In eight weeks, we built four fundamentally different prototypes for an industrial AI product. Three got thrown away. None took more than 48 hours to build. Here's what that does to venture building.

I spent a week in the research literature on cognitive bias, innovation methodology, and interface design. What I found is reshaping how I build Assumption Mapper, and what it means to track assumptions at all.

We built a beautiful dashboard for industrial AI. Operators loved the data. Then we threw it away and built a chat interface instead. Here's what four prototypes taught us about where AI actually belongs in plant operations.

Chinese markets move at a pace most European corporates can't comprehend. What 13 years taught me about speed, customer proximity, and the 'good enough' philosophy, and how Western companies can adapt without copying.

Evidence from interviews and notes never reaches your system. We built a tool that auto-links raw conversations to your riskiest assumptions.

Every morning I confront my riskiest assumptions before checking email. When spreadsheets failed, I built Assumption Mapper.

We built Forward in 75 days, from first commit to beta. Every mistake we made is a lesson the startup sim tries to teach.

A simple time tracker grew into a personal operating system with AI meetings, CRM, and weekly reviews, built by following the friction.

Every venture runs on untested beliefs. Most teams never write them down. Here's why assumption management is the discipline you're missing — and the tool I built to fix it.

Most customer interviews generate polite lies. Here's how to structure conversations that surface the truth.