Last updated: October 2026.

Still in Berlin, still working on Laboperator

Six years on, I'm still working on Laboperator, though the company around it has changed. I joined it at Labforward. Since then it has moved to Xavo, and today the team is part of LabAthena. The mission hasn't changed: a plug & play platform that connects all of a lab's instruments, collects their data automatically and automates the devices themselves, so scientists can spend their time on science.

These days I'm the CTO and CISO, which means pushing the technology forward and keeping it secure while doing it. Underneath the titles, the work is the same thing that's always pulled me in: sitting where people and technology meet. Listening to the scientists and lab teams who use what we build, digging past the first description of a problem to the one that's really there, and finding the solution that doesn't just fix it but lets them do more than before. I still write code, and still find the instruments, the data and the edge cases where they meet more interesting than any org chart.

I'm still not a scientist, and I still find it exciting that something I helped build might sit somewhere in the path of a discovery.

Building with agentic coding tools

Most of the code I've shipped this year, I didn't type. I described it, read it, argued with it and sent it back. It's changed how I work more than any tool since I started programming, and it's changed which ideas I bother trying. I've been writing about what that means for the people who build software.

Making things for fun again

That cheaper path from idea to prototype finally got me to build Voice Runner, an endless runner whose stages are generated from whatever music is playing around you, and which you steer by shouting at it. It had been sitting in my head for a year. More small playgrounds like it are likely to show up here.

Writing again

After more than a decade of silence, this blog is alive again. All of that has me thinking hard again about what it means to be an engineer, and to build a product. When writing the code stops being the hard part, what's left is deciding what's worth building at all. Building the right thing matters more than ever. Ideas are cheap. They only become valuable once they're in people's hands, and they're only worth putting there if they solve something real.

That's the thread running through what I've been writing lately:

• Say the Magic Word: computers asked us to talk to them in 1977, then spent forty years building interfaces that made sure we couldn't. Now we finally can, which means rethinking what an interface is for.
• Technology for the People: nobody has ever been forced to play Tetris. Most software doesn't have that luxury, and the industry has gotten good at not noticing whether people actually want what we ship. Cheap prototypes might make us notice.
• Above and Below the Code: as typing code shrinks, the job moves to the work around it, above the code (knowing what to build and why) and beneath it (infrastructure, data and security).