Patternflow

---
title: "It's a Failure"
date: "2026-09-01T16:40:00+09:00"
excerpt: "In a different direction"
cover: "/journal/its-a-failure/pdm-mic.jpg"
coverCaption: "A PDM mic wired onto the ESP32"
hideCover: true
draft: false
---

Through last weekend I posted about Patternflow to several subreddits and sent cold emails to about fifteen magazines. I was trying somehow to drive traffic to the Crowd Supply funding page and create conversions. Naturally the response in the communities was good, and more than the cold emails, the Reddit posts that got a lot of upvotes spawned plenty of purely organic articles. Or did Crowd Supply take some action on their end? I can't tell whether they're doing anything.

Anyway, I got picked up by several electronic music and synthesizer magazines like Mixmag, Synthtopia, MATRIXSYNTH, and MMR Magazine, and traffic came in. Conversions didn't happen. Is the price the problem? GitHub stars and Discord arrivals come, but they don't buy.

The PETG 3D enclosures I'd ordered in advance, anticipating a successful funding, arrived yesterday. PCBway sponsored again, and the quality is very good. It feels much better than last time. I'd worried the material would feel more plastic-y than PLA, that it would be glossy, but it doesn't look like a problem at all. It feels cleaner, which is great. And the shipping was, again, carefully packed to avoid breakage and arrived safely.

<figure className="journal-md-figure">
  <img src="/journal/its-a-failure/petg-enclosure.jpg" alt="The PETG enclosure parts printed by PCBway" />
  <figcaption>The PETG enclosure that arrived. None of the gloss I'd worried about; if anything it looks cleaner.</figcaption>
</figure>

But this too only becomes useful if the funding ends successfully… I hope it doesn't turn out to have been for nothing. To make sure it isn't, I'll have to put in a lot of effort and cost, and I've been thinking about that a lot. Do I use the new grant to force the funding into success somehow and use that as an asset for the next steps? Or, based on the data that conversion isn't happening even after generating sufficient traffic from precisely the right target, do I quickly admit failure and pivot to a different direction?

Actually, "pivot" isn't the accurate word. It's more like deciding whether to skip a bonus stage. The plan was to use the funding to bring in people who don't have time to build a Patternflow themselves, and specifically to activate the pattern community. What's already been proven wrong is that owning a Patternflow device has no connection to creating and sharing patterns. For everyone besides me, even though Pattern Lab's AI system made making patterns easy, the reason to make them is lacking. There's no reason to make a new pattern. The person who's made the most, most consistently, doesn't even own a device. Anyway, what I mean is that this is a separate problem.

On the other hand, the number of people building Patternflow themselves keeps growing. From March, when I started publicizing the project, until now, about 50 people have built their own Patternflow. Amusingly, the evidence for this came from an unintended mechanism. It comes from the AliExpress affiliate link and from a structure born out of a chronic limitation in building Patternflow right now. Whatever else you use for the other parts, I'd noted that for the LED panel specifically you should buy exactly that company's model to make building easy, so the majority of builders end up recorded as a number.

There were 64 orders in total, and assuming one person builds several units, around 50 people seems about right. One disappointing part is that the people who voluntarily share their Patternflow on Discord number only about ten. So soon I'm going to put out a request to tally up the people who've built one. Of course there's a good chance they won't bother. In which case, nothing to be done.

So the foundation for growing Patternflow as a community and an open-source ecosystem can be considered mostly in place. Which means there's enough to attempt the next steps even without the funding.

I have to build a structure where countless variations and experiments happen voluntarily within the community and get shared. To that end I recently tore up the firmware on a large scale. You can see the details in [this issue](https://github.com/engmung/Patternflow/issues/349). It made me quite uncomfortable to do. The truth is I forced the people who had already built and contributed to follow my changes, and there was a clash of opinions in the process. Whether it was really a clash or just a miscommunication, and whether the relationship has recovered, I can't tell. On the surface things are better, but when it comes to people I tend to carry a lot of worry and anxiety, so I can't easily put my mind at ease.

Now there's a foundation for people to add the features they want to the firmware and share them. And the area I'll focus my own development on is audio. Ninety percent of the people who've built and used a Patternflow were most interested in syncing with audio. I've gone back to improving the features I'd dropped while preparing the funding, and I tested attaching a PDM microphone to the ESP32. It works well, and I think there's a lot of potential this way. Except the reel I posted yesterday with high hopes came out with the worst results yet, so I do doubt a little whether this is really right.

<figure className="journal-md-figure">
  <img src="/journal/its-a-failure/pdm-mic.jpg" alt="A PDM mic module wired to the ESP32 on the Patternflow board" />
  <figcaption>A PDM mic wired onto the ESP32. It works.</figcaption>
</figure>

How to document it and let people try it themselves, I don't know. There are plenty of features I want to build right now, plenty of places that need fixing, and I still don't want to give up on the funding either. And in the middle of all this, organizing what I've made and writing it up is the thing I find most tedious… but it also seems like what's most needed right now. I don't know. The project feels disorganized at the moment. I need to open up the paths…

Coming back around, then, I have to think about how to make money. Today is the first day of the semester, and I took a one-year startup leave of absence. Which means I have to make money with Patternflow. If I were just going to end it as an art project, there'd be no reason to take leave. If I sell few at a high price, the buyer could be a company or an individual. Going after companies means building up actual cases little by little; going after individuals means becoming known. I have to make good use of what I've achieved so far.

Anyway, that's how it is. My thoughts have piled up again and it's irritating. I wish some brilliant idea would just pop out, but I don't see room for that. I already know all the directions; what I lack is the conviction to boldly choose one and push. If I were doing this alone I'd just act first and think later, but with several people involved now, it seems I can't. The thing that catches most is the funding, of course. I signed a contract, and much of what I've achieved so far was built on that funding, so is giving it up right? People spent time and did work assuming the funding would succeed, so is it right to render that meaningless?

I should separate the rational perspective from the emotional one, and think about whether my discomfort right now stems from relationships with people, and whether I'm overthinking it on my own again. For now, having decided on the leave of absence, it's right to treat this not as a simple project but as a matter of survival.