The Big Vibe Code Lie

That smell. What is that smell? That's the smell of These are the things that I wish I knew a little bit more. I don't know, like a year ago. But also something that I think most folks who don't make these videos haven't ever been a creator and haven't had seen the amount of inbound that you get from companies. They just wouldn't know what's going on. So, I wanted to expose it. Not expose, but I wanted to at least highlight it. And I think the world of AI engineering is kind of waking up to this where there's a lot of traps that you can fall into. So I hope by watching this video you'll avoid some of them and it will be helpful for you. What am I talking about? Well, if you want to have zero nuance, you can just read this. But the idea is there are lots of claims all over the internet about how easy it is to build something. And so you'll notice this on X and on YouTube where there's kind of this like paradox going on where they call it the yap to ship ratio. So you have shipping and you have yapping. Yapping. If you yap 10 times for every one time that you ship or push a commit or anything, then you are very much so skewed on the side of yapping. And the thing is is this is a direct tie to your time. So if you're spending all your time yapping and not actually shipping, then you really aren't that qualified to talk about things that work for shipping. So why do I mention all that? Because when you go onto a platform, there's a really good chance that if someone is talking up a company a lot, like I don't know, every other day or even if they're not explicitly mentioning that the video has some sort of like monetary tie to it, chances are that's happening. So, I just wanted to highlight that as someone who made a video every single day for I don't know like 40 days in a row. didn't take a long time to get a ton of different inbound. And it's not all bad, right? It's a cool thing how it works where if you are creating and the company is someone that you already like, that you already work with, that's cool. But that's not the case for most of the YouTube videos that are out there. So, you have it's it's kind of like disingenuous and it's kind of a crazy thing. I want to just show like I know this story because I made these videos and I stopped making these videos because of this problem where you go and you make a workflow video and it assumes that the person that's watching it already kind of knows how to do this without AI. That's at least how I do it. And all the projects that you're doing in these videos, they're very very small. So, it's not something that is going to have, I don't know, like payments. You need to make money, right? You need to have payments. You shouldn't try to oneshot that with AI. You look at all the best engineers that have a better yap to ship ratio. They're talking less. They're shipping more. They have a product that's doing real dollars per month and paying for their bills. They're not going to go and put the AI on doing that. So, that's just one of the big traps. And so just have a good filter. When you go on to YouTube and you see someone that has a giant audience, chances are that there's a little bit, it's a little bit misguided. That's all. And the second thing I wanted to kind of address is this anxiety that a lot of people get around the new tool or, oh, I'm going to get left behind. I'm going to get left behind. The amount of times if you had a dollar for every time that an AI bro said, you're going to get left behind, you'd literally be Jeff Bezos. It's the most cringe phrase. Now, it's partially true. Doesn't mean that it's not, you know, at all true because you should be using these tools. They should be augmenting your workflow. You should know when to take the wheel, when not to take the wheel. But scare tacticing, like uh I don't know, fearmongering, getting people scared that they have to use this thing. I don't think it's like a great way to do it. And it just shows you the highlights. Like there's really good demos. Like I had one video where I was doing a client project and this is literally like a glorified Squarespace website. It was in Next.js JS no backend all like let's put the blog together and I used a tool won't name it but I used a tool in that video and it pulled like 30,000 views and I don't know it was crazy because I also run a community no it's not for you it's for people that actually build software I'm not disqualifying you I'm not trying to pull tactic but I'm not trying to plug it either anyways I had a community I still have a community it's for builders but all these people that watch the video they're like ooh Ooh, that's the workflow guy. Ooh, let me join. Ooh, what do you mean I have to learn how to code? And it's just this crazy thing. And I see it in the marketing for all the Vibe coding tools. There's different classes of these things. I think that if you look at like I don't know, let's get let's get some drawings going. I think that there's different classes of tool where if you have like let's let's call it like some stairs. Let's make some stairs. So, we'll put like V0 here. We're going to put bolt here. We're going to put lovable here. Actually, Lovable maybe is like a step down. But anyways, some of the marketing on this like I have uh my wife's friend like hit us up and I was like, "Oh, I want to make an app." And we're like, "Okay, cool. You should go use like Figma make or something and try to make the interface for it." She comes back and she's like, "I have the entire app. I build it in lovable." Because the marketing site says that you can type in a prompt and get your app in one shot, including payments, including like all this stuff. And it's such a ridiculous promise. Like, think about that. If you're tied up in this idea of, I've always wanted to build an app, but I just don't know how to code it. Like, if I just knew how to code, if I had this app, like this thing, it would it would get me out of my job, I would feel so good about it. And then you see that, you see that pain, you see that like dream getting unleashed on onto your capabilities. And you go and you do it. And then you, I don't know, you get a link that then you share with your friend. It's not your app. It says like the same thing in iMessage when you go to share, which is like type in a prompt and get your app in a weekend. That's not true. That's literally a lie. And then the person finds out, oh crap, I've made a glorified Squarespace website. I've made a thing that nobody wants to use because it feels awful and I have to actually sign up for lovable. I don't have to sign up, you know, no hate on lovable. I keep naming companies. But it's just this trap that I see and I wish people were more honest. I wish companies were more honest about it really really bothers me. Part of like why I stopped making videos all the time. But another reason was I was just trying to get these numbers up the ship uh numbers. Now I'm looking back at the board. But those are the two things. Also just this wall around like okay let's take a real use case. We have people in our group and this I'm like the most qualified person to talk about people who have to hit this like reality wall, right? So let's let's say this is your experience. You have 1, two, and three levels or three years of experience. Those are years. This is experience. So high, medium, low, low, medium, high. Cool. Now, when you get to one year of experience, and we're going to do orange is no AI. No AI. And then you have red is AI. All these people, they get SCOP into the fact that they can like oneshot an app in a weekend and they're going to be the next Zuck. And they come in and they're their experience even after a year isn't even low. It's like right here because they're just waiting for the model to get better and blaming the product for not being good enough to oneshot the app or deliver on the promise that they were sold versus like if you had literally zero AI. I think I flipped these colors. Well, you understand the point. That's no AI. Oh, yeah. No. So, let's just do this then. With AI, you'd have even less. Let's just put it that way. In the same amount of time. And so, that's a big problem that I see that's going on. And when I do funnel people through the channel and they end up learning about like what's going on and this group that I have this little logo that's on all the videos, it's this thing right here. Whoops, wrong way. I originally I had to change like the whole thing because at the start of it, I kind of bought into it too. I only have 3 years of engineering experience on top of like I don't know 15 years of like fiddling around with like front end, but not really. like I was a PM and a designer and like was a script kitty when I was younger. So I only have like three years of like actual every single day working doubles very little AI and I so I'm just like really happy that there wasn't AI when I was learning cuz it like it's such a trick like you you get a little bit of the sauce, you get a little bit of the the dopamine and it's like oof that's awesome but then it'll mess things up. And when people started to join this group that I started, I just had the same thing over and over and over again where they're like, "Okay, making the leap from let's say I make a lovable app." Going from this to actually like deploy your app. This is a huge leap because there's all these little things that you've never even seen or you've never heard about that exist behind the fancy covers that are covering the files. Like you go into the loable and you want to get it into there. I'm sure they have a new thing and it's fantastic and you can connect it to Super Maze and whatnot, but like even if you could do that, then the next step is okay, I want to deploy it. Well, do I actually need an API? What is an API? Then you have this next leap. And it's like, oh, it can't do the one-click thing that it promised me now. I need to go over and I need to learn superbase. And Superbase actually has like all this other stuff in it where it's like, oh, it's DB. Oh, it's Oth. Oh, it's edge functions. Like, what's an edge function? How does OH work versus authorization, authentication, RBAC, multi-tenant support? What are all these things? I don't know what to do. And so, all that said is like all these tools, they are awesome when they are more honest. And I wish they could be more honest with, hey, you need to learn a lot of this stuff that's underneath what's going on in here. And it is, it's crazy because I like it because you can move, you can pull the future forward, so to speak, where the hello world right now is like way better than it used to be, but you don't have that period of time where you're like grinding. And I made this video right now because, you know, one, I had to change the community up where it's like, hey, no, like this is for people who want to actually use the tool to grow their own SAS or they're at a company and they want to build something. They know and they recognize that like, yo, there's a lot of stuff I don't know and I just want to learn it faster. That's a mindset that you have to have. Like, if you just think you can ride the coattails of AI providers and let your brain turn into mush, you're not going to win. And um that's one reason why this is really pertinent right now. And then another is I've just noticed this with my own stuff where it's like if I looked at my tooling over the last 3 years, it's like oh cool like I was building next websites beyond 3 years ago but it was like nothing. It was like oh it's like like grab a template and there's tailwind and you're just building basically a landing page. And then after that it's like oh I'm going to take this serious and the second I want to take it serious and it's like oh I need to have an API that is going to be always on. Oh, you can't do that cuz they don't have web hook support. Like if I need web hooks, what's a web hook? Oh, I need to go learn Express. Oh, what's HANA? Oh, how do I deploy that? And um same thing goes with like Superbase where it's, you know, there's this thing that happens where you go and you use it and you're like, this is awesome. Like there's so much hype about it. It's so cool. All I do is I, you know, I click on the thing and there's this pretty table and it's really nice. But then you learn the tool enough that you realize, wait a second, I don't need to learn all this stuff that is like a layer on top of the actual thing which is Postgress and authentication. And uh it's just a crazy thing. So I encourage everybody to use the tools, get better at knowing when to give it the reigns, when not to give it the reigns. And this is coming from someone who's not perfect, right? I I have these same issues where I set a 30-day challenge. I was trying to get to 10K in MR. And that was 2 days ago. I was supposed to hit that and I'm just now getting to deploy the product because 2 weeks in I realized oh even though I did the research even though the PRD worked perfectly and there were no bugs when I went to actually use the thing I had not had the experience of architecting a system that could do the thing at the scale I wanted to in the performance that I wanted to. I hadn't done that before. So it's like my input to the AI, it's garbage in, garbage out, right? And it didn't press me. Yes, man. Me. So that's really the biggest learning is when you're getting into this, it's trying so hard to just identify like what are the things that I don't know? And then you look those up. And then in my case, if I were to start the project over again, of course, I'd be way off to the races, way faster. But that next time that I want to build something for my product Zeke and it's not something that I've done before, I won't just go get Yesmaned by an AI. Instead, I'll look at the features that I want to build that solve the problems that my customers have and I'll think of what other companies have done something similar that I can learn from and then I can go and do the research. And a lot of people forget that there's an R and R&D. They just think of like the the development part. So that's it for this video. If you guys want to join a bunch of engineers and builders, we have a private club. That's this thing. We have five events a month. It's pretty great. And I'll see you in the next video. Oh, like and subscribe. X.