← All posts

Feb 2026 · 3 min read

86 out of 100

I finished the Breakthrough course in February with 86 out of 100. I've kept the number the way you keep a first payslip — not because it's huge, but because of what it's the first of.

Five weeks earlier I'd walked in knowing nothing about AI. What clicked, somewhere in the middle, was that building with AI is mostly thinking clearly: describe the problem well and the tools meet you halfway. Describing problems clearly, it turns out, is a skill you can carry in from anywhere — including a bus cab.

The moment it became more than a course wasn't the certificate. It was realising, while building the final project, that I had a list in my head titled 'things that are broken that I could now actually fix' — scam job ads, AI hype, the chaos of planning rail trips. The list hasn't stopped growing since.

A course ends with a mark. This one ended with a direction. 86/100, and — more usefully — a plan for March: build something real and ship it.

I was also already trying to start something of my own: RailRoam, a platform for Interrail and Eurail travellers. The idea came from my own Interrail trips and the gaps I kept hitting — delays, seat reservations, sleeper reservations that at times seem to need their own qualification. You basically need to study before even getting an Interrail pass. RailRoam would use AI on historical data — delay patterns, busy routes — and auto-suggest what to book and when: everything I'd have loved in my arsenal on an Interrail trip. I had the business plan written with Claude and a prototype app built on Google Gemini Canvas. The course didn't just teach me tools — it gave me the nerve to use them on my own ideas.

— Michael

Next in the seriesMarch — Building my first app