Solo UnicornThe Story of the Solo Unicorn Club — Why I Built This Community
The Story of the Solo Unicorn Club — Why I Built This Community
Opening
In late 2024, I was running two early-stage projects simultaneously — ArkTop AI (luxury retail) and JewelFlow (a SaaS tool for the jewelry industry). I was working 60 hours a week, most of it spent on operational details, with painfully little time left for actually thinking about direction.
What I needed most at the time wasn't a tool. It wasn't a tutorial. It was someone to talk to.
Someone who was also using AI to build things independently, who was also stuck on certain problems, who also had no plans to join a big company.
I looked through WeChat groups, Discord channels, LinkedIn communities. What I found was either people selling courses, people sharing news back and forth, or a few dozen people asking questions that no one answered. The kind of community I was looking for didn't exist.
So I built it myself.
That's how the Solo Unicorn Club started.
Why I Built It
I work in AI in Silicon Valley, and I'm also active on Chinese social media. I have friends on both sides who are "building with AI," but the conversations are completely different.
Silicon Valley: AI is eating the world. New valuations every week. The focus is fundraising and scale.
China: On one side, anxiety ("Will AI take my job?"). On the other, courses designed to exploit that fear. And in the middle — the people actually using AI to quietly build things — are almost entirely silent.
I'm one of those people in the middle. CS PhD, NYU Stern business background, working on both tech and product, with no plans to raise funding or hire employees. The work I do with AI already replaces what would have required 2–3 full-time hires.
I figured I couldn't be the only one.
In December 2024, I posted on my WeChat Moments saying I wanted to start a small circle specifically for young people using AI to build things independently. No courses, no "thought leadership" — just exchange and mutual support.
Within three days, 180 people signed up.
I knew the demand was real.
Positioning and Values
Solo Unicorn Club — the name comes from a straightforward idea: a unicorn doesn't need a hundred employees. One person, with the right AI tools and methods, can accomplish what used to require an entire team.
This isn't a story about "AI replacing humans." It's a story about "Humans decide, AI executes."
I established three clear rules for the community:
First, no fear-mongering. Many communities keep people hooked with "If you don't do this now, you'll be left behind." I refuse to play that game. We talk about specific tools, real projects, and replicable methods — not abstract dread.
Second, cross-industry. We have members running law firms, jewelry e-commerce shops, content studios, and financial analysis practices. The core logic of AI tools is transferable across industries, and cross-pollination between different fields often produces the most valuable insights.
Third, practice over theory. When you join, you share what you're using, how you're using it, and what problems you're running into. Too many lurkers and the community dies. One of our entry requirements is posting a self-introduction in the first week — who you are and what project you're working on.
These three rules are simple, but they define the community's character.
How to Run a 700-Person Community
This is the question I get asked the most.
One person, already working full-time on two AI projects — how is it possible to simultaneously manage a 700-person community?
The answer: I don't manage it full-time. My AI system does.
Here's how the architecture works:
Admission screening: I created a questionnaire with 5 questions, asking applicants to describe their project and how they currently use AI. The responses are initially sorted by an AI Agent — determining whether the applicant is a genuine independent builder or someone looking to promote their services. I spend 10 minutes a day reviewing the classifications and making the final call. I haven't fully handed this step to AI, because it involves a judgment call about community character.
Content scheduling: I write weekly discussion prompts in advance — things like "This week, share one mistake you've made with AI" or "Share a tool you recently discovered." A scheduled Agent posts them at set times.
Highlight archiving: When someone shares something valuable, the AI automatically identifies it and files it into a knowledge base for new members to browse. The system curates 30–50 high-quality entries per month — there's no way I could comb through every message to do this manually.
Event notifications: Biweekly voice-sharing session invitations, topic summaries, recording transcriptions — all automated.
The entire system's API cost is roughly $50/month.
This doesn't mean humans are no longer important. Every time a truly meaningful conversation happens — someone's project is stuck, someone's looking for a collaborator, someone is building their first AI system and doesn't know where to start — I respond personally. AI handles the information volume; human judgment and cultural stewardship are still on me.
Three Member Stories
Jessica, Law Firm Partner, Shanghai
She joined with a very specific problem: how to use AI for large-scale contract review without violating client data confidentiality requirements. This sparked one of the highest-quality discussions we've had. She ultimately adopted a private deployment solution that processes documents entirely on local infrastructure. A month later, she came back to report that her contract review time had dropped from 15 hours a week to 4.
She's now one of the community's most active members, sharing new discoveries almost every week.
Marcus, Independent Photographer, New York
When he first joined, he was skeptical — he thought AI was "a programmer thing" that had nothing to do with him. Then another member who does content creation helped him set up a simple workflow: after a shoot, it automatically organizes photo metadata, generates delivery folders, and sends a client email template. The process went from 2 hours of post-shoot admin to 20 minutes.
Marcus said he used to think AI entrepreneurship was "a game for people with money and technical skills." Now he realizes that anyone with repetitive work can find value in it.
Annie, Jewelry Brand Founder, Shenzhen
Annie was among the earliest members. She was running an independent designer brand, and her core challenge was customer service — high inquiry volume, but every order required customization, making it hard to standardize. She collaborated with a few technically-minded members and, within three weeks, built a WeChat customer service Agent that automatically handles 80% of common questions. Only the genuinely complex inquiries reach her.
Later she said this community helped her find people to "share a boat with," not just learn about tools. That was a pretty apt way to put it.
Mistakes I've Made
Mistake 1: Growth too fast, quality dropped.
Going from 180 to 500 members, I didn't tighten the admission criteria. Some people joined solely to promote their own services or products, and the community atmosphere started to shift. I spent a month doing a "soft reset" — reiterating the rules, asking dormant members to confirm whether they were still active, and removing accounts that clearly didn't fit the community's purpose. The count went from 520 down to 470, but quality noticeably improved.
This taught me: density matters more than scale. 700 people who are actually building things are worth far more than 2,000 who are just passing through.
Mistake 2: Too much information, not enough structure.
Early on, I didn't categorize content, so valuable discussions were quickly buried. New members couldn't find past highlights; veteran members didn't want to re-explain the same things. After building a simple knowledge base that organized frequently asked questions and the most valuable contributions, new members' sense of "this is useful" went up markedly.
Mistake 3: I accidentally became customer support.
After 500 members, the volume of direct messages exceeded what I could handle. I tried to respond to every one, which consumed enormous time while my response quality also declined. The fix: I realized that 80% of DMs were "What tool should I use?" or "Can my project use AI?" — so I wrote a two-page getting-started guide with answers to common questions and set it up as the auto-reply. Now I only deal with genuinely complex questions that can't be resolved in the community discussion.
Advice for Those Thinking About Building a Community
If you're considering building something similar, here are the three things I most want to tell you:
1. Start by defining who you don't want.
A positive definition ("I want young people who are building with AI") is too vague. A negative definition is sharper — no people who watch but never do, no people who are here to promote their services, no people who share anxiety but take no action. The sooner you clarify these filters, the less work you'll have down the line.
2. The first 100 members set the tone more than the first 1,000.
My first 100 members were practically hand-picked. I personally reached out to people I'd met online who were actively working on projects and invited them in. That group established the "practice-first, mutual support, no fear-mongering" culture, and every member who joined after was shaped by that atmosphere. Community culture is self-replicating.
3. Use AI for operations, but don't use AI to replace your judgment.
Scheduled topic posts, highlight curation, initial screening of new members — all fine for AI. But deciding whether someone truly belongs in this community, handling friction between members, choosing the community's future direction — those require you. Humans decide, AI executes. That division of labor holds just as true in community building.
Closing
The Solo Unicorn Club now has 700+ members spanning tech, law, design, finance, retail, and more. Every two weeks, we hold a voice-sharing session covering everything from AI tool reviews to the psychology of independent entrepreneurship.
I didn't expect it to grow this fast, and I didn't expect it to become a resource for my own projects. More often than I'd have guessed, a problem I'm wrestling with on JewelFlow gets a better answer within two hours of posting in the community than I'd managed after three days of thinking on my own.
I built this community because I needed it. It grew because many others needed it too.
What are you working on independently right now? Drop a line in the comments.