The game lab

Lulu and Otto build games.

My niece Lulu is 10. My nephew Otto is 11. Over a handful of afternoons this summer they each finished a game they actually wanted to play: they picked the idea, argued about the rules, broke the code, and fixed it. Everything on this page runs in your browser, on a phone or a laptop. Play it.

  • P1Luluage 10
  • P2Ottoage 11
  • P3insert coin to join

Lulu, 10 Puzzle

Habitat Quest

An animal shows up in disguise with three cryptic clues. Build the climate, land, water, plants and shelter it needs, release it, and find out whether you read the clues right.

Off screen

Coding together
Coding together
Lulu thinks about what habitat to quiz for next
Lulu thinks about what habitat to quiz for next
Otto builds Lego's while he thinks
Otto builds Lego's while he thinks
Wow that was fast!
Wow that was fast!
I'm really just a kid
I'm really just a kid
My sister and amazing Family
My sister and amazing Family

How we work

I do not hand them code. They say what they want in plain language, we ask a model together, and then the part that matters starts: reading what came back, deciding whether it does the thing they asked for, and changing one value to find out what it was holding up.

Otto's first game has a block at the top of the file called TWEAK. Bullet speed, reload time, enemy health. He edits one number, reloads the browser, plays for thirty seconds, and keeps it if the game got better. That loop is not a teaching exercise. It is the same loop I run on work people pay me for.

Lulu went the other way. She wanted twenty-four levels across six continents, each with an animal hiding behind three clues, so nearly all of her time went into writing clues that were fair without being obvious. The hard part of a puzzle game turned out to be prose.

What I want them to leave with is smaller than "learn to code" and more useful: a model will hand you something plausible in four seconds, and it is your job to find out whether it works.