SKYWATCH: The AI-Native Build
Build the same real-time flight tracker as the original SKYWATCH course, this time by directing an AI coding assistant: real prompts, real output, and how to review, steer, and verify what it builds.
Let’s build SKYWATCH again. Same real-time flight tracker as the original course, except this time you are not writing the code by hand. You are directing an AI coding assistant through it: real prompts, the real code it produces, and how to review, steer, and verify every piece before it becomes part of your app.
⚠️ Prerequisite: Comfortable with TypeScript and React, and ideally familiar with the original SKYWATCH course or a similar full-stack project. This course assumes you already know what correct looks like, so you can tell when the AI got it wrong.
Built with: Next.js · TypeScript · Fastify · PostgreSQL · Drizzle ORM · WebSockets · Zustand · Leaflet · Tailwind CSS · an AI coding assistant (Claude Code or similar)
No blind trust module. Every lesson shows the actual prompt, the actual output, and a real review pass, including the mistakes an AI assistant plausibly makes at that step and how you would catch them.
By the end, you have the same finished app as the original course, and a repeatable workflow for directing an AI assistant through a real, production-shaped codebase without losing track of what it built.
That’s the plan. Let’s write it.