Back to Blog
Week 4: The Reality of Building (And Why You Should Build)
Building in Public

Week 4: The Reality of Building (And Why You Should Build)

July 22, 20256 min read
#sound design#ai music#video editing#building in public#product development#mvp#coding#cursor#node.js

Week 4: The Reality of Building (And Why You Should Build)

Week 4 of building an app that will replace me (as a sound designer). Follow along as I either succeed or fail - probably the latter.

By

Tomas Havranek

Jul 22, 2025

When Life Gets in the Way

For the first 5 days of the week I didn't build - life can get in the way sometimes. Not much you can do about it.

Saturday and Sunday I tried to get the app online and working. I somehow managed, but the whole codebase is a mess and if I just publish it now I'd probably end up exposing API keys and user data because I still need to understand how the code actually works.

But essentially I turned the whole backend into an API which I'm now using with my front end. Here's the very basic UI:

Basic UI screenshot

The Reality of Building

This is what building actually looks like. Most weeks you don't make dramatic breakthroughs. Most weeks you're just trying to get stuff to work without breaking everything else.

The app is online and technically works, but the code is barely holding together. While I want to get it out ASAP rather than perfecting, releasing an app full of bugs does not make sense. The goal of this phase is to validate an idea on a small scale. But if users leave before they even get to use the app, I'm not validating anything other than my ability to build broken apps.


Why You Should Build

Since this week was weak on actual progress, here's something that might be more valuable - a nudge to get you to try building yourself.

We're all creative, and to me, building is the ultimate act of creativity. You can literally create anything. Out of nothing. A thing that millions of people can find useful or beautiful.

I've never coded before this project (as you can definitely tell if you've read any of my previous weeks). And two years back, we were highly limited by the learning curve of actually learning to code. Nowadays, you simply need to be open to using new tools and you can build amazing apps using natural language.


The Two Tools You Actually Need

You've heard of different programming languages, and with AI, which one you use matters less and less. What matters is how you get to your desired result the fastest. I'd suggest getting started with web apps. They're the simplest to use, test and publish.

My suggestion - you essentially only need two tools to build anything on the web: Node.js and Cursor.

Node.js is just a platform that lets you run stuff locally. You install it once and then can forget about it forever.

Cursor is where your code gets put together. In programming terminology these would be called code editors, but with AI you're not really the one editing the code. Cursor comes with an AI agent built in (a simple text window which you prompt with instructions) and the agent executes changes in the code for you.

The pro tier on Cursor uses better AI models, so you simply tell it what you want to build, iterate and the results are great. If you're on the free tier, use your favourite AI chatbot to build a step-by-step plan on how to execute what you want to build. Then you just paste those steps to Cursor which will actually write the code for you.

If there's a bug, you paste it back into your chatbot and keep going like this.

The reason I choose Cursor, which can seem a bit more intimidating over tools like Lovable or Bolt.new, is that once you know your way with Cursor, you can truly build anything. In any programming language. Bolt.new and Lovable still lock you just to web development.


Getting Your Stuff Online

To publish your site, you can do it for free on Vercel (hosts the front end - the visual aspect of your website). If you have any backend implemented (in most cases this usually means storing user data like logins, usage etc.) I use Firebase.

Don't worry about these terms too much. You'll naturally make your way toward them when you need them.


The Best Time to Build

We're going to see 10 person 1B companies pretty soon… In a group chat I have with my tech CEO friends, there's a betting pool for the first year that there will be a one-person 1B company.

– Sam Altman

Building might be one of the best things you can possibly do right now.


Next Week

Goal for week 5 is simple: clean up the code enough that I won't accidentally leak my API keys, get rid of bugs and get this thing in front of actual users.

Follow along as I either succeed wildly or fail spectacularly - probably the latter.

Subscribe to my Substack