Can this be radically simpler?
Rules, documents and workflows should not consume attention just because the system behind them is complicated.
The four flagship projects on my homepage are the clearest proof of work. This page shows the wider field behind them — civic systems, reliable AI, human wellbeing, games and product experiments — without pretending every experiment is equally mature.
The technologies change. The questions underneath them are surprisingly stable.
Rules, documents and workflows should not consume attention just because the system behind them is complicated.
Automation is most useful when evidence, uncertainty and consequential decisions remain visible.
Useful products save effort. Great products can also create delight, connection, curiosity or a reason to return.
This is an archive, not a ranking. The coloured clusters show why each area exists and the projects that currently live there.
Rights and public services only create value when people can find them, understand them and take the next step.
Automation should remove repetitive work without hiding uncertainty, evidence or responsibility.
Technology can help people understand themselves and one another, not only optimise their output.
Games are not a detour. They teach feedback, timing, balance and the unforgiving product question: is this actually fun?
A good product has to earn attention — by being useful, delightful or valuable enough that someone wants to return.
No percentages and no commit leaderboard. Just the projects where the next piece of evidence matters.
Mietrecht pilot: turn real user questions, missing sources and reviewer corrections into permanent regression cases.
real-user proofGrow the public watchdog fleet from working runs into useful daily briefings people can actually discover, trust and receive.
working civic pilotThe synthetic reliability layer works; the meaningful next step is one real office workflow in shadow mode.
real workflow nextShow what a truly understandable public-service journey could feel like from citizen to caseworker.
public-service proofYou do not need to inspect the whole archive. These are the shortest paths into the work.
Search, grounding, inspectable sources and real-user failure loops.
Try the Mietrecht tool ↗GOVTECHDocuments, rules, evidence gaps and human-reviewable next actions.
Open the prototype ↗AGENT SYSTEMSHuman approval, explicit policy, observable traces and replayable failures.
Open technical demo ↗PUBLIC SERVICESA citizen-to-caseworker service journey designed around clarity and responsibility.
Try the service flow →I keep old experiments because they show learning and range. I promote only the work that makes the current direction easier to understand.