Beginner: Build a safe, tested first slice from a real user problem.
Bring: No coding background is required. Bring one real problem worth solving.
Level result: A focused project brief, a working first slice, and evidence that the slice solves the chosen user task safely.
- Frame a realistic project around a specific user result.
- Prompt with constraints, tasks, examples, and privacy boundaries.
- Deliver a small working slice with a clear evidence packet.
You can review the curriculum here. A free member account saves lesson work, answers, and progress.
What Vibe Coding Is
Explain the human-and-AI build loop and choose one low-risk project to carry through Level 1.
Build: A Level 1 project card naming one user, one observable result, one boundary, and one failure case.
Sign in to beginWhat Generated Code Is
Inspect generated code as an editable draft rather than a verified product.
Build: A generated-change inspection note with files, behavior, assumptions, evidence, and limits.
Sign in to beginWhat AI Can Misunderstand
Identify unstated requirements, invented details, and confident errors before they enter the project.
Build: A misunderstanding register listing assumptions, consequences, verification sources, and owners.
Sign in to beginChoose a Realistic First Project
Reduce the project to one useful, finishable, low-risk journey.
Build: A project scope decision with included journey, non-goals, time box, and stop rule.
Sign in to beginIdentify the User
Define a reachable primary user, their moment of need, and current workaround.
Build: A one-page user snapshot grounded in observation rather than assumptions.
Sign in to beginDefine the Desired Result
Translate the user’s problem into one observable outcome and success measure.
Build: An outcome statement with baseline, target behavior, evidence source, and failure condition.
Sign in to beginComputer Requirements
Confirm that the device, storage, memory, browser, and access controls fit the planned work.
Build: A readiness inventory with available resources, gaps, workarounds, and purchasing boundary.
Sign in to beginCloud, Local, and Hybrid Options
Choose where AI work runs based on privacy, quality, availability, setup, and cost.
Build: A cloud/local/hybrid decision matrix with test results, data boundary, and fallback.
Sign in to beginSubscription and API Costs
Estimate the visible and usage-based costs of the Level 1 project before committing.
Build: A beginner cost sheet with setup, monthly, usage, limit, and review trigger.
Sign in to beginProtect Private Information
Minimize data collection and keep secrets and private information out of unsafe locations.
Build: A data-boundary sheet listing permitted, prohibited, protected, and deleted information.
Sign in to beginWrite a One-Page Build Brief
Combine the Level 1 evidence into one stable source of truth for the first build.
Build: A one-page brief covering user, result, inputs, rules, risks, tests, and definition of done.
Sign in to beginBreak Features into Tasks
Turn the brief into small, independently testable changes with explicit dependencies.
Build: A ranked task board with thin slices, acceptance evidence, dependencies, and rollback points.
Sign in to beginWrite Clear Prompts
Translate one task into an explicit prompt with boundaries, examples, tests, and requested evidence.
Build: A reusable task prompt that preserves the brief and asks for verifiable output.
Sign in to beginUse Examples Well
Use representative examples to clarify behavior without hiding edge cases or leaking private data.
Build: A safe example set covering ordinary, boundary, invalid, accessibility, and recovery cases.
Sign in to begin