Vibe coding 101 · October 19, 2026
What Is Vibe Coding? A Plain-English Guide for Business Owners
Vibe coding lets you build software by describing it in plain English. What it is, where it came from, and what it can and cannot do for a service business.
You run a service business. You have an idea for a tool your clients would use: an instant quote calculator, an intake assistant, a portal that answers the questions your team answers fifty times a week. A year ago, building it meant hiring developers. Now you keep hearing that you can just describe it to an AI and it gets built.
That is vibe coding. This guide explains what it is, where the term came from, what it is actually good for, and where it stops being enough. No jargon you will need to look up.
Where the term came from
In February 2025, Andrej Karpathy, a co-founder of OpenAI and the former head of AI at Tesla, posted a short note describing a new way he had started building software. He called it vibe coding: you "fully give in to the vibes" and "forget that the code even exists."
His workflow was simple. He talked to an AI coding tool, often by voice. He accepted every change it suggested without reading it. When something broke, he pasted the error message back in and let the AI fix it. The software got built without him writing, or even reading, much of the code.
The term caught on fast. By November 2025, Collins Dictionary had named "vibe coding" its word of the year. Tools built around the idea, like Lovable, Bolt, Replit and v0, grew into businesses with millions of users. Lovable alone reported more than $600 million in annualized revenue by September 2026.
The part of Karpathy's post that gets quoted less: he said it was "not too bad for throwaway weekend projects." The person who coined the term was describing a way to build things that do not matter much if they break. Keep that in mind. It comes up again below.
What vibe coding actually is
The clearest definition comes from the developer and writer Simon Willison. In his words, vibe coding means building software with an AI "without reviewing the code it writes."
That last part is the important bit. Developers have used AI to help them write code for years. If you read what the AI produces, test it, and understand it, that is just software development with a faster assistant. Vibe coding is specifically the version where you do not look under the hood. You judge the result by whether it seems to work.
For a business owner without a technical background, that is exactly the appeal. You never needed to read the code. You needed the thing to exist.
How it works in practice
Every vibe coding tool follows roughly the same loop:
- You describe what you want. In plain English. "Build a page where a homeowner enters their square footage, number of bathrooms and how often they want cleaning, and gets an instant price."
- The AI writes the software. It produces the screens, the logic, and often a database to store what people enter.
- You look at the result and react. "Make the price update as they type. Add a button to book a first visit. Use our brand colors."
- The AI changes it. You repeat until it looks right.
The tools differ in how much they handle for you. Some, like Lovable and Bolt, run entirely in your browser and can put your app on the internet with one click. Others, like Claude Code and Cursor, work with files on your computer and give you more control, with a steeper learning curve. We compare them in detail in Lovable vs Bolt vs Replit vs v0 vs Claude Code.
What a service business can realistically build this way
Vibe coding is genuinely good at some things. Be honest with yourself about which category your idea falls into.
It is excellent for prototypes. If you want to show your partners, your best client or an investor what your idea looks like, you can have a clickable version in an afternoon. A law firm can mock up a client intake flow. A logistics company can show what a shipment tracking page for customers would feel like. Nothing beats a working demo for getting people to understand an idea.
It is good for internal tools. A dashboard your team uses to see which jobs are overdue, a calculator that estimates project hours, a page that turns call notes into a summary. If it breaks, your own staff notice and work around it. The stakes are low.
It is risky for anything clients rely on. The moment real customers log in, enter personal details, pay you money or depend on the tool to do their own work, you are in a different situation. That is where most vibe-coded projects run into trouble, and it is exactly the category most service businesses actually want to build.
Where it gets hard
The most-shared reality check on vibe coding came from Karpathy himself. A few months after coining the term, he tried to ship a small real app, MenuGen, which turns a photo of a restaurant menu into pictures of each dish. He wrote up the experience. The demo on his own computer came together quickly. Then he tried to put it online with logins and payments.
His summary: he felt 80% done, but it was "a bit closer to 20%." He spent most of his time not on the app itself but on hosting, user accounts, payment setup and configuration. His conclusion was that vibe coding full web apps "is kind of messy and not a good idea for anything of actual importance."
That is someone with decades of engineering experience. The same gap shows up for everyone. Addy Osmani, an engineering lead at Google, calls it the 70% problem: AI gets non-engineers about 70% of the way, fast, and the last 30% still takes real engineering knowledge. We dig into what lives in that last 30% in Your AI prototype is 20% done.
The specific things that tend to go wrong:
- Security. AI-built apps regularly expose the data they store. In one 2025 study, researchers found that 170 of 1,645 apps showcased on Lovable let anyone read or change their data. For a clinic or an accounting firm, that is not a bug. It is a breach. See our security checklist.
- The fix loop. You ask the AI to fix one thing and it breaks another. Without understanding the code, you cannot tell which change caused it.
- Costs. Most tools charge per message or per unit of work. Fixing problems often costs more than building the first version. See what vibe coding really costs.
- Everything around the app. Hosting, domains, logins, payments, email. None of it is hard individually. Together, it is where weekends disappear. We explain each piece in Where does a vibe-coded app actually live?.
The term is already changing
In early 2026, on the first anniversary of his original post, Karpathy suggested vibe coding had become "passé" and proposed a new term: agentic engineering. His framing is useful. Vibe coding "raises the floor": almost anyone can now build something that works on a good day. Agentic engineering, where people who understand software direct AI agents with careful plans, reviews and tests, "raises the ceiling."
For you, the practical meaning is this. The floor has risen for everyone. Your competitors can vibe code the same prototype you can, this afternoon, for the same subscription fee. What separates a toy from a product people pay for is still the part that needs judgment: what to build, how to make it safe, and how to get it in front of customers.
How to start, if you want to try
If you want to try it yourself, and you should, here is a sensible first week:
- Pick something small and low-stakes. An internal calculator or a prototype for a client-facing idea. Nothing that stores personal data yet.
- Write down what it should do before you open any tool. One page. Who uses it, what they do, what they get. Our guide to writing a PRD before you prompt has a template.
- Use a browser-based tool first. Lovable or Bolt will get you to something visible fastest.
- Show it to three real users. A colleague, a friendly client, someone who would actually use it. Watch what they do.
- Decide what it is. If it stays internal, you may be done. If clients will rely on it, treat the prototype as a sketch, not a foundation.
The question that matters more than the tool
Most of the searching people do around vibe coding is about tools: Lovable or Bolt, Claude Code or Cursor, Vercel or Netlify. Those questions have answers, and we cover them across this blog. But they are the second question.
The first is what to build. Automating your back office with AI is worth doing. It is cheap and it works, and so every competitor you have will do it too. The tools that actually move your position in the market are the ones your customers touch, the ones you can price and sell, and the ones nobody could have built a year ago.
That is the work we do at Veyra. We start with a working prototype you can click through before any money changes hands, so you see the idea for real before deciding anything. If you want to talk through what your business could offer that it could not before, book a call. And if you would rather build it yourself, the rest of this blog is written to help you do that well.