Neurocourse

What vibe coding is and why it works

Vibe coding means building sites and apps through a dialogue with AI: you describe the product in words, the model writes the code, you look at the result and request changes. This lesson covers how the process works, which tools to pick and what you can realistically build with zero experience.

Not long ago, the gap between "I have an idea" and "I have a website" was months of learning to program. Now the gap is a conversation.

What the process looks like

  1. You describe it: "Build a landing page for a yoga studio: schedule, prices, a booking form, calm warm colours."
  2. The AI builds it: a minute later you have a working page — not a picture, a real site.
  3. You iterate: "Move the form higher, add a testimonials block" — and on until it's right.

The code is still there. The AI just writes it, while you play client and project lead.

Your toolkit

  • Lovable — describe the product, get an app with a database and sign-in. The friendliest way in; our first projects live here.
  • Replit — an AI agent writes code and runs it in the cloud right away. For the trickier stuff.
  • ChatGPT / Claude — pages and games generated straight in the chat. Good for quick experiments.

What you can actually build (thousands of beginners have)

A landing page in an evening. A browser game over a weekend. An app with forms and a database after a week or two of practice. A product that takes payments once you have a couple of projects behind you.

Where the phrase came from

"Vibe coding" was coined by Andrej Karpathy — an OpenAI co-founder and Tesla's former director of AI. In February 2025 he posted a short note online: there's a new way to program, he wrote, where "you fully give in to the vibes and forget that the code even exists" and simply talk to the model. The phrase spread in days: by spring 2025 it was on T-shirts, and the Merriam-Webster editors added "vibe coding" to their words-to-watch list. In six months, one engineer's joke turned into the way thousands of people with no CS degree ship working products.

A real example: Pieter Levels and his flight simulator

On 22 February 2025 the indie founder Pieter Levels (@levelsio online) typed one request into Cursor: make a 3D flying game in the browser. He had never built a game before. A playable version existed in about three hours. He put it online as fly.pieter.com and kept building it live.

Then it got away from him: Elon Musk retweeted it, thousands of people piled in, and 17 days later the thing was doing $1M a year. One person, no team, not a line of code written by hand.

Don't plan for that ending — it's rare, and ten years of his reputation sit behind it. The part that matters is the other one: going from “I want a game” to a working game took an evening. A year earlier that was a team and six months.

2025 was full of stories like this: journalists and hobbyists raced to build apps in a single evening and posted the videos. One person made a habit tracker for himself and his wife over a weekend. Another built a menu generator for a small coffee shop. Another made a platformer his eight-year-old actually played. Not one of them typed a line of code by hand. They all had one thing in common: they knew exactly who the product was for and why — and handed the "how" to the machine.

Common myths

  • "This isn't serious — real products aren't made this way." They are. Startups already ship first versions on these tools, show them to investors and early customers, and only rebuild "properly" once scale forces the issue.
  • "The AI will make a mistake and I won't catch it." You will, because you're looking at the live result, not the code: the button doesn't click, the form doesn't send, the text is misaligned. You check it with your eyes, like any user would.
  • "Learn to program first, then vibe code." It works the other way round. Plenty of people finally understand what a database or sign-in actually is by building a project — a real example beats dry theory.

A Socratic pause

if the AI writes the code, what exactly is your job, and what is a client paying you for? The answer's in the next paragraph — but put your own into words first.

The big shift in your head

Your value isn't writing code, it's clarity about what you're making: who the user is, what they need to do, how it should look. That's the skill this whole course trains — let the machine handle the code. Which is why we don't start with tools. We start with a brief: a short product description you'll learn to write so the AI lands close to your intent on the first try. That's the next lesson, and it saves you dozens of edits.

Why now is the moment to learn this

Explaining a task to a machine and carrying the result through to a finished product went from exotic to basic in 2026, roughly the way spreadsheets did once. Employers increasingly aren't hiring "a developer" or "a designer" — they want someone who takes an idea all the way to a working site or app on their own. And the best part: the barrier has never been lower. In a couple of years half the room will be able to do this. Learn it now, while it still buys you a head start. Later in the course you'll build a real product with your own hands — or rather, your own words.

Practice · 4 задачи

Short questions on the lesson — with an explanation for every answer.