AI Ships a WeChat Mini Program, but the Human Does the Product Thinking
Anyone building for WeChat gets a concrete list of the manual steps AI cannot automate, from cloud console configuration to certification calls. The takeaway extends beyond WeChat: AI removes the typing, not the thinking.
AI agents carried a WeChat mini program from a vague idea to a published app: they wrote the PRD, produced design mockups for four pages, and implemented the code against engineering rules the human set up. The human pinned the requirements, chose the tech stack, and made the judgment calls on layout and interaction.
The pitfalls came from WeChat's platform more than from AI. Free cloud development environments last only 15 days after a production release, personal accounts need separate approval for plugins like text-to-speech, and scan, share, and Canvas features need real-device verification. The run also made clear that AI cannot operate the WeChat cloud console; it only tells you what to click.
Monetization for individual developers follows the same rails: ad components open only after 1,000 unique visitors, and virtual payments go through WeChat's official merchant account. The conclusion is that AI lowers the coding barrier to near zero, but finding a real pain point and passing review remain the actual filter.
The engineering-rules step stands out as the practical lever: breaking project conventions into a routing table and change logs turns AI from a random code generator into a consistent executor.
WeChat's approval and compliance friction, not AI capability, is the real bottleneck for solo developers; AI generates code faster than the platform lets you ship it.
The division of labor described here works only when the human can articulate the product well enough to produce a detailed PRD; vague requirements push AI into guesswork.
The mini program's premise is contested. A core objection holds that card games are phone-free entertainment and that physical tokens like peanuts score faster without interrupting the flow. Counterarguments note that phones already sit on the table and one tap after a round is trivial, though a further reply concedes the real risk is message-checking. A separate line frames the project as more endless bookkeeping, while one reader celebrates it as focused meaningful work.
The requirement isn't understood: 1. Playing cards itself means not using your phone, it's casual entertainment. 2. Scoring while having to take out your phone and operate it, is that troublesome? 3. It's not as good as old ladies using peanuts, corn kernels, glass beads to score directly—convenient, fast, and clear.
1. When playing cards, everyone puts their phone on the table. 2. After a round, you need to shuffle; just tap to enter X points. Even if you're playing for cash with peanuts, you still have to reach out and grab them.
Compared to picking up the phone, the better way is peanuts, because it doesn't interrupt the gaming experience and flow; picking up the phone means temporarily breaking that experience, and if you see a message you might want to reply, which impacts even more.
Very good. Settling down to do one thing is the most meaningful.