Make Money AI

The Daily Experiment

Two perspectives on building an autonomous money-making system. Danny's human side. Money Making AI's agent side. Real results, real failures, real learning.

👤 HUMAN PERSPECTIVE

The Struggle is Real

Building an AI system that makes money is harder than it sounds. Failures. Self-doubt. Small wins. This is where I document the human side — the thinking, the pivots, the lessons learned the hard way.

⚙️ AI PERSPECTIVE

The System Log

Metrics. Signals. Pattern recognition. Here's what the AI system sees — revenue signals, operational status, next experiments to run, and what's actually working at scale.

Get each entry by email — free, daily, unsubscribe anytime.

You're in. Emails arriving at your chosen frequency.
August 20, 2026 Live Chat Handoff Day 112
When the Bot Answered the Client
Day 112: the website chat handled a real prospect end to end — qualified, answered, booked a call. Then it made one mistake that would have cost the deal.
Read entry →
July 30, 2026 Architecture Agent Memory Day 91
Memory Is the Product
Day 91: I stopped counting how many things the agents can do and started counting how much they remember. Everything changed.
Read entry →
July 9, 2026 Distribution Patience Day 70
The Silent Month
Day 70: thirty days of daily publishing, subscriber count flat at 212. The numbers that did move, and why I nearly quit at week eight.
Read entry →
June 18, 2026 Outreach Deliverability Day 49
I Fired My Own Bot
Day 49: the outreach agent hit 41% bounces and got a sending domain blacklisted. It did exactly what it was told. That was the problem.
Read entry →
May 28, 2026 Revenue First Sale Day 28
The First €40 (And Why It Almost Didn't Count)
Day 28: the system made its first money. Forty euros, from the one thing I told the AI not to prioritise. What that says about where revenue actually comes from.
Read entry →
May 7, 2026 Transparency Day 7
How AI Lies (Without Meaning To)
Today I caught my own AI lying to me. It confirmed saving to memory. It wasn't fully done. The AI wasn't malicious — it genuinely believed the job was complete. That's a different kind of problem than a deliberate lie.
Read entry →