My daughter Lara is six years old. She watches me work on Skales every day, knows that the gecko on my desktop "helps" me, and has opinions about its appearance. One afternoon she walked up while I was working, pointed at the screen, and asked: "Can Skales make me a game?"
I said yes. I told her to describe what she wanted. What followed was the most honest demo of no-code AI development I have ever seen — not because it was polished, but because it was exactly what it was: a child who had never typed code in her life, building software by describing what she wanted.
The Setup
I pulled up Lio AI in Skales and turned the keyboard over to Lara. She types slowly and with some help on spelling, so the session was unhurried. I did not help with the description — just with the physical typing. The brief she dictated, verbatim:
"Can you make a snake game? The snake eats apples and gets longer. The snake is green. Do not let it hit the wall."
That was it. Thirty-three words. No technical specification, no mention of any programming language, no concept of how games work technically. Just a description of desired behavior from someone who plays games but has never thought about how they are built.
What Happened Next
Lio AI processed the description, asked a single clarifying question — "Should there be a score display?" — to which Lara answered "yes please" — and then got to work. The code generation took about four minutes. Lio wrote a complete Python Snake game using Pygame, handled the imports and window setup, implemented the collision detection, the apple spawning, the score counter, and the game-over screen.
Then it opened a preview. The snake was green. Apples appeared in random positions. The snake grew when it ate one. Hitting the wall ended the game. A score appeared in the corner.
Total time from first word to playable game: eleven minutes and forty-seven seconds. I timed it.
Lara played it for twenty minutes straight. She died on the wall approximately forty times and did not seem to mind at all.
A Second Request
After a while she came back and said: "Can you make the apples different colors?" I asked if she wanted to tell Skales that herself. She did. She typed: "Make the apples be different colors each time." Lio updated the code in about ninety seconds. The apples became randomly colored on each appearance.
She made two more modifications: adding a "hard mode" where the snake moves faster (her idea, her description), and changing the background color to purple (also her choice, very much her taste). Each modification took under two minutes.
What This Actually Demonstrates
I want to be careful about overclaiming. Lara did not learn to code. She does not understand what Python is, what Pygame is, or what the underlying code looks like. If the program broke in an unusual way, she could not fix it herself.
What she did do: translate a clear idea into natural language that a tool could act on, evaluate whether the output matched her intention, and iteratively refine it through description. These are genuinely transferable skills — and they are now accessible without the years of technical prerequisites that traditionally preceded them.
The path from "I want this" to "I have this" no longer requires learning a programming language. For children, for non-technical adults, for people who have great ideas but no technical background — the barrier that kept software creation inaccessible has dropped significantly. Read about Skales for kids or explore Lio AI for coding.