I Decided Not to Build the Eighth

Once you learn to build services, you catch a strange disease.

The disease of seeing every problem as "couldn't I just build a new service for this?" Before, I'd endure the inconvenience, look for an existing tool, or just do it by hand. But now, whenever something is inconvenient, this thought arrives automatically: "I could just build a service to automate this."

One evening in May, I experienced a flare-up of this disease.

Using StackTube, a problem emerged. Analysis notes just kept piling up. Analyze a YouTube video and a note is generated, accumulating in the folder as PDF, as markdown. But I almost never reopened those notes. Creating was automatic, reviewing was manual, and manual tasks eventually don't get done.

Digital hoarding. The compulsion to store digitally. The thing you download but never look at again. The tool I'd built was actually accelerating that problem. Before, I skipped videos without watching; now I stockpile analysis notes without reading. Only the form had changed.

The disease flared. "Couldn't I just build a new service to solve this?"

In my head, it already had a name. StackBrain. A knowledge-review service that makes you revisit accumulated notes. Spaced repetition, AI tagging, weekly digests, vector search. Features tumbled out one after another. I explained it excitedly to the control tower. I asked it to research eleven competitor services. Readwise, NotebookLM, Recall, Heptabase, Mem... and to organize nine knowledge-management methodologies. PARA, Zettelkasten, BASB...

The control tower did the research. And at the end of it, it said this:

"This isn't a problem to build a separate service for. It should go inside StackTube."

The reasoning was clear. I had a rule I'd set for myself. No more than three projects in active development simultaneously. And if a project doesn't clear 100 users or $200 monthly revenue within twelve weeks, I fold it. This rule was set in advance, without emotion.

Building StackBrain would make it an eighth service. There's already a seven-service roadmap, and this would stack one more on top. And this while the first service—StackTube—hadn't even properly settled.

The control tower's logic: digital hoarding is a real problem. But it's a problem StackTube users have. Let StackTube make you revisit the notes StackTube created. No need to build a new house. Just add a room to the existing one.

Hearing this, I felt a little ashamed.

Hadn't I leaned toward building a new service because it was more fun than solving the problem? That excitement when naming it "StackBrain"—that wasn't a desire to solve a problem. It was a desire to build something new. The problem was an excuse; the real motive was "the thrill of starting something new."

This was the trap solo founders must guard against most. Product sprawl. Products spreading endlessly. The master plan listed it as risk number one. The danger of "managing seven products simultaneously." I couldn't even handle one properly, yet I was trying to build the eighth.

I dropped StackBrain. Instead, I decided to integrate those features—review, tagging, digest—as functions within StackTube.

But this disease was recurrent.

The same day, another flare-up came. This time it was mobile. I wanted to use StackTube on my phone. Which would mean building a native app. iPhone app, Android app, even a Mac app.

I asked the control tower. "Should I build native apps?"

Dissuaded again.

The control tower's analysis was cold. StackTube is a "set it once and it runs on its own" service. Register channels, and every time a new video appears, it analyzes in the background and sends results to email or Obsidian. It's not a service you open daily to operate something. It's closer to a newsletter.

Native apps are overkill for such a service, it said. Eighty percent of native app benefits—push notifications, home screen installation, offline access—can be achieved via PWA. A PWA is a website that installs like an app. And going through Apple's App Store brings in-app purchase fees. Development takes four to ten weeks, and that time would push out other roadmap items—the Japan launch, the next service.

Two flare-ups, two dissuasions. Twice in one day, AI told me "don't build."

Strange. AI is a tool that builds code for you. Say "build it" and it builds. Yet this AI was saying "don't build." And its advice was right.

Here's what I learned. Once you gain the ability to build, the ability to not build becomes just as important. There's a saying: hold a hammer and everything looks like a nail. Learn to build services and every problem looks like a new service. But most problems are solved not by a new service, but by improving an existing one—or by not building at all.

Restraint is strategy.

This is harder than writing code. AI writes the code for you. But the judgment "let's not build this" is mine to make. And to make that judgment, I have to overcome the urge to build something new. Overcoming urges isn't skill. It's discipline.

The night I decided not to build the eighth, I built nothing. And that was the best decision of the day.

---

🔧 Technical Terms in This Episode

Digital Hoarding The habit of continuously storing digital files but only piling them up without reviewing or organizing. Downloaded PDFs, saved bookmarks, captured screenshots that only accumulate.

Product Sprawl The phenomenon of products or features growing uncontrollably. Fatal for solo founders, because building more than you can manage makes everything shoddy.

Kill Criteria Setting a numerical threshold in advance for stopping a project. "Stop if under 100 users or $200/month by week 12." A mechanism for judging by criteria rather than emotion.

PWA (Progressive Web App) A technology where a website behaves like an app. It can be installed on the home screen, receive push notifications, and work partially offline—providing much of the app experience without building a separate native app.

Native App Apps built separately for iPhone and Android, distributed through app stores. Powerful in function but costly to develop, with app store payment fees (typically 15–30%).

Spaced Repetition A learning method of reviewing material at progressively longer intervals to retain it in memory. Apps like Anki are representative.