1.1 Two definitions
The short definition: vibe coding is building software with the help of AI. The long definition: vibe coding is directing capable AI agents as they plan, write, test and deploy code for you. An AI agent is an AI that does not only answer; it makes a plan, takes several steps on its own, and reports back.
Most of it is a conversation: you say what you want in everyday language, the AI writes the Python, and you say what to change. Success depends on how clearly you communicate and how much context you give, not on knowing syntax.
1.2 Most of the work, not all of it
The AI does perhaps 80 to 90 percent of the work. The rest is yours, and it is why this course exists:
| The AI is good at | You still own |
|---|---|
| Writing code that runs | Deciding what to build and for whom |
| Explaining code and error messages | Checking the result is right, not only that it runs |
| Doing setup steps | Keeping secrets such as API keys safe |
| Trying again when something fails | Saving your work and putting it online |
1.3 Why it matters now
Speed: an app that once needed a team and months can be built by one person in hours. Who can build: if you can describe a problem clearly, you can solve it with a small tool of your own, such as turning a weekly spreadsheet chore into a web page.
1.4 Today's finish line and roadmap
Do it watch the instructor open the finished IDX Stock Snapshot on a phone. Type a ticker such as BBCA and it shows the company, its latest closing price, a price chart and a savings calculator, or a friendly message if the ticker does not exist. By the end of the session you will have your own, live: understand (Modules 1 to 3), build one project step by step (4 to 6), save and ship (8, 9), and a bonus look at MCP (10). Module 7 is extra practice for fast finishers and for home.