A few weeks ago, I sat down with Claude and spent a few hours working on content for my Chinese social media accounts.
By the end of the day, I had 60 complete posts ready to publish.
Not outlines. Not rough drafts. Sixty finished pieces of content, each with a hook, a main idea, and a call to action — ready to copy, paste, and schedule.
I want to tell you honestly what that experience was like, because the reality is more interesting and more complicated than the headline suggests.
The Setup
I run two accounts on Xiaohongshu (China’s version of Instagram meets Pinterest). One is focused on value investing — I share AI-assisted stock analysis and investment thinking. The other is about side hustles and information arbitrage — opportunities that most Chinese people don’t know about because the information lives on foreign platforms.
For both accounts, content creation was the bottleneck. I have the ideas. I have the research. But turning research into polished, readable posts takes time I don’t always have after a full day of work.
So I decided to test what would happen if I treated AI as a real content collaborator, not just a writing assistant.
What I Actually Did
I didn’t just ask Claude to “write me 60 posts.” That produces garbage.
What I did was feed it real research — information about specific side hustle opportunities, their mechanisms, realistic income ranges, China-specific versions of foreign strategies. Then I asked it to write each post in a specific format: a curiosity-driven opening line, three to five concrete points, and a first action step the reader could take today.
The first few posts needed refinement. I’d read them, note what felt off, explain the problem, and ask for a revision. After about ten iterations, I had a template that was producing good results consistently.
Then the real work began. I went through each of the 60 topics I’d planned and let Claude draft them one by one, reviewing as I went.
The Honest Results
The posts were good. Not all of them — maybe 80% were genuinely usable without significant editing. The other 20% needed real work, either because the topic was too complex for a short post or because Claude made assumptions that didn’t match reality.
A few things surprised me:
The speed was real, but the thinking wasn’t eliminated. I still had to decide what to write about, what angle to take, what information was accurate versus plausible-sounding. AI accelerated the execution, not the judgment.
Quality varied by topic type. Posts about concrete, specific strategies (“here’s how to earn money from X platform”) came out better than posts about abstract concepts (“why information arbitrage matters”). The more specific the input, the better the output.
I learned things in the process. Because I was reading and evaluating 60 posts in one sitting, I started to notice patterns — which hooks worked better, which structures were more readable, which topics generated more interesting angles. I ended up with a much clearer sense of my own content strategy.
What 60 Posts Actually Gets You
Sixty posts, published at two per week, is about seven months of content.
That sounds great. But here’s the more honest version: it’s seven months of content at the cost of one intense day, plus you still have to actually publish, engage with comments, track what performs well, and adjust.
The content creation bottleneck is real, and AI genuinely helps with it. But content creation was never the only bottleneck.
Building an audience takes consistency, engagement, and time. AI can write the posts but it can’t respond to your comments, build your reputation, or decide what direction to take the account based on what’s resonating.
What I’d Do Differently
Next time, I’d spend more time upfront on topic research and less time in the generation phase. The quality of output is almost entirely determined by the quality of input — which means the thinking you do before you open the AI tool matters more than anything you do during.
I’d also be more selective. Sixty posts sounds productive. Forty excellent posts would have been more valuable.
Have you tried using AI for content creation at scale? What’s your experience been — does the volume actually help, or does it create other problems? I’d love to hear.
Share your experience or thoughts below.
Related Reading
- I Translate Western Side Hustle Stories for Chinese Audiences. Here’s How the Gap Works.
- I Asked AI to Help Me Start a Side Hustle on China’s Biggest App. My Account Was Banned Before I Got a Single Follower.
- Week 1 of Running This Blog: What I Expected vs. What Actually Happened
Leave a Reply