Nice job, but it's definitely lacking a reward mechanism to get this to qualify even loosely as "edutainment". Maybe add a particle emitter glitter gun or some kind of flashy CSS animation. You'd be surprised at how even the smallest things can help keep a child engaged.
About a decade ago my little brother was struggling with the multiplication table so I built an education game based off the magic carpet escape sequence in the Sega Genesis game for Aladdin.
Khan Academy has pretty simple rewards. A correct answer gets you a short chirpy melody/sound effect and some confetti. My kid likes it.
blaix · 16d ago
I played this for way too long... nice work!
jumploops · 20d ago
This is awesome.
My wife built a matching game[0] for our toddler, because she doesn't like the flashy content and addictive nature of the games on the app store (and neither do I!)
It's very simple, but it's exactly what she envisioned.
Disclaimer: I work on the vibe code tool she built this with.
Done similar to help kids with learning beginner math.
Should now be possible to make an app that specializes in making new apps on demand based on simple requirements like these :)
As Khan Academy is experimenting with 1:1 tutoring using LLMs... this may be an adjacent space for them to explore: let users generate apps/games on a math or science topic of choice, beyond the regular quizzes.
It is 95% written by Claude with Cursor free tier. The good part, 5 yrs old want more features like more levels, nice graphics, they basically want all their book in this style but I don't have resources to take this further.
an_aparallel · 21d ago
The noise when you hit a letter is so loud and grating :(
dessimus · 20d ago
>The noise .... is so loud and grating :(
So, like basically every kids toy that's battery-powered?
mushufasa · 21d ago
I've had similar success vibe coding small, self-contained projects.
Where it lacks is wisdom + taste. Not a problem for small greenfield projects. A very real problem for large codebases where it quickly implements silly technical approaches. For example, building an authentication system from scratch for a feature, rather than extending the existing authentication system.
keyle · 21d ago
That's nice, I create birthday invites in 5 mins now for my kids. An ordeal that used to take hours.
The last time I asked Chad Gepeto to make sudoku puzzles though, it went horribly wrong. At least it got my kids to prove that this and that puzzle by AI were broken, which is a net win. Early skepticism against the machine!
About a decade ago my little brother was struggling with the multiplication table so I built an education game based off the magic carpet escape sequence in the Sega Genesis game for Aladdin.
https://specularrealms.com/alad
After a few weeks, he had those tables down cold!
My wife built a matching game[0] for our toddler, because she doesn't like the flashy content and addictive nature of the games on the app store (and neither do I!)
It's very simple, but it's exactly what she envisioned.
Disclaimer: I work on the vibe code tool she built this with.
[0]https://toddler-matching-game.magicloops.app
Should now be possible to make an app that specializes in making new apps on demand based on simple requirements like these :)
As Khan Academy is experimenting with 1:1 tutoring using LLMs... this may be an adjacent space for them to explore: let users generate apps/games on a math or science topic of choice, beyond the regular quizzes.
Without Claude so it took much longer than 10 minutes.
"A shout of joy or excitement"
My answer: "cheer" - expected answer: "cheers"
How hard is it to change it for a different language? Looks fun for practicing words
It is 95% written by Claude with Cursor free tier. The good part, 5 yrs old want more features like more levels, nice graphics, they basically want all their book in this style but I don't have resources to take this further.
So, like basically every kids toy that's battery-powered?
Where it lacks is wisdom + taste. Not a problem for small greenfield projects. A very real problem for large codebases where it quickly implements silly technical approaches. For example, building an authentication system from scratch for a feature, rather than extending the existing authentication system.
The last time I asked Chad Gepeto to make sudoku puzzles though, it went horribly wrong. At least it got my kids to prove that this and that puzzle by AI were broken, which is a net win. Early skepticism against the machine!