Interactive Lesson Experience

The Vibe Coding Landscape

Vibe coding has gone mainstream. This lesson turns the original presentation into an explorable website so learners can quickly understand the tools, compare categories, and see when each approach makes sense.

$1B+ industry segment within months
25% of YC W25 startups with codebases 95% AI-generated
4 major categories of vibe coding tool
Plain English → Software

Describe what you want, and AI writes the code, scaffolding, and often deployment flow for you.

From Idea to App

Web app builders can assemble frontend, backend, authentication, and database layers from a single prompt.

Beyond ChatGPT Alone

The strongest vibe coders know which tool to reach for — and when — depending on speed, budget, skill, and project complexity.

Part 1

Four Types of Vibe Coding Tool

Not all AI coding tools solve the same problem. Some are best for non-developers, some for professionals inside an editor, and some for technical power users who want control and efficiency.

AI-Native IDEs

Editors built around AI

These tools understand your whole codebase rather than just the file in front of you.

  • Best for: professional developers
  • Strength: project-wide understanding
  • Examples: Cursor, Windsurf
IDE Extensions

AI inside tools you already use

Broad compatibility and low switching cost. Ideal when you want AI help without changing your workflow completely.

  • Best for: existing editor users
  • Strength: familiarity and flexibility
  • Examples: GitHub Copilot, Cline
Web App Builders

Browser-based app creation

Describe the application in plain English and the platform builds the frontend, backend, and database.

  • Best for: non-developers and fast prototypes
  • Strength: lowest barrier to entry
  • Examples: Bolt.new, Lovable, Replit
Terminal / CLI

Command-line AI agents

Powerful, often very cost-effective tools for users comfortable working directly in the terminal.

  • Best for: power users
  • Strength: efficiency and cost control
  • Examples: Claude Code, Aider
Part 2

Who Should Use What?

Use this section to match the tool category to the learner. It reframes the lesson from a list of products into a practical decision-making guide.

Non-developers

If you want to build a real app without writing code, start with web app builders. Bolt.new and Lovable can create full-stack applications from a single paragraph.

Developers

If you already work in VS Code or a similar environment, AI-native IDEs and extensions can transform your workflow and speed up multi-file work.

Budget-conscious builders

If cost matters, open-source options such as Cline and Aider can be extremely affordable because you typically only pay model usage costs.

Part 3

Two Interactive Examples

Switch between the examples below to see how the lesson moves from abstract tool categories to realistic scenarios, prompts, and outcomes.

Web App Builders

Build a consultancy client portal in an afternoon

A small business owner needs a client portal where people can log in, view active projects, see invoices, and download deliverables. Existing SaaS tools are too generic or too expensive, and custom development would take weeks.

1

The Problem

They need something purpose-built for their business, not a bloated subscription product or a costly bespoke build.

2

The Prompt

A single plain-English brief outlines the login flow, dashboard, invoices, file downloads, and visual style.

3

The Result

The platform produces a full application with authentication, database, and polished UI in minutes.

The Bolt.new Prompt

Build me a client portal web app. It needs: - A login page for clients - A dashboard showing their active projects and recent invoices - An invoice table with amount, date, and status (paid / outstanding) - A file download section for project deliverables Clean, professional design. Blue and white colour scheme.
Secure login
Individual client accounts and authentication flow
Live dashboard
Project status and invoice visibility from a real database
File library
Deliverables presented with download access
Instant deployment
Built and published without installation or a traditional dev process
AI-Powered IDE

Upgrade a whole codebase, not just one file

A developer needs to migrate a project from SQLite to PostgreSQL across more than 50 files. Manually, this means days of careful refactoring and retesting.

1

The Problem

The change touches many parts of the system, so isolated code suggestions are not enough.

2

The Prompt

Agent Mode is asked to update database queries, adjust syntax, add error handling, and flag manual review points.

3

The Result

Cursor edits 53 files, runs tests, and fixes the errors it introduced, compressing days of work into hours.

The Cursor Prompt

In Cursor's Agent Mode: "I'm migrating this project from SQLite to PostgreSQL. Update all database queries across the codebase to use the new connection string and syntax. Add proper error handling, and flag anything that needs manual review."
Codebase indexing
Builds a project-level map so the AI understands how files connect
Multi-file edits
Changes dozens of files consistently instead of one snippet at a time
Test execution
Runs the test suite and self-corrects introduced issues
Practical outcome
Roughly three days of work reduced to around two hours
Part 4

How to Choose Your Tool

This decision framework keeps the lesson practical. Learners can start from their own experience level, budget, and working style rather than chasing hype.

Bolt.new and Lovable offer one of the lowest barriers to entry. You describe the app in natural language and the platform assembles the experience for you. Both offer free tiers suitable for small experiments.

For developers who want a mainstream, affordable starting point inside an existing editor, Copilot is a simple on-ramp. It lowers switching friction and helps you test whether AI-assisted coding fits your process.

Almost every major tool now has a free tier. The fastest way to develop intuition is to build a small, familiar project in two different tools and compare quality, speed, and control.

Part 5

The Vibe Coding Hangover

Speed is not the same as quality. This section gives the lesson necessary balance by surfacing the risks that come with blindly trusting generated code.

A December 2025 analysis found that AI-generated code contained 1.7× more major issues than human-written code, including 75% more logic errors and 2.74× more security vulnerabilities. These tools are extraordinary accelerators, but they are not a replacement for understanding what you are building.

Why teams get caught out

Organisations build faster than ever, then discover maintenance debt, security weaknesses, and code nobody fully understands.

The right mindset

Use AI to go faster, widen your ambition, and reduce friction — but keep human review, testing, and architectural understanding in the loop.

Part 6

See It in Action

The best next step is not to memorise a list of tools — it is to build. Choose one small project you understand clearly, try it in one platform, then compare the result with another.