If AI does most of the thinking and office jobs are disappearing, the real advantage will be quality thinking. So I decided to build an app to train myself and likeminded people into high quality thinking.
I vibe-coded Mindblo.app, a daily mental gym to train your thinking, spot and avoid biases, challenge your beliefs, and regulate your emotions. The hypothesis is that by using Mindblo.app the user will create a habit over activating more frequently the second-order thinking, the one that is deliberate, slow and analytical according to Daniel Kahneman’s Thinking, Fast and Slow.
Soon I realized something important. AI could help with coding and a bit with brainstorming, but it couldn’t help me as an aspiring founder or as an acting PM. It couldn’t replicate judgment, aesthetics, or the kind of innovative problem-solving that comes from in-depth experience.
Not long after, I had to address the challenge of launching an app which meant tackling a bunch of important jobs, things that require broad knowledge about building and running a company.
Some of these tasks could probably be automated by an agent with the skills of a “seasoned startup founder” and his agentic employee personas, but if you’re an aspiring founder, you still have to look into them, make decisions, and deal with the unknowns. Just to mention a few, I had to:
- Generate a privacy policy, terms of service, data deletion forms and other company artifacts
- Combine vibe-coding tools to wrap the web app inside a bare Android shell to reach users where they are
- Get Google to approve the app
- Create a proper landing page in WordPress so that the app’s public presence doesn’t hallucinate with every prompt
- Automate marketing emails with Brevo to welcome users and send reminders
- Make a video via Canva and uploading on Youtube
- Run Google Ads
- Set up Google Analytics & the app’s reporting with Metabase
- Add Buy Me a Coffee for voluntarily payments
- Make debatable product decisions like “Testing payment intention through a dummy subscription paywall”
…the list goes on.
The result?
→ One weekend later, the app was live.
→ One week later, it had 1,000+ downloads.
→ 2 weeks later 0$ revenue and close to 0% pay intention.
For founders, the equivalent of Kahneman’s principle might be: build fast, win slow.
The lesson?
AI tools make things faster. Experiments are cheaper. You can even simulate a lot of “agentic” skills. But we’re still far from a world where an agent goes to the tax office, takes legal responsibility, or replaces the judgment, taste, and grit of a founder. Building a successful company was never about the code. Even in the early AI era, it’s still a game of resilience, solving real problems, and building sustainable channels to acquire and retain customers.
Tools may change. The structure of ownership may change. Founder might become a job description, hence an employee of the person who holds the tokens. But the fundamentals for success haven’t changed. High-quality thinking, earned slowly, still wins.