I ran what I thought was a genuinely interesting experiment.
I gave the same World Cup prompt to 6 AI models — 3 Chinese (Kimi, Doubao, Qianwen) and 3 American (ChatGPT, Gemini, Claude) — and asked each one to predict 14 real matches. Then I bought actual lottery tickets based on the results.
They agreed on 12 out of 14. The one match where they split clean down the middle? Norway vs Senegal. Every Chinese AI picked Norway to win. Every American AI picked a draw. No crossover at all.
I thought: this is interesting. People will want to read this.
I was wrong about the second part.
First stop: Reddit
The experiment felt like a perfect fit for r/ChatGPT. AI content, real data, a genuine result. I created an account, wrote up the post, and hit submit.
The AutoModerator replied within seconds.
“Unfortunately, it has been removed because we noticed that you have an account that is under 4 months old.”
Four months. Not four days. Four months.
I didn’t know this rule existed. I’m not sure how anyone new to Reddit is supposed to find an audience there without already having an audience somewhere else first.
Second stop: Hacker News
HN felt like the right crowd — technical, curious, interested in AI behavior and international comparisons. I found the submit link and tried to post.
One sentence came back:
“Sorry, your account isn’t able to submit this site.”
No explanation. No threshold mentioned. Just: no.
Third stop: Twitter / X
I had an account. I posted the experiment with the photo of the real lottery tickets — China AI on one side, US AI on the other.
I checked the analytics the next morning.
1 impression. 0 clicks. 0 shares.
One person saw it. Probably me.
What this actually means
There’s a name for this: the cold start problem.
Every platform that protects its quality by requiring established credibility also makes it structurally impossible for new voices to establish that credibility. Reddit wants 4 months of history. HN wants karma. Twitter’s algorithm simply doesn’t show your posts to anyone until you already have followers.
It’s not personal. It’s the same wall every new creator hits. But knowing that doesn’t make the wall shorter.
I’m an ordinary person — I don’t have a tech audience, a media connection, or a friend who runs a newsletter with 50,000 subscribers. I have an experiment and a WordPress blog that’s been running for a few weeks.
What actually worked
After the three dead ends, I tried two platforms I hadn’t thought of first.
Dev.to — a community for developers and tech-curious people. New accounts can publish immediately. I posted the same piece. It went live.
Medium — same story. Import the original post, set the canonical URL back to my blog, publish. Done.
Neither went viral. I’m not writing this to tell you I found a magic channel. I’m writing it because the process of failing in three places before finding two that worked is the actual story of what building in public looks like when you’re starting from zero.
The honest takeaway
I used to assume that if you made something genuinely interesting, people would find it. The internet would do its job.
That’s not how it works. Distribution is its own skill, separate from creation, and it turns out most of the places where distribution happens are gated by the credibility you only get from having already distributed things.
So the plan, such as it is:
Comment on HN before trying to submit again. Let the Reddit account age. Keep posting on Dev.to and Medium where new voices are allowed in. Accept that the first few months of any new site look like this.
The experiment was real. The result was interesting. The audience was, for now, approximately one person on Twitter.
That’s fine. I’m writing it down anyway.
Related Reading
I Ran 6 AI Models on 14 World Cup Matches — Here’s Where China and the US Split — the experiment this post is about
Xiaohongshu Banned Me on Day One — the last time a platform said no before I’d even started
I Built This Blog for ¥700 Using AI — where this whole experiment began
Leave a Reply