for my business, and you would press enter,

it would generate the code, and then it would open the artifact and render the front end code right there so you could see it. And that created a ton of people who had never had an experience with code. And I think that is the key, right? To have an experience with code, because that was like this massive onboard. They're like, oh my God, now I can do all of these different things. I can create websites. I can create components, animations, all these different things. And me, I'm the most curious person ever. And I think that's one of the reasons I've done well with content. It's like when I find something interesting, I go to the bottom of it. Like in a week, I learn five months worth of stuff in it because I'm so obsessed. And so,

So immediately after Cloud Artifacts came out, Amar Reshi, who was the head designer at 11 Labs, until he just went to DeepMind, he posted this video where he took the code that he generated on Cloud Artifax and pasted it into Replit, which is one of the top vibe coding tools right now. And on this app, I could not only see the code being rendered, I could also deploy it to the internet very easily. And I could also set it up to AI API. So I could ask for a chat GBT clone and then paste in my API key and tell the AI to create an app that allows me to chat with AI. And I did that. And basically the experiences and types of apps that I could create for myself started to increase more and more. And then I hit this natural ceiling because you can only go so far copying and pasting from Claude into Replit. and then that is when cursor became really good.

I think it took about three months after Claude 3.5 came out and went into cursor, and then all of a sudden you could generate insane apps, and I was generating image generators. I made a Figma clone, and then I started making content about how I was doing this on X, and I think so, like thousands of people were excited about it, and then I just started a community with no real plan. It was just a free community, and hundreds of people per day. day at the beginning, we're just submitting their vibe coding projects. At the time, we called it software composing. It wasn't until Andre Carpathy renamed it to vibe coding. We just called it software composing or AI coding. And yeah, I just kind of fell in love with it. And then I started

building apps that I would put on the internet. Some of them made money. Some of them was just to help people. And some of them were just to have some fun on the internet. I'm, no, it's awesome. As you were kind of diving in, what was the, or maybe I'll ask, in a way that's sort of broadly useful as well. How have the band of possible apps that can be created or sort of just web app, you know, experiences that have been created expanded over the time that you've been in this space. You know, coming up on a year, but maybe especially concentrated in the last six months. Yeah, I think, first of all, I think it's transitioning into more mobile. Mobile apps are inherently harder to build than web apps. And I think there's a lot more

fun things you can do on a phone, such as haptics, such as immediate access to your camera role. There's just inherent fun things. And I think the younger generation is just very mobile first. They spend more time on their phone. Some people only have a phone. They don't even really have a laptop. And so I think that is becoming more and more of a thing, especially as I've learned more about React Native and there's other tools out there now that create mobile apps. I think that's becoming a big thing. In terms of like general capability, we've seen with, Gemini, the new Gemini 2.5 model, people are just succeeding at a higher rate to create the app that they have in their mind. People are now actually creating games, right? That wasn't really a popular

thing until Claude 3.7s on it, where if the games you were creating were like Flappy Bird, but now, as we've seen, we're seeing people without any coding experience at all, they're creating 3D, 3JS sailing games, or flight simulator games, or first person shooter games. that look like Minecraft. And this has kind of started this movement and obviously Bolt created this $1 million hackathon event. And there's a lot of culture around it. And in my opinion, it feels like it's this big movement

when you're on Twitter. But I don't think it's a global movement yet. I don't think it's even touched the masses yet because it's still unapproachable. And that's where I think the market opportunity is right now is building the Canva for vibe coding where you don't even think about the code. You don't, it's just intuitive. It actually, like, helps you.

It gives you suggestions. It allows you to, like, very easily put your brand assets within apps, games, websites, that type of thing. And so that is the most exciting opportunity in the space that I see for me. So I love this. And we're going to dig in more to sort of mobile, mobile specifically and some of the things that you're trying to build. I think that you're right that most.

people, even though, even with this, this audience is obviously extremely enfranchised and still probably the majority of these folks haven't tried one of these tools yet. And maybe they've tried to sort of do directly within Claude or chat GPT some version of this, but haven't necessarily gone out and, you know, and used a specific text to code type tool. What is the experience as it is right now for people? Like where, you know, what do you do, how do you interact with it? And then where do you start to get stuck? Where are their bottlenecks? Where are there stops? And how does that constrain what people are building so far. Yes. Okay. That's a really, really good question because a lot of people get stuck. To first answer your question, I think the best place to get started, honestly, is v0.v.

V0.dev. I think it's the best place to start because it's the easiest to create something that is fun. And I think you should start with fun. The odds that you're going to turn your first app that you vibe code into a company that brings in any revenue is not the way you should approach this. you should tinker for like two or three months and actually really get used to the possibilities and then you have a good idea of like what you can build and then you should go from there in my opinion. Just have fun.

And if a business comes out of it, you're way more likely to find it along the way than you are to like plan it before you've ever tried it. That's the first thing. V0 now also allows you to upload images directly into the chat. So you can go to GPT40, for example, create a logo and then ask GPT40 to remove the back. and you can paste it directly into V0 and say create a landing page and use this logo as for the website.

You can also generate background gradients and say use this as the background gradient. And then you can also upload sound effects, which I just made a video on that I'll be posting next week, where you can actually add 11 lab sound effects and you can add your own music. And you can make it this like fun experience, which is what I originally thought made the web fun was that it was kind of this personalized. Now there's like design rules and it's kind of. of boring. So just have fun with it to start. Where you are likely to get stuck is one, AI features and other APIs are when it gets a little bit harder. You're also likely to get stuck on

authentication and database. You know, if you want to create an app that users can sign into, and then also like on Twitter, if you make a post, that data has to be stored somewhere, you know, and that's a database, right? And people get stuck. there because it's like which database do you use? Do you use superbase, Firebase, something like InstantDB. There's a bunch of different options. And when you're faced with options and you're already in a space that you don't fully understand, like most vibe coders who are building only with AI, inherently don't know about programming. Therefore, how are they supposed to make a decision on which to use? And I think people get stuck there. The final one that I'll say is like Stripe Integrations,

where like adding payment options and then making sure that it's secure and then also that people can't find savvy ways around your payment to like use your product without paying and then just running into errors. And I think the deeper you go down the stack, which is like authentication, database, storage and payments, the more likely you are to run into errors and therefore get fully stuck on your project. And so it's for that reason why I still think we're in a prototyping phase. I've seen people fully vibe code apps that make a ton of money. I've made a product myself that's made some money, but the more, I still think the most effective way to do it, especially if you're working at a job, maybe you have some money saved up, is to use it as a prototype tool

and use it to explain the exact app that you want and then get someone to come in and help you finish it. I think we're going to move past that where anyone will be able to just build apps fully end-to-end, and that's why we're starting a company to solve that. And I think that's right around the corner. Six months to 12 months, I think the models will get better to the point where they can do it. It might be expensive at first with just how much the models will have to look at the entire codebase. But yeah, that's kind of how I look at it. Yeah, no, I think it's super, super interesting.

So I definitely validate the prototyping as the low hanging fruit and a good place to start, particularly if you're thinking about it kind of from a business context. So one of the shifts that we've made internal to superintelligence, and that I think is you're just going to see everywhere is we effectively have a fairly hard ban at this point on feature writeups or like ideas for features that aren't just, you know, prototyped somewhere because like there's two layers of radical improvement by plugging it into bolt or level or whatever you're using to describe the feature. One is it's going to way more quickly clarify what you're actually trying to do for yourself, right? If you're thinking through

it. You start with a really simple, kind of very basic prompt, and you see it manifest and you're like, oh, wait a second, that's not at all what I meant. You kind of have to refine it and you work with it. So you're refining what it is that you're trying to communicate to everyone else. And then second, when it comes time to actually communicate it to other people, you can just, I mean, it's the classic aphorism, you know, a picture's worth a thousand words, you know, an app's worth a thousand Slack messages, basically, right? Where people can just go see it themselves. We saw the Shopify AI memo say something similar around their product. product team. I actually think, ironically, one of the things I talk about with enterprises a lot is

the fact that it is so good at prototyping but still has limitations with other things, that actually is a hugely invitational thing potentially for enterprises who have real security concerns. They're not ready to just let these tools interact with their legacy code bases yet. Let your people just start messing around with the prototyping kind of feature sharing place because it's super low stakes. It's nothing that's ever seeing the light of day. You don't have to care about it from a sort of risk or compliance perspective because it's just internal. So it's, you know, it could be a place where people are building muscles. You know, so I think I think that that's right. I think that prototyping is a great place to start, although I do think it's, you know, incredibly small relative to where it's

going to be in terms of the capabilities. The other piece that I wanted to sort of affirm and share something is the starting with fun. So I think that one of the things that actually quite frequently to me gets lost in discourse of AI because it's so clearly revolutionizing how we work is that the sort of immersion experience or the experience that many of us have that kind of addicks us is that it's really fun to create, right? It's really fun to feel like a wizard, right? I remember the first time I created, you know, mid-jorney images. There's like, you know, Hemingway in Paris in the 1920s. You just feel like it's this totally like capital F fun kind of experience. And I hadn't had that for a long time until I started vibe coding. And so, you know,

my project a couple of weekends ago was I was thinking about, you know, could I create some sort of like, you know, Eldritch-themed HP Lovecraft style, you know, game that integrated tabletop gaming and role-playing and card games and stuff. This was a Friday night. I was just kind of like wanted to basically zone out, but not all the way. And then I was like, well, you know, I love the experience of like the shi-old Oregon Trail when I was, you know, in school, like the black on green screen version. So I was like, let's see if I could first like just create a version of Oregon Trail that is, you know, but like, you know, HP Lovecraft themed. And, and, pulled up, I think lovable. So I tried lovable and both. Loveable did a better job. So I was using

lovable for this. And, you know, over the course of like two hours created this fully fully working thing. I mean, it's like, it's ridiculous, but it's kind of fun, you know? And, and it's that, that experience is like, I'm not going on and trying to monetize this now. It's just this like really fun thing that I could never have done before that now end to end could actually do. You know, and in the process, I think to your point, like learned a ton about how to actually use. these tools and what their capabilities are. So I super agree with it. Just start by trying out fun things because it's a, you know, even more than some other types of AI, the capability to just do something that's awesome is very different. Today's episode is brought to you by Super

Intelligent and more specifically Super's Agent Readiness Audits. If you've been listening for a while, you have probably heard me talk about this. But basically the idea of the Agent Readiness audit is that this is a system that we've created to help you benchmark and map opportunities in your organizations where agents could specifically help you solve your problems, create new opportunities in a way that, again, is completely customized to you. When you do one of these audits, what you're going to do is a voice-based agent interview where we work with some number of your leadership and employees to map what's going on inside the organization and to figure out where you are in your agent journey.

That's going to produce an agent readiness score that comes with a deep set of explanations, strength, weaknesses, key findings, and of course a set of very specific recommendations that then we have the ability to help you go find the right partners to actually fulfill. So if you are looking for a way to jumpstart your agent strategy, send us an email at agent at B-supert.aI, and let's get you plugged into the agentic era. Today's episode is brought to you by Vanta. Vanta is a trust management platform that helps businesses automate security and compliance, enabling them to demonstrate strong security practices and scale.

In today's business landscape, businesses can't just claim security, they have to prove it. Achieving compliance with a framework like SOC2, ISO-27-01, HIPAA, GDPR, and more, is how businesses can demonstrate strong security practices. And we see how much this matters every time we connect enterprises with agent services providers at Superintelligent. Many of these compliance frameworks are simply not negotiable for enterprises. The problem is that navigating security and compliance is time-consuming and complicated. It can take months of work and use up valuable time.

and resources. Vanta makes it easy and faster by automating compliance across 35 plus frameworks. It gets you audit ready in weeks instead of months and saves you up to 85% of associated costs. In fact, a recent IDC White Paper found that Vanta customers achieved $535,000 per year in benefits, and the platform pays for itself in just three months. The proof is in the numbers. More than 10,000 global companies trust Vanta, including Atlassian, Cora, and more. For a limited time, listeners get $1,000 off at vanta.com slash NLW. That's VANTA.com for $1,000 off. Hey listeners, want a supercharge your business with AI?

In our fast-paced world, having a solid AI plan can make all the difference. Enabling organizations to create new value, grow, and stay ahead of the competition is what it's all about. KPMG is here to help you create an AI strategy that really works. Don't wait, now is the time to get ahead. Check out real stories from KPMG of how AI is driving success with its clients at KPM. Now, U.S. slash AI. Again, that's

www. kpmg. dot us slash AI. Now, back to the show. I think the reason why people's intuition, they're like,

oh, why would you spend time making a game, but you're not like trying to make money on it? It seems counterintuitive because of how hard programming has been traditionally. Whereas if you think about what happened in the early 2000s

with content, with video production in general, in general. Like, I remember my dad when I was a kid had this massive video camera and he would take videos of me as a kid. And if I asked him at the time, like, how are you going to edit that? He's going to be like, what, what does that even mean? You don't know what how to edit it. You don't know what software. He just pulls it out. It's a VHS tape, which whatever, I don't even know what that is really now. But my point is, his video production used to be a production. It used to be a thing only exclusive to people with the tools that know how to, like, edit a video.

I don't even know what probably was Adobe around then. But like now, we have these phones in our pocket that have cap cut. And now 13-year-olds know how to use cap-cut. They can edit a video from their phone. Like, they can edit a video while they're just watching TV on the couch in 10, 15 minutes, and they can post it on the internet, which is like this democratizing thing that if you, and you can start a company. I literally did this where I started a company based on me just pulling out a phone and talking to the camera on the internet. And so I think we're seeing this shift. I think it's like the cap-cutification of software where I don't know what the final form is, but I do know that it's getting significantly easier to create experiences with code. The same way that all of this software,

the phone and Cap-cut, allowed you to create experiences with video where, like, you didn't need funding or you didn't need like all these checklist of items in order to create an experience with code. Anyone can create an experience with code. And so that's what I'm most focused on is the experiences that you can create. And I think that the best content creators, I know, have this love for the game. You know, they love creating vlogs because they love it. And they've created thousands of videos.

They didn't even really intend to get views. They just kind of like learn the culture. They learn techniques for fun, for the sake of it. And I think having that mentality now, especially if you're young and like you want to do this for the long haul, I really think you should have this long term perspective of like, like figuring out what the first platform for code, because it's coming. There's going to be many platforms to create experiences with code.

It's just too fun and too easy. And so that's what I would pay attention to. It's like, how can you use, you don't need to create a SaaS company right away. Just create fun experiences for yourself. And then something will come of it. And that's what I've noticed in general right now in AI and in content. And all of the people that I've seen successful are always tinkering. They're experimenting.

sometimes for no reason, and then you will be presented with the reason when an opportunity comes if you put yourself out there. Sorry if that's like a long-winded ramble, but like, that's kind of a mentality that I have and I think is good for the state of vibe coding. Because no one really knows where it's going, honestly. Yeah, no, absolutely. I think that it's, it actually, listening to you talk about that, it, I think it helps explain why this has captured your attention because there is this through line, this continuity. through line between you had, there was a certain set of tools of creation. You liked creating things. So you gravitated towards the tools of creation that were available to you, which is content videos,

new platforms like TikTok, Instagram, right? Like that's where you were. But it probably wasn't because you saw your dad using video and you're like, I have to make videos. It's because that's what you could create. That's the expansionary creative opportunity has been in video and media. And now it's moving. And I think what you're on the vanguard of is it's moving to anything that you can build with code. That is the new creative canvas. This new creative template. And you know what? It's not like a new, I do, to me, it's like the connective tissue between everything. And this will make sense. For example, for example. So obviously we're trying to grow and we've been successful so far in the month that we've created the Twitter account for Vibecode app. We haven't launched the product

yet, but we've grown to 10,000 followers. And I didn't even announce it to my audience until we were at 8,000. And what, what I did to grow it quickly was I created a script. I vibecoded an app that automatically put the screen recordings with an outline that looks like an iPhone. And that's how I've created those viral videos. And I've been able to vibe code a solution for myself to help me with marketing. And obviously, I've been making content on using chatGBT, mid-jurney, runway, claying all these tools for all these different things. And code is just kind of the connective tissue between all of it, because now I can create video tools. I have a really good intuition on how these tools can interact with each other and what APIs are useful.

And for those of you who are untechnical, APIs are just like, I think of them as power-ups that you can use technology that you can use. Like, you can use the OpenAI API API. So anything you can do on chatGBT, you can fully emulate in a software. And a lot of these, what custom GPTs failed to do was create a good interface for the power-ups that is AI chat, right? Not everything is made for a chat interface. You need to create new interfaces, and you can do that.

You can vibe code new interfaces that makes writing more fun, that makes video editing more fun. Like, you can literally make a custom video. Like, if you have one specific type of video you make over and over again, you can make your own custom workflow for that. And so that's where I've seen all the value is as I've become more entrepreneurial, starting my own company, I vibe code mini solutions for a lot of things. I think that is just like a super powerful thing that people, have never had before in history, and it just connects all of this technology together. And you're able to vibe code that connection, if that makes sense. Absolutely. I think it's interesting. I think that most people who are deep on this would sort of bristle against the characterization

of coding or agentic coding as like a use case for AI, because kind of what you'd arched articulated is that use case as the thing that actually enables every other use case for AI, right? is the thing that unlocks of the sort of full potentiality of AI is, is to use your words, this connective tissue that comes in the form of coding and vibe coding being the thing that unlocks it for a greater set of people rather than the previous, you know, kind of the set of people who had those skills before. Absolutely. Agreed.

So let's talk about vibe code specifically, the app that you guys are building. I know it's not out yet, but, you know, there are, there, it feels to me like it's a combination of, there's a set of barriers that you guys ran into yourself. that you ran into as you were building that you wanted to try to address. But then also, I think a different kind of emphasis, potentially a different audience, you know, the choice to go after mobile clearly is not just about, you know, solving a problem. It's also about going out and seeking an opportunity. So talk a little bit more about kind of how you guys are thinking about vibe code and that opportunity and what you're hoping to bring to market with it.

Yeah. I think you could say our spirit company is Canva. Like we want to create the Canva. the same way Canva created the design tool for the masses, we want to create the vibe code tool for the masses. And I think what Canva does extremely well is they make it very easy for anyone to like make a design. And I think there's ways that they could actually make it easier. But the point is that vibe code, I think that the existing vibe coding tools are like straddling the world of like developers and the world of vibe coders. And I think like like for instance, lovable like announces a. super base integration. Like, no one knows what that means. Like, yes, on Twitter. And so they're, they're kind of catering to their existing audience, which is developers who want to be more efficient,

but what if you created a company where you only catered towards the masses and you focused only on serving people who are vibe coders? I think you can move significantly, you can move much faster. You can build solutions that solve 99% of problems instead of like custom building your software for edge cases, like, oh, you can use this database or this database, and then all of a sudden, all of your work as a code builder gets more complicated. That's one thing. But in terms of the use of product, like, I want to build an app that allows you to vibe code, and it will offer you suggestions when you need it. So to make it feel like your own vibecoded app, but like always there to like offer you suggestions, and it's very aware of the app that you're building. And we are

starting with mobile apps because I don't think anyone has done it well. No one has made a good mobile app builder. And I think there's so many fun experiences you can create with a mobile app. You just can't quite get on web apps right now. And we will, and everything in the app is React native. So you're going to, like any app you make on mobile, it's very easy to get that on web as well. We just think the best entry point from a company perspective is mobile. And so we're starting with mobile and it's really, really fun, the apps that, like, we spent all last night, just all we did was made apps for three hours and then just talk to our first wave of users, you know, what are they making? How can we make the agent better at making these apps? And like, that's the phase that

we're in. It's just like, focused on making the product better. And that's like all I've been thinking about for the last, you know, week and a half, two weeks. That's awesome. Is there anything that you've created so far just personally or seen created that you just think is super fun and a good example of like a thing that wouldn't have been possible? Or it was just, isn't well suited to the web, but is amazing for mobile. Yeah, I mean, there's a famous app out. Blake Anderson created it. It's Cal AI. And it's this app that allows you to pull out your phone, take a photo of like a Snickers bar, and it'll tell you the exact calories and it will add it to your daily calories in one click. And yeah, I created that in four prompts with the app. And like a

highly designed. It worked after one prompt. And maybe it wasn't perfect with the API. But we were able to create an app like that that was personalized for me. And other apps that people are creating, like one where you can take a photo of someone and it roasts them. Some people are creating content workflows. Some people are, I think that is a huge thing for like productivity. Whenever you do a certain task, there's always questions you ask yourself as before you do them or that is useful for context. And so I've seen a lot of people create apps that like, okay, I'm creating content.

AI takes that in and then asks the user three questions. and then based on those answers, it gives you a clean output in your own tone of voice. You can very easily create voice apps, any apps that generate images. Going back to the API conversation, we want to give users credits so that they can actually use AI features without having to get API keys. And I think once they experience the power of using AI and APIs, then they'll be motivated to go get the API keys. And so I think if we can allow users to use the chat GBT API, GBT40, image model, once it comes out, you can just show up and build design apps on the device.

And I think that'll be super powerful. One of my favorite, favorite small pieces of content from you was, I don't even know the dates exactly, but maybe a couple months ago, you tweeted or X or whatever the hell we call it, how you tweeted, you know, basically a Google search results comparison of prompt engineering versus. vibe coding and you were like gonna come back and check on this in a year and I think it was like three weeks later like literally three weeks later or something like that where the the lines it actually crossed and I think you updated it maybe a week or two ago and it just had continued to sort of of you know a vibe coding going to the stratosphere compared to prompt engineering yeah not even just prompt engineering look up vibe coding and prompting like just prompting which is just like a more common term than prompt engineering and it's vibe coding is past prompting I just had the

intuition, though, that this was going to be a massive trend, I did not think it would grow this fast, honestly. Yeah, it's really fascinating. There's a whole separate, like, PhD thesis to be written about the resonance of the term, you know, that Andre introduced, which is it even, like, we've kind of taken it and used it in, I think, a slightly different way than he even intended with that original post. But there's something so resonant about the sort of vibe idea that even though he wasn't talking about just referring to people who don't code, being able to code, it feels so invitational and accessible to people. It's very rare that a description of something hits that people so intuitively just latch on to that it works for them. But I think that's a piece of it.

But anyways, the question that I had for you is zooming out then, you know, three months, six months, 12 months from now, how different do you think vibe coding looks, let's call it six months and a year from now, as opposed to what it is now. And what do you think the biggest differences are? So first of all, some things that I will make sure happen is that there are tools that make it 10 times more accessible. So those are the things that I can control and make it more fun and more vibey.

Because I 100% agree with what you said about the approachable nature of this term vibe coding. I think there's something about the word vibe that is really resonant. I think it's the word of the decade, if I'm being honest. Like something about the word vibes, I don't think it's something I heard like growing up in like junior high, like in the, you know, early 2010s. And for some reason, there's something very resonant about the word vibe.

In general, I think everything is based on vibes, like just like the way people create content, people latch onto people's vibe on social media. I think, I don't know. There's something we can talk about that forever. That's the first thing. And so, yeah, I want to make vibe coding more vibe. And that means reducing errors. And that also means making it more fun and making it more intuitive and making it like a collaborative process with what feels like your friend.

Like you and a buddy, which is an AI building an app. And I think that is ideally the end state of the vibe code app is just like the best AI agent for anyone to create apps and have fun creating apps or experiences with code is what I call it. So that's what I can control. The other things that I think other companies are going to do is just without the obvious answer of like these models will get better, cursor will obviously get better, windsurf will get better. I think that there's going to be, we're coming up on a time where I think we're going to see some advantages of scale. Like I think there's going to be prompts that you can run that cost $100, which basically look at your entire code base, look for bugs and iteratively debug your app. I think Manus was a huge eye-opener for me to see how long an AI agent could work and have like kind of an overall understanding of the full mission. And so I think the longer those loops get, and Andrew Yang actually retweeted a post a few weeks ago talking about how like every period of time, the length of time that AI agents can work is increasing.

And so I think we're going to see who's super long task AI agents in the coding space that you're going to be. to be able to accomplish a lot more and do a lot of the things that the vibe coding cynics say you're not going to be able to do, like get rid of bugs or do proper security rules. Because at the end of the day, they are deterministic lists. There is a list of things you should do to make your software secure. And I see zero evidence that AI won't be able to do that. I think the issue is context, which we're seeing get solved, and also just like macro understanding of a larger task, which we're seeing being.

solved. And so I think companies, it's going to be a no-brainer to spend $100, $200 on a single task you give in AI. And so I think we might see, we'll call it the Devin shift, because Devin obviously has, it feels like a longer workflow. And yeah, I think that'll be a massive thing. Awesome. No, this is so fun. Let's make sure to come back in six months or 12 months or whatever it is to check in on some of this. In the meantime, for people who want to follow along, who want to check out vibe code, where can they find you or the app. Yeah. So on X, VibeCode app, if you want to go to our website, right now we have early access.

So vibecode app.com. We had way more demand than we anticipated, which is why it's been longer for us to roll it out to more test users because we have thousands of signups. But we are trying to get it out as quickly as possible. If you want to follow me, just, I don't even know what platform to, I would say, say I am most myself on X. I, that's where I just, I don't even think before I post. That is my notes taking app and where I post my thoughts as soon as I have them. So follow me on X. It's Riley Brown underscore AI. And if you want to watch my videos, you can find me on YouTube as well.

Awesome. All right, Riley, well, thanks so much for hanging out today. Love the, love the work. I'm glad you're doing this. And I can't wait to see what you guys built. Appreciate it. This is awesome. Thank you.