I am Pascal.
What fascinates me is how much you can build today if you have a clear idea and can explain well what it should become. The part I enjoy most is the path from the first thought to the moment when something actually works and I like using it myself.
Most of it starts in my own daily life
Heyapply came out of my own job search, when I realised how much time a genuinely good, well-matched application takes. I am building Memember because I sometimes leave meetings with plenty of thoughts but not with all the notes that mattered. And I built the Kids Music Player because my son wanted to listen to music himself and we could not find a solution that really worked for him.
Many of my projects start exactly like that: something bothers me, is missing, or will not leave me alone. At some point I no longer want to just think about it, but find out whether I can build a better solution myself.
Not every idea has to become a company. Some projects go public, others stay small tools for me or my family. And sometimes it is enough to have seriously tried an idea and understood why it works, or why it does not.
What building has taught me
01
Do not start every idea right away
I would happily build almost everything that goes through my head. But if I start on every new idea immediately, I quickly end up with too many projects at once. So I write them down first and come back to them a few days or weeks later.
If an idea is still on my mind then, and I catch myself thinking through possible flows, that is usually a good sign. Others lose their appeal on their own. That does not mean they were bad. They just did not have to be built right away.
02
Not everything has to become a business
AI has made building software far easier. But a working product alone is nowhere near enough to reach a lot of people. Attention, trust and an existing user base cannot be built up nearly as fast.
So I try not to judge every idea by whether it could turn into a big business. What matters more to me is that the problem interests me and that I enjoy the product myself. Once the first excitement wears off, it takes enough personal motivation to keep going and to look after the many small things that make a product good in the end.
03
Start small
With a first version I no longer try to build the whole product already. To begin with, it should only show whether the most important part of the idea works at all.
That does not need every setting, every edge case and every feature that might make sense later. The smaller I start, the faster I find out whether there is really something to the idea, or whether it only worked well in my head.
04
Use the product myself
As long as I only look at a product, I can talk myself into a lot. I usually find the real problems once I need it in daily life and have no patience for unnecessary steps.
So I use my own products myself as early as possible. That shows me quickly which features actually help, what is still missing, and which things I built but never really need.
How I work on an idea
Before I start building, I try to understand as precisely as possible which problem I want to solve and what the simplest useful version could look like. I play through different flows, question my own assumptions and try to work out what the product really has to do.
I get as concrete as I can about the design too. Words like «modern» or «clean» only help so much. So I give Claude Code plenty of reference and inspiration images. That can be single elements, whole apps, dashboards or landing pages. The point is not to copy anything. The images help me show more precisely what kind of structure, style and handling I am after.
Many of my recurring preferences now live in my own skills. They hold things like how I like to build landing pages, what to watch for in the handling, and which design solutions usually do not suit me. Claude Code therefore does not start from zero on every new project. A new landing page is often already about eighty percent in the direction I had in mind with the first version. The last twenty percent usually still take a few rounds.
For important decisions I also have things researched thoroughly. I want to know how the best comparable products solve a problem, which approaches have proven themselves, and where other solutions have their weaknesses. The AI is not supposed to simply confirm my idea. I deliberately have it look for counterarguments and better options before I settle on a direction.
What I currently build with
Most of the actual building happens in Claude Code today. I use ChatGPT mostly for brainstorming, writing, research, and to develop loose thoughts further.
Which technologies sit underneath depends on the project. Often I work with Next.js, TypeScript, React Native, SwiftUI, Supabase and Vercel. These tools keep changing and usually follow from whatever I want to build at the time.