Vibe Coding: A Reluctant Love Story
I heard about vibe coding on a podcast. It seemed like science fiction. It seemed like hype. It didn’t seem like something that could change the course of my career and potentially my life. But it just might. I’ll explain.
Ok, not quite love at first sight. But amazing nonetheless to a non-tech novice like me.
First and foremost, you have to understand the level of computer coding skills I’m bringing to the table. Here’s the resume:
I learned how to do a simple Excel macro in a class in college, and promptly forgot everything about it. (I was a Psychology major, so not much applicability later on.)
I am very good at pivot tables and Vlookups (for an HR guy)
I can string together enough formulas in Excel to do cool things, but in a pretty inefficient way.
I was 10 when we got our first computer at home and I promptly figured out how to change the sounds when certain actions occurred on the computer. My favorite examples were on start-up (Gene Wilder in Young Frankenstein shouting “It’s alive!!!” and shut down (Billy Crystal in The Princess Bride saying “Bye bye boys, have fun storming the castle!”)
And…that’s it. That’s my full computer coding literacy and skill. To put it another way, I have practically none.
So when I talk about vibe coding, I’m not coming from the perspective of an engineer or computer scientist. I’m not coming from a hobbyist that has done some low key experimenting. I know literally nothing.
That’s why vibe coding is so incredibly powerful, especially for startup founders and small business owners. It opens a world of possibility that was until very recently only available to the highly skilled (coders) or the highly wealthy (with enough money to pay for said coders). So let’s talk about what it is, how I learned it, and why its so impactful. Below I will share my opinion and more specifically, my personal experience. I am not an expert. I hope people in the comments actually add some color and detail. But I want to share what I personally have seen and learned.
What is vibe coding?
The simplest way I can explain vibe coding is that it works like an AI chatbot that takes your inputs and creates an app, website, or program based on what you tell it you want. You can give ChatGPT some ideas for a poem and it will write one. Prompt Claude with a song concept and genre and it’ll put together something pretty solid. That’s what vibe coding is essentially, but the output isn’t an email or ballad, it’s an app. You talk to the AI coder(s), through something like Replit or Antigravity or one of many others, and you describe what you want. Like prompting ChatGPT, the more detail and specificity, the better. It takes your inputs, researches options, makes a LOT of assumptions, and builds you something you can see and click and navigate. Then you iterate. Sometimes the AI will ask you to test it, tell it what you like and don’t like. Sometimes it’ll be eager to launch it because it thinks it did a great job and it’s production ready (it’s not). But it generally works, or at least presents a design you can react to and improve.
You tell it you want an app that reminds you seven times a day to do jumping jacks. And to always include pictures of kangaroos. And to add a funny quote each time. You call it Kanga-Jump-aLaugh-aRoo. It makes the app. Ready to go. For basically $0. Something simple, but impossible for a non-coder to create on their own.
Why is vibe coding so valuable to entrepreneurs?
The value to an entrepreneur isn’t necessarily the speed at which an app can be built, nor the quality of the product itself. The value is in the ability to quickly and effectively prototype an idea. That’s the magic. That’s the power vibe coding unlocks. Here’s an example:
How an entrepreneur would have done things 12 months ago:
You have an idea for an app. You really believe in it. Believe it has the ability to be a highly successful, highly profitable business. You would build this thing now if you could. But you can’t. You’re not a capable coder or programmer. You need someone (or a team) to build it so you can launch it, sell it, and prove the potential of this product. So you need to hire that someone to build it.
If you don’t have the money to pay salary or development fees, you have to borrow it. From an investor, a bank, even a friend. But before they hand over the substantial sum of money required, they need to at least “see” what I’m talking about. So you get to work. On PowerPoint slides. You mock up the app, write out workflows and use cases. You project where users will spend their time and eventually, their money. You do everything you can to help the investor visualize what it can become, if you just had the capital to hire the people to build the thing.
This takes a lot of effort and time. And it’s a hard hill to climb. But many do. They secure the funds. What next?
Well, then you hire the brains to build the thing. And you have to explain all over again what you’re envisioning and what it needs to be. And that takes time. Because they ask 100 times more detail than the investor. Because they have to write every single line of code. To build the app takes a monumental amount of detail and decision-making. Debug. Iterate. Improve. Enhance. So you have to translate what’s in your mind to the team. Sometimes you have to wait until they are free for a meeting. Sometimes they are on the other side of the world and you’ve got to talk to them at 11pm your time, 9am their time. Constant back and forth. It’s slow. It’s painful. And it relies on you being capable of perfectly articulating the vision, and them fully understanding it.
Then, finally, you have a product in your hands. So now you can test it yourself. See what works, what doesn’t. See what you forgot about or want to add now. And you go around and around again, getting it right.
Then eventually, you’ve got your MVP, the Minimum Viable Product. So you hand it to your first batch of testers. And they poke holes in it. They break it. They tell you what they love, what they hate, and what’s missing. All that feedback is data. It’s important. And you take all of that, and you go back to the builders, and you get this thing better and better.
After months of work, hundreds of thousands of dollars (if not more), and an ocean worth of emails and Slack chats, your app is ready for public launch. Best of luck. I truly hope your idea was the right one. You’re about to find out.
That is so much time and money and risk, just to take an idea and make it a physical thing. And that was the only way to do it. Until now.
How I would do it today
I have an idea for an app. I haven’t fully thought through all the features or the workflows, or even simple things like the interface colors and font and experience. But I have an idea. I want to build this thing and then use that to further ideate and design. So I do.
You know all those steps I listed above? Yeah, I skip most of them. I don’t make PowerPoint decks. I don’t hunt for funding. I don’t explain to a team of coders over and over again what exactly I want. I don’t wait for user testing to tell me how the app functions and flows.
I just build it. I go on Replit (replit.com). I create an account and subscribe for $25 a month (or even free at first). And I describe to it what I want. I design the thing in real time, add my ideas, and watch the idea take it and build what I describe. It’s not perfect at first. Not functional. But it’s pretty dramatic how far it can take my first idea. So then I click around. I see what I like and what I don’t. I ask myself “what do I want to happen next when I click this?” because I hadn’t really gotten that far in thinking it through. And I keep talking to the Replit agent.
Sometimes it asks me direct questions. Sometimes it says to test it out and tell me what’s next. Sometimes it takes a few minutes to really think through what I’ve asked for (thinking mode engaged) and sometimes it’s really quick and easy (change the reds to blues).
Now I have something in my hand, on my computer and on my phone. I can test it. I can play with it. I can hand it to other people to do the same. It’s not a concept on a page. It’s a real thing you can interact with. That’s powerful.
And if you’re wrong, if your hypothesis was off, so be it. Think of a new one and try again. You’re paying $25 a month. Go crazy. Create all you want.
How did I learn how to vibe code?
Practice and play. I would think of an idea. Something I needed to make my life a little easier. And then I just wrote that in Replit and asked me to build an app. And then I did it over and over again. My first app, I spent 20 hours over the course of a couple days. Because I just kept getting ideas, kept wanting to improve and tweak and try new things. And that was 6 months ago, when the tech was dumber and slower. And I was dumber and slower. One has improved rapidly. And I’m trying.
Now the ability to rapidly prototype and build, with both the improvements in tech and my experience, has reduced that time even more. Probably by half, at least. But it takes practice.
When did you know this was really a big deal? What was your “ah ha” moment?
There were two moments that really had a big impact on me with regards to the potential this vibe coding thing could unlock. One was with a mutual acquaintance at Microsoft. Another was with my 9 year old.
The Microsoft engineer
After my first ever vibe coding deep dive, I showed it to a friend. He was impressed. So we gave a demo to a solutions engineer we knew at Microsoft. After walking him through the app and the functionality, he asked me three direct questions, to which I gave three direct answers:
Q: “How long did it take you to make this?”
A: “20 hours.”
Q: “How much coding experience do you have?”
A: “Is negative zero an answer? Because that’s my answer.”
Q: “How much did this cost you to build?”
A: “About $25.”
He looked genuinely shocked. He said to build this thing would have taken 3–6 weeks and cost at least $100,000. And I did it, as an absolute novice, 20 hours and $25. That’s when I knew I had something powerful. I could skip the whole early stage prototyping and explaining steps and just build it.
The 9 year old
I had built a working app, but I wasn’t sure how well it would really work. Me knowing exactly what I wanted and navigating what I designed is one thing. Building something and handing it to another user is a whole different story. I need to test this out. And that’s when my 9 year old’s messy room gave me an idea.
My 9 year old is a brilliant, wonderful kid. They also are very disorganized and messy. Cleaning their room is a constant battle. And it had reached a point where we needed to get rid of some things. So I did the thing that annoys my wife and child alike. I went and got BOXES.
I labelled one as Keep, one as Sell, and one as Give. If it was something she wanted, it goes in Keep. If it’s for the neighborhood garage sale, Sell. And for Goodwill? Donate. Simple. Messy and cluttered, but simple.
That’s when I thought about creating an app to manage this instead.
I hopped on Replit and explained what I wanted. I wanted her to take a picture of an item, label it, and then categorize it. So it built it. It worked. But it then left the next step incomplete. What happens after each item was labelled? Well, it depends.
If it goes in Keep, where does it go? What room? So I asked Replit to build icons of rooms in the house (bedroom, basement, play room, etc.) and give the option to select one for each item.
If it goes in Sell, how much would it be listed for at the garage sale? So Replit created that option. Hit Sell, then enter a price.
And if Donate? Click the button and leave it as is.
I had the physical boxes turned into virtual ones. But then I realized as a parent, I needed to know what the kid had put where. Some things were definitely not donate (that teddy bear from grandma stays). Some were way overpriced (sorry, I can’t sign off on $55 for a used Bubble Yum container).
So I asked Replit to build me a parent login portal in addition to the kid one. Now I saw everything my kid had put in there. I could press a button and see the entire inventory. I called it the Kidventory. I was even playing with marketing and branding ideas just because I could prototype things real time.
I ended up spending 4 hours on this. Building a login feature for parents and kids alike. Fixing up some of the photo capture functionality, etc. But I built it.
So then it was time for the real task. I loaded it onto the 9 year old’s tablet, told them what the app was, and said go and do it. I didn’t give any more instruction than showing one time how to take a picture and label it. Halfway through I was waved away. The kid understood the assignment.
A couple hours later, I had a 30 item “kidventory” in my app. And I had a child that had done what I asked and had fun doing it. The idea wasn’t brilliant (feel free to steal this and give me 50% of your profits). But the ability to prototype something that quickly, iterate and improve it before “launch”, and then hand it to a user with little instruction was absolutely mind blowing. And powerful. And fun.
That’s why vibe coding is such a valuable, impactful tool.
I am, by nature, a creative experimenter. I like to try things. Throw ideas out there. Iterate quick. Fail fast. All that. I’ve had ideas that required technical solutions and always just shut them down because I didn’t have the capability or capital to make it a reality. But now I do. I don’t have that barrier of entry. I don’t have the gatekeeper. I can do what I need to do to put my vision into reality. To build something real.
It’s not perfect. Far from it. You still need professionals to build things that are scalable and secure and truly great. But if you ask me to choose between building a slide deck to present an idea, or building an app to truly show what I mean, well, it’s not even close.
So yeah, vibe coding is a big deal. It unlocks so much possibility if you are willing to try. And if you’re an entrepreneur, starting something or building upon what you already have, it’s worth the investment. Trust me.
