Last Wednesday at 10 p.m., I was stuck on my nephew’s request
My nephew is 11 and has been talking about one game for months: a flying cat that jumps between rooftops. He drew it on the back of his homework notebook—more than 20 pages of it. I promised to make a playable version over the summer. By August, I was still figuring out where to start. I don’t code; the closest I’ve come is aligning columns in a Word spreadsheet.
That night, I saw someone build a complete little game in one night with Fable 5.2, a new tool that can read a brief, generate an entire app, and run it. I thought, maybe this is finally my turn. So I tried it.
What it is, and who’s already using it
Tools like Fable 5.2 compress the steps of writing code, setting up the environment, and running the app into a single conversation. I describe what I want in everyday language, and it generates an app I can open directly. I didn’t write any code. I explained everything in Chinese and made three rounds of changes: the cat flew too slowly in the first, the jumps didn’t feel right in the second, and the third finally worked.
I shared my experience in a small indie-game community. A reader named A-Yan, who is based in Hangzhou and has done indie-game contract work for two years, sent me a message. Last year, he used a similar tool to make a small game with sound effects in two days and sold 40 copies for 12 yuan each. He said the most valuable part wasn’t the code; it was that “something that used to take three weeks to run now takes one night.” He described spending that night with “sweaty palms,” because the first thing that ran looked different from the image in his head.
What it cost me to reproduce it
Money: 0 yuan. I used the free allowance and paid nothing extra. Time: 3 hours, from 10:30 p.m. until 1:30 a.m. The first two attempts went off course; the third worked. Technical barrier: knowing how to type and describe things in plain language is enough. When I got stuck, pasting screenshots and the exact error messages back into the chat was much faster than guessing. First step: open its chat window and start with one sentence explaining “what to make, who it’s for, and what style,” rather than opening with “make me a game.”
If I were at this stage, I’d…
If I were just starting out with one vague idea, I’d use it to make an ugly version that runs. I wouldn’t worry about the art yet; if it’s playable, that’s enough. I’d fix what goes off course instead of losing a month waiting for perfection.
If I already had one or two clients and needed something visible within two days, I’d use it as an accelerator. I’d prototype one version by hand, then let the tool handle the repetitive parts. That way, I could still explain every line of what I delivered.
If I were scaling up and turning a small tool into a product, I’d remind myself not to mistake a demo for a product. A free allowance and “something runnable in one night” aren’t the same thing. Once real user data, payments, and long-term maintenance are involved, I’d still ask someone who genuinely knows the work to take a look. Not everyone needs this tool, and it’s fine not to try it now. I’d first make sure I can really free up those three hours.