June brought my spark back.
For most of the last year I built quietly. I learned, I shipped, I rewrote systems, I taught myself the parts of engineering I did not have yet, and I kept circling the same question: what kind of company do I actually want to build. That work mattered. But June reminded me that I am not at my best when I only build from my room.
I like being in the world. I like rooms full of people making strange things. I like stages, hackathons, demo nights, the late conversations, and the feeling that something could happen if you just show up and dare a little more. That is not a side of me to hide. It is most of me.
So this is the first of a monthly builder update. Not a highlight reel, the actual process while it is still unfinished. I have always learned more from watching friends figure things out than from polished founder stories, and I want to do that in the open too.
The short version of June: I was in Berlin twice, spent time in Munich's builder rooms, spoke at Anthropic's first Claude Founder House, pitched at a demo night, attended an AI Tinkerers meetup, the {Tech: Europe} Applied AI Conference, and the Europe Embodied AI Summit, and won a hackathon with my team at IPAI in Heilbronn.
The longer version is the one that actually moved something: June showed me the gap between what I can build and how clearly I say it out loud. And it made me want to close that gap, loudly.
The work finally went into rooms
Building from home is still where my deepest work happens. I need the long, quiet sessions where I can disappear into a system until it starts making sense. Coworking sessions and builder conversations keep the work from getting too sealed off. But June was a different kind of input: room after room of people who are actually building.
One of them was the Europe Embodied AI Summit, a one-day event that made the physical side feel close: robot arms, robots moving across the floor, live demos, and Andreas Klinger reminding the room that the world is bigger than software. I walked out charged up, with a useful reminder not to let my world get too flat or too screen-shaped.
Other rooms pushed on something harder. I kept hearing a version of the same thing: say what you are doing more clearly. Show the technical depth. Do not make people guess your ability from a nice interface. Say, out loud, that you are on a founder path and looking for the people to build it with.
It was uncomfortable, because it was true.
What I shipped
Most of June's real work does not show up as a headline. It shows up as commits.
I pushed about 70 commits to Just Move to Europe. The visible changes were cleaner mobile navigation, capital results in global search, a sharper investor directory, and better event and resource surfaces. The work I care more about was underneath: source-check dates and provenance, an event import review queue, classifier evals, promotion guards, and signal labels, so the map stays current without turning into an auto-published mess. I wrote up how that system actually works in How I Use Agents to Maintain JMTE.
I also rebuilt this site in another 30 commits, because my public surface was quietly making me look more junior than the work is. I started showing more of the technical layer: how JMTE runs behind the scenes, and how Mini-Me, my private context layer for agent work carries project memory, skills, and routing so I can hand agents real context instead of re-explaining a project every time. Mini-Me itself got more capable this month too, with more skills and tighter eval loops.
And my team and I won the HackNation hackathon at IPAI in Heilbronn. The part I am proud of is the direction we took: instead of overfitting to the sample documents, we built evidence-backed checks against held-out and public reports, with traceable rules, designed to fail on fake success rather than look good in a demo.
The JMTE roadmap came into focus
First, people do not only want the map. They want the room around it. The strongest pull inside JMTE is not the directory itself. It is the question of who else is building from Europe and which rooms are worth showing up in. So JMTE should become a living signal layer for builders here, not a static guide: city playbooks, events, capital routes, resources, and eventually a lightweight place where people can save what matters and find each other.
Second, people are genuinely curious about the technical system. They asked how it works because the mechanics are interesting in their own right, and because trust gets built there too: what gets sourced, what gets checked, and where a human decides. That surprised me, and it changed my mind about what to write.
Third, content is not a side quest. If JMTE is going to be the signal layer, then writing and video are part of the product. The map shows the state of things. The writing, video, events, and changing signals are what make people come back.
Saying the work out loud
For a long time, the work was changing faster than my public explanation of it. I was not trying to hide it. I just did not always know which part was ready to name yet, while I was still building the foundation under all of it.
What I underestimated was how much gets lost when the work stays too quiet. People could see pieces of it, but not always what they added up to. It took hearing that reflected back before it clicked: building the ability was never the whole job. I had to say it out loud, far more often than I had been.
June made it unavoidable. I was invited to speak at Claude Founder House about the system behind Just Move to Europe. I had not trained that presentation muscle in a long time, so I expected to be far more nervous than I was. Instead, I enjoyed it, standing there with a mic and talking about a system I had built. I wrote about how I got there in Treating Fear Like a Research Sprint, but the short version is that I was already someone who could do this. I had just been treating her like a future version of me.
After that, something in me got lighter, and I am more excited to share the journey itself, not only the clean proof at the end. I want the work to be seen while it is still moving.
So the lesson is not that everything is clean now. It is that I can say the unfinished part without sounding lost. The range of directions, the open questions, and the things that shift month to month are the work itself, and this is the stage I actually want to be in, moving parts and all.
The focus question
The hard part now is focus, not ideas.
Several directions are pulling at me: JMTE as a signal layer and community, Mini-Me as context infrastructure for agent work, more technical writing, and the more playful community and media experiments I keep starting. I even began prototyping a collector app for builder artifacts: part merch, part provenance trail for the objects that move through this culture. I can see a world where the limited-edition Nano Banana pullover from the Google Gemini merger is not only merch, but a traceable piece of builder culture. That one is a joke and also not a joke. To keep it from staying only a thesis, I learned to print my own merch at a studio near me, and wore a t-shirt with my site's QR code to the {Tech: Europe} Applied AI Conference.
I am still drawn to deeper infrastructure and technical products, and eventually maybe even hardware.
I also had to let go of a project I was working on because it depended on someone else's infrastructure, and that ground shifted. That is builder life in its less romantic form. Sometimes something is real, and you still have to stop, move on, and make space for the next thing. It stings, but it is also okay.
Resource constraints make the focus question sharper. I cannot let every interesting direction stay a parallel project forever. What I need now is clearer sprints, real deadlines, and more people around the work who can build with me, challenge the direction, or become early believers.
So the question is not whether I have ideas. It is which direction deserves the next serious sprint, and who I build it with. That is what I want July to answer.
What I have been reading and watching
One idea from Amy Poehler's MasterClass, "Prepare to Be Unprepared," kept following me around: you cannot be halfway in. The useful part was not the line as motivation, it was the operating rule underneath it. Do the work, show up prepared, then leave room to respond to what is actually happening. That was the whole June lesson, in rooms, on stage, and in the product: less waiting to feel ready, more showing up prepared enough to stay present. It also landed where I already was: people and process matter more than the idea. I did not expect an improv class to be the clearest founder advice I got all month.
A few other inputs kept feeding the month. Mike Cessario's "Branding with Disruption," which speaks straight to the community and merch side I keep circling. Annie Duke's "Thinking in Bets," read in the gaps while my coding agents finish a run, which turns out to be the perfect use of that time. And a book a stranger handed me on the train, "Der träumende Delphin" by Sergio Bambaren, with a note saying I seemed like the kind of person who would enjoy it. I am taking that as a good sign.
Also, life
June had good texture off the screen. My first cucumber grew in the greenhouse. A friend opened his kitchen store, Marqise, and the mayor gave us a tour of the city.
I also brought back the old habit of calling friends on runs, which still feels like one of the best uses of both. And I spent real time on communication in the most literal sense: pacing, speaking with a lower register, and learning how much voice changes the way people listen.
July
For JMTE: sharpen the city playbooks, keep improving the event and resource automation, make field content that lets people feel an ecosystem in real life (the rooms, the buildings, how people move and gather), and bring the community in more, so I can learn what is actually needed next and make the whole thing more useful.
For me: more builder and co-founder conversations, more grant and fellowship applications, real deadlines around my research sprints, a get-together I am organizing, sharing the journey more consistently, and being far clearer in public about what I am building and who I want around it.
So here is the plain version, since apparently I need to say it out loud: I am building, I want to found, and there is nothing I would rather be doing right now. I am looking for the problems and the people worth going all in on. And I am done making anyone work this hard to see it.
The spark is back. Now I want the outside to match the inside.