Community

For three or four days I did nothing but develop. Vibe coding, sure, but it takes plenty of energy in its own way. As for what I built: the groundwork for Patternflow community users to easily share patterns and try them on their devices. And I made the pattern editor a bit more extensible, structurally. There's a mountain of things left to improve, of course, but let me rest from development around here. Doing only this was driving me up the wall… and writing gets me some distance on it. Let me recover a little of the health I wrecked pulling all-nighters, then get back to it.

To put a pattern I'd made, or someone else's, onto my own Patternflow, you had to open the Arduino IDE, paste the code, edit about one line in the registration file, then upload, and after about two minutes the compile would finish and it would upload. Now that process takes 20 seconds at most, and it's all doable through the Patternflow web, with no need to open an IDE at all. Much easier, and much more fun. You can also load a pattern someone shared in the community onto your own Patternflow with a few clicks, and later I'll make it possible to assemble your own favorite pattern setups like a deck and share those. Because even now, sifting out the good patterns from the many is a hassle. Something like Pinterest, where you can assemble picks to your taste and share them, would be even better.

Not just the patterns; the community around Patternflow in general feels like it's slowly coming alive. On Discord people answer each other's questions, introduce themselves, improve and test the 3D model, propose fixes for PCB issues. Above all, the feature I just added, uploading patterns quickly and easily from the web, started from someone else's suggestion. I'd meant to do it too, but I hadn't planned to take it as far as it's now come (afraid it'd be hard) and had only implemented up to the previous stage. Then they finished a PoC on their own and shared the code, and there was no reason not to do it. So I kept working on it for two days and made it a main feature. Good. If I'd done it later, after the features and structure had settled, it would definitely have been much harder. That people from all kinds of backgrounds, gathered around a shared interest called Patternflow, put their time and effort into improving it, is astonishing. I did think it would turn out this way, but having it actually happen is good. Exciting. I hope more people draw on their own backgrounds and share all kinds of experiments.

It would be good if everyone could see what goals Patternflow has in each area, how difficult they're expected to be, where other people are working, and what the overall situation is right now. That would surely help a lot with experimentation and contribution, and additional connections and communication between people could grow out of it. It's a goal I've had since the very early days of going open source, but I had no feel for how to implement it, so I never did. The closest thing, made because I figured I should at least organize the past, is this roadmap in the Inside panel of the site. The content isn't all that accurate, though. For one, I don't much like the design, and going through and fixing all the content takes time, so it's been left more or less untended. Still, the concept itself feels like enough to show that I'll visualize the road traveled, the current state, and the roads ahead this way, and let other people take part in it. I do want to design it properly, though, and I have no feel for it at all. Well, someday.

Ah, this just came to mind: Claude Max really is worth the money. I expected to be developing a lot for a while so I paid for just one month, and it's clearly good. Without it, developing at this speed would be impossible. Not just the speed; in terms of implementation, if there'd been no AI it would have gone past dreadful into impossible. Maybe I could even build the firmware for Patternflow down at a more fundamental level. Or not.

As for the Crowd Supply launch, it's being pushed back. I gather they're running some conference, Teardown 2026, starting this weekend. Because of that, guidance on Patternflow's launch prep keeps getting delayed, and at first I felt a little let down. But looking at it now it's actually good, because it freed up time and let me lay the groundwork for the pattern community, which I'd have had to do eventually. That's partly down to my disposition. I always have to be doing something. When something I have to do comes up, I try to handle it right away instead of putting it off. It isn't that I'm especially diligent or anything. I'm an INTP and lazy to match. Back in school my homeroom teacher always called me lazy, too. But when something is circling in my head I can't get away from it. It keeps nagging at me. That's so tiring that I want out of it fast. So with university assignments and the like, I'm always the first to submit. Whether it's a visual communication design thing or just a university student thing, most people submit in a rush starting about ten minutes before the last day. But I don't even need to, and I work on it the day I get it and turn it in that night. If I were being diligent I'd review it several times and submit comfortably the day before, but I just make it fast and hand it in right away. And then I don't think about it at all. It's a disposition with clear upsides and downsides. Well, what isn't…

Writing after being cooped up at home, losing track of time, doing nothing but development, it feels like I'm putting down all sorts of things. Come to think of it, these are things anyone could make if they ran an AI, even without me. But other people don't go this far. Because I do it at nearly an obsessive level. Is that because I have that much affection for Patternflow? It's already been about four months since I started. Almost exactly four months, I think? By my memory I bought the LED matrix and hooked it up around March 28th. Is that right…? It's the first time I've worked on a single project for this long. That doesn't mean I'm confident in its current level of completeness. Things to improve keep appearing, and the more I improve, the more I see what's lacking. I think I've written this before, but the more I do, the lower the completeness I feel becomes. Maybe the dream is just getting bigger.

Honestly, that's how it is. The feeling that anything works if you just do it steadily, a lot, like a madman. That talent and all that doesn't matter, you just have to do it. Though if you call the ability to keep going a talent too, then I've got nothing to say. For whatever reason, everyone probably has one thing they get obsessive about. I wonder if I just found mine relatively early. Well, this kind of talk should really come after achieving a lot, and I'm not quite there yet. I don't know. Sometimes there are people who envy the ones who found their thing and go at it like mad, and wonder how they can do that, and, well, is there some other method? You just try this and try that and try everything. It feels like I'm only writing obvious things, which I don't love.

After nothing but coding for a while, honestly I don't know what to do next. Ah, Patternflow even keeps showing up in my dreams. I pulled all-nighters and wasn't sleeping healthily, so right after waking I'd often be in a haze. You know those moments where you can't tell dream from reality. Things from the dream felt truly real, and doing something similar would give me a strange déjà vu. Let me take care of my health. I've got to live long.