Pick a stack you can ship on in a weekend
Choosing tools is the most fun way to avoid building. Skip the month of research: here's one opinionated, modern stack, why each piece earns its place, what it costs at zero users and at a thousand, and a real app deployed to the internet before you close the laptop.
- A list of things you must not build yourself
- A Next.js app in a private GitHub repo, with environment variables set up properly and live on Vercel
- A validated problem from Prove people will pay before you write code. You can do this mission first, but don't start building features until you've validated
- A computer you can install software on, and a free GitHub ↗ account
- Comfort opening a terminal. If that's new, do Terminal survival kit first (15 minutes)
The four rules for picking a stack
Know why the choices below are the choices, so you don't get dragged into tool debates.
Your customers will never ask what database you used. They'll ask whether it works and whether it solves their problem. The stack's job is to get out of your way. So every pick below follows four rules:
- Boring and popular. Huge communities, years of answers online, and loads of examples in AI coding assistants' training. When you get stuck at 11pm, someone has hit the same error before.
- One language everywhere. TypeScript for the front end, the back end and the scripts. One language means less context switching and an AI assistant that's good at all of it.
- Managed, not self-hosted. Someone else patches servers, runs backups and wakes up at 3am. You build product. You can self-host later if the bill ever justifies it.
- Generous free tiers and easy exits. Nothing costs real money until you have users, and the important parts (Postgres, standard auth, Stripe) can be moved if you outgrow a provider.
There are other perfectly good stacks: Rails, Laravel, Django and Phoenix have all launched thousands of profitable products. If you already know one of them well, use it. The best stack is the one you can ship on fastest. This mission is for everyone who doesn't have a strong preference yet, and it's the stack the rest of this track uses.
CarefulDon't spend more than this one mission choosing tools. Every hour comparing frameworks is an hour not talking to customers. You can change almost anything later. You can't get back the weeks you spent deciding.