The Biggest Pivot I’ve Made As A Product Builder

I hit a frustrating point with my app in April 2026: no matter how much I tweaked it, users were not showing up.

New features, changes in the copy, a structured onboarding to retain new users…

Nothing.

But what was wrong?

The epiphany came to me while reading an essay from Paul Graham (Y Combinator): How to earn a billion dollars.

Graham, always eloquent, proposes the following:

“Since your intuitions about other people’s needs are usually a crap signal, and your own needs are an especially valuable one, you should usually listen to the second signal; you should make something you and your friends want.”

Looking back, it’s obvious: my intuition about what others wanted from my app was definitely a crap signal.

So what did I actually want from it?

That’s the interesting question, and it took me a while to discover it.

I spend three weeks pondering this question. Mulling over it. Suffering it.

“What the hell do I want?”

It takes a lot to patience to discover this, and I actually got to it indirectly.

The first idea that came to mind was to build an AI chat to simulate a conversation with Socrates. This is something I’ve wanted for long.

This idea moved me forward, it motivated me, but the real revelation came when thinking about how I wanted to interact with this Socrates-like character.

“Not another chat,” I thought, as the issue with chats is that they open more topics than they close.

When you write 10 words to an LLM, it returns 1,000 words back to you, and I figured that if you talked to Socrates, it would be the opposite: you said 1,000 words to him and he would give 10 words back to you. A single question; a short and unexpected comment.

This is what I wanted. Something I could read in less than 5 minutes instead of an open conversation.

So this was the initial idea: a sort of sage within the app that shared very concise observations about your writing.

After weeks of thinking, I finally started building, and surprisingly, the Socratic character ended up being a secondary feature.

The main feature became the platform to communicate with Socrates: the new Journaling module of Detox Mental—a place to store your journaling entries, reflect upon them, and if you wish, to even write your own book.

This is something I find incredibly cool and that my close ones have also approved as something worth exploring. This was the valuable signal I was looking for.

—

Building products is hard, but as time goes by, the building part is becoming less difficult. The value is shifting toward DECIDING what to build.

I totally agree with Paul Graham: the best way to build is to find something that you and your close ones think is cool.

This is the compass guiding me in the last few months while building Detox, and based on my own usage, I’m getting more insights in a week than in months developing without any feedback from users.

If you have also wondered what it would be like to get comments from Socrates, feel free to write your first entry in Detox Mental today. You may be surprised by how insightful an LLM can be when constraint to respond in 10 words instead of 1,000.