AI Without the Hype.
Chapter 14 of 21
Part Five · Build Real Things · Chapter 14

Build it yourself
before you buy

By the end of this chapter you can look at a tool or subscription you pay for and judge honestly whether the simple version of it is something you could build in an hour and own outright.

You built a real thing last chapter. Now turn that power on your own expenses. A surprising number of the apps and subscriptions people pay for monthly are, underneath, something you could build a good-enough version of in an afternoon, and then own forever. Not everything. But more than you would guess.

This is not about being cheap. It is about noticing that the gap between you and a paid tool got much smaller. The viral AI app, the niche calculator, the little tracker, the formatter you pay for: a lot of them are a thin layer of logic you can now direct into existence. When you build it, it stays yours, it bends to your exact need, and it does not bill you next month.

AVATAR OPENER · ~90s
Watch: rebuilding a thing you pay for, in an hour, and keeping it
HeyGen avatar · generated, consistent presenter

The honest version, because this course does not oversell: some things are absolutely worth buying. The judgment is the skill, not a blanket answer. So this chapter gives you the question to ask, not a rule to follow. When something is mostly your own data plus simple logic, building wins. When it is heavy infrastructure, other people\u2019s constantly-updated data, or real liability, buying wins. Learning to tell those apart is the point.

Worth buying
Heavy infrastructure, other people’s live data, real liability, or something a whole team maintains.
Result: let someone else carry it. Pay and move on.
VS
Worth building
Mostly your own data plus simple logic. A thin layer you can direct into existence and own.
Result: build it once, bend it to you, stop paying.
Building is free now, and the thing you build stays yours. When it is your data plus simple logic, building wins. When it is heavy infrastructure or live outside data, buying wins. Learn to tell them apart.
THE BUILD-OR-BUY QUESTION

Run any tool you pay for through these four questions. They turn a vague feeling into a clear call.

Whose data?
Mostly yours? Lean build. Mostly the world’s, kept fresh? Lean buy.
How complex?
A few rules and a tidy layout? Build. A deep, evolving system? Buy.
Any liability?
If a bug costs real money or breaks trust, buying the tested thing is wiser.
Do you want it?
Building it teaches you the thing. Sometimes that is the real return.
SEE IT

Here is the move in practice. You pay a small monthly fee for an app that turns your expenses into a tidy monthly summary. That is mostly your data plus simple logic. A strong build candidate. Here is how you start.

Rebuilding a simple tool you pay for
I pay monthly for an app that takes my expenses and gives me a clean monthly summary by category. I want to build my own simple version. Before building, design it with me: what do I put in, what do I want out, and what is the simplest version that would actually replace the app for me? Then build it as something I can run myself with my own data. Keep it simple. Ask me one question at a time first.
Try it in Claude

Here is the part that makes build it yourself real for someone who has never written a line of code. When Claude makes something substantial, it does not bury the result in the chat. It opens a separate window beside the conversation, called an artifact, and the thing renders there ready to use: a working page, a small calculator, a chart you can read, a document you can download. Your expense summary is not a wall of code you have to host somewhere. It is a thing you can click, try, and keep.

When you are happy with it you can download the file and own it outright, or publish it to a link other people can open. One honest note so you are not caught out: publishing makes a genuinely public link. Your actual chat stays private, but the thing you published can be opened by anyone who has the link and can even turn up in a search, so only publish something you are happy for the world to see. That is the quiet shift this chapter has been pointing at. Building something usable is no longer a thing you watch other people do. It happens in the window in front of you.

You will often find the version you actually need is simpler than the app you were paying for, because the app had to serve everyone and yours only has to serve you. That is the quiet advantage of building: it fits one person perfectly, and that person is you.

NOW YOU TRY · EVALUATE
Run one subscription through the question

List two or three things you pay for monthly. Run each through the four questions: whose data, how complex, any liability, do you want to learn it. Pick the one that leans most toward build, and actually build a rough first version with Claude. Compare it honestly to the paid tool. Decide whether to keep yours, keep paying, or use both.

Right if you can explain, for each tool, why it leans build or buy, and you have one honest verdict backed by a thing you actually made.
Show the worked solution
The honest answer is usually a mix, and that is the right answer. Say your three are: a note app, a tool that checks your writing, and an accounting service. Walk them. The note app: mostly your data, simple logic, you could build a focused version, but a polished app you already love is worth keeping for the convenience, so this one is a fair buy. The writing checker: mostly logic plus your own voice rules, low liability. Strong build, and the version you make is better because it knows your specific never-list, which the generic tool never will. The accounting service: real liability, tax rules that change, other people’s requirements. Clear buy, do not touch it. So you build the writing checker, keep the accounting service, and make a judgment call on the notes. The lesson is not build everything. It is that you now get a real say, because building moved from impossible to an afternoon. The skill is the honest call, and you can only make it once you have actually tried to build the thing.
WATCH FOR
✗You assume everything is worth building. Heavy infrastructure and live outside data are worth buying. Build is not a blanket rule.
✗You assume everything is worth buying. A lot of what you pay for is your data plus simple logic. That leans build.
✗You build a perfect clone instead of what you need. Your version only has to serve you. It is usually simpler than the app you paid for.
✗You build something with real liability. If a bug costs real money or breaks trust, the tested, supported product is the wiser buy.
WHAT YOU LEARNED
The takeaways
  • The gap between you and a paid tool shrank. Many subscriptions are a thin layer of logic you can now build and own.
  • It is not build-everything. The skill is the honest call: your data plus simple logic leans build, heavy infrastructure and live outside data lean buy.
  • Four questions decide it: whose data, how complex, any liability, do you want to learn it.
  • What you build fits one person perfectly, because that person is you. The paid tool had to serve everyone.
Your project · step fourteen

Look at your thread project and ask whether any part of it is something you were about to go pay for. If it is your data plus simple logic, fold it in and build it instead. Your project grows by absorbing the small tools you would otherwise rent. Next chapter takes this furthest: building the one thing nobody can sell you, your own second brain.

For a long time, build it yourself was advice for people with a team. Now it is a question you get to ask about your own monthly bill. Ask it more often than you think.

← Chapter 13: Your first real build Next → Chapter 15: Your second brain