I'm trying to build an AI-first climate business. In public.

I come from twenty years in urban innovation, culture, sustainable construction, and design — not from climate finance or climate tech. At some point I had a climate panic moment that changed everything. I believe we can't solve this by relying on goodwill alone — we need systems that make money — even when people don't care. This experiment is the result: finding real business opportunities in the climate economy, testing them with AI, documenting everything — including the mistakes.

This is the log.

Read the latest entry →
The thesis

Climate action doesn't need only better people. It needs better incentives.

The climate economy is growing whether we're ready or not. The question is whether the systems built around it work with people as they actually are — busy, distracted, not always paying attention — not just the most committed ones.

The method

Almost entirely run by AI. Everything documented.

Every step of the research, the tools used, the dead ends, the findings — written down and published. Including the AI slop, the bad prompts, and the things that only worked on the third try.

The constraints ~8h

Human work per week. 95% of every business explored must run almost entirely on AI. One filter for every opportunity: does it make money without relying on goodwill?

What you'll read

A weekly log. Not a newsletter.

01

The opportunity. What I found this week in the climate economy — at the intersection of cities, territory, construction, and architecture. Why it might be a real business.

02

The test. How I'm validating it — with AI, with research, with real numbers. What passed the filter and what didn't.

03

The mistakes. What went wrong, what I misunderstood, what the AI got completely wrong. The stuff most people don't publish.

04

The numbers. Real revenue, real costs, real metrics. Not projections — actuals. Starting from zero.

Free, every Thursday.

The week's experiment, the method behind it, what I found — and what I got wrong.

Follow the experiment → Free. No spam. Unsubscribe anytime.