If you’ve ever been the only person in a room using a weird tool—a strange text editor, an obscure version control system—you know the strange mix of pride and loneliness. You feel like a pioneer. But mostly, you feel like a weirdo. Then someone else shows up, and they quote the meme: “There are dozens of us. Dozens!”
That joke holds the secret to how real adoption happens—and most AI companies are missing it.
Here’s the scene. On the Codex GitHub repo, a developer posted a request for first-class support for Jujutsu—a version control tool so niche it makes Git look mainstream. The top comment? “There are dozens of us. Dozens!” It’s a joke. But it’s also a confession of fierce loyalty.
Supporting an obscure tool isn’t a feature request—it’s a loyalty test. When an AI assistant acknowledges Jujutsu, it signals: We see you. We are one of you. That signal is worth more than ten thousand generic features.
The $100 billion AI coding revolution is obsessed with the median developer. The one who uses VS Code, Git, and demands a ChatGPT plugin. But the median developer doesn’t evangelize. The niche user does. The user of a tool with 50 active contributors? They will tweet, blog, and drag their entire community into the ecosystem.
You’ve probably been told to ignore obscure tools. “Focus on what everyone uses,” the advice goes. But that’s great advice if you’re building a product. It’s terrible if you’re building a movement.
Neutrality is death. Pick a side—even if it’s the side of the dozens. The AI companies that win long-term won’t be the ones with the most users today. They’ll be the ones whose tools pass the loyalty test of the smallest, most passionate tribes.
The developer who posted that GitHub issue didn’t just ask for a feature. He declared allegiance. And if Codex builds for him, he’ll return that allegiance a thousandfold. That’s the Mimeng principle in action: emotion first, logic second. The logic says Jujutsu has 50 users. The emotion says those 50 will make you a believer.
So next time you debate which tool to support, look for the joke about the dozens. That joke is a map to where real loyalty lives.
FAQ
Q: Isn't it wasteful to support a tool used by less than 0.1% of developers?
A: Not if that 0.1% includes the early adopters and influencers in a critical niche. Their loyalty is worth more than 10,000 indifferent users. A single passionate tribe can create a moat that scales.
Q: What should AI assistant builders do differently based on this insight?
A: Audit your users' most passionate tools, not just the most popular ones. Support the weird ones first. Run the loyalty test: if a user would proudly share that your assistant works with their obscure tool, you've earned a brand advocate for life.
Q: Maybe it's better to ignore obscure tools and dominate the mainstream?
A: Dominating the mainstream is a commodity game—everyone is fighting for the same user. Winning the hearts of a few passionate tribes creates defensible moats that competitors can't replicate. The dozens are a strategic asset, not a bug.