▶ Play the game
No. 01 · Labyrinth
The Minotaur's Maze
Top-down maze crawler. Procedurally generated each run. You are Theseus; the Minotaur hears your footsteps.
A four-week workshop for ages 8–12. Kids describe what they want to make — a game, a site, a story. With a mentor on the call, they build it and ship it.
Spring cohort · starts May · Thursdays, 4–5:15pm PT · 8 kids · $495 tuition ($150 deposit to apply). Also offering 1:1 sessions and a 4-week 1:1 Builder course.
Thirty-two bottles. Tested. Reviewed. Ranked. Curated by an 11-year-old who will not stop.
Three steps. One mentor. A real URL at the end.
Any idea. Game, site, story, sim. We don't pick the project — they do.
Hands-on with the tools real builders use. Mentor on the call, every step.
Week four, every kid presents a working thing at a URL friends and family can open.
One 75-minute live session a week, plus a little independent work only if the kid is excited about it. Cohorts cap at eight kids so the mentor can see every screen.
Sketch, pitch, and get a first working version up by the end of the week. Not the game — but something real to react to.
Hands-on with the tools real builders use. Kids learn to drive them — not to read every line underneath.
Playtest with the cohort. Notice what's not right yet, describe the change, try again. Taste over technique.
Demo night for friends and family. Every kid shares a real URL and walks through what they made.
In our pilot cohort, a student named Ellis decided he didn't want to make one game — he wanted to make six, one per myth, and put them on a single site. Here they are. They all work. You can play them right now.
Came in obsessed with Percy Jackson. Left with a playable mythology anthology, a vocabulary that now includes "refactor," and a new strong opinion that Hades is the most misunderstood god, actually.
Led by one person. Every call, every kid.
The program was founded by a working software engineer and former middle-school teacher who spent a decade watching the gap between what kids can imagine and what they're allowed to make. AI tools narrowed that gap overnight. Tell and Show exists to teach kids to use that narrowing well — with taste, judgment, and a point of view.
The approach is borrowed from studio art programs, not coding bootcamps: critique, iteration, public demo. Kids are treated as the authors of their work, because they are.
A single four-week cohort is enough for most kids to ship a thing they're proud of. The other tiers are for families who want more time, more scope, or ongoing practice.
The core program. One kid, one cohort, one shipped thing. Best first step.
For kids who catch the bug. Finish one project, return with a bigger one. Depth over breadth.
A single session, one kid. For a specific project, an exploratory chat, or a check-in between cohorts.
Solo and intensive. For a kid with a specific ambitious project who wants the mentor's full attention for a month.
Already done a cohort? Ongoing monthly studio is $295/mo — email the mentor to pick it up. Requires at least one prior cohort.
Direct answers. If you have a tenth, the form at the bottom goes straight to the mentor.
No, and this is the most important thing we'll tell you. Kids who hand AI a one-line prompt get a generic, broken result — and learn nothing. The program is entirely structured around the opposite skill: noticing what's wrong, describing exactly how it should be different, and iterating until the thing matches the picture in their head.
We call this taste and judgment, and it turns out to be the hardest thing to teach — and the most durable, because it's orthogonal to which tool they'll use in five years.
Three things, specifically: (1) how to describe what they want in enough detail to be built — a writing skill, essentially. (2) how to drive real tools — the same kind adult builders use. (3) how to tell when something is wrong, and what "wrong" means to them.
Those three skills transfer directly to every future tool, every future AI model, every creative or professional context. The specific framework they use during the workshop will be obsolete in two years. The skills won't.
It's the exact right window. Old enough to have strong opinions about what's fun, young enough to not be afraid of making something bad first. We've had eight-year-olds ship games that play cleanly and twelve-year-olds ship short novels with working hyperlinked footnotes. The kids self-select on ambition, not age.
Scratch and Code.org teach a platform — you learn their blocks, you build inside their sandbox. They're excellent; many of our kids come from them. Outschool is a marketplace of one-off classes.
Tell and Show is a small, mentor-led studio with one explicit goal: in four weeks, your kid ships a real, non-sandboxed thing to a real URL, using the same kinds of tools adult builders use. It's more demanding and more personal. Eight kids, one mentor, every week.
Most of our strongest students aren't. The kids who thrive are the ones with something specific they want to exist — a book, a game, a site about their interests — and the patience to keep describing it until it shows up. That's a language-and-taste skill more than a tech skill. If your kid can write a convincing book report, they can do this.
One 75-minute live session per week, plus whatever independent work the kid chooses to do — often none, sometimes a lot, their call. The live session is highly active: talking, sketching, demoing to each other. It's closer to a drama class than a passive screen.
Parents tell us this is the rare screen time they feel completely fine about, because at the end of it there's a thing that exists.
A working software engineer and former middle-school CS teacher. See the mentor section above for the long version. Short version: they've mentored 48 kids 1:1, taught in classrooms for six years, and ship software for a living. Every session is led by them personally — there is no rotating staff, no TAs teaching the curriculum.
Three paths, roughly: (a) most kids take a break, keep tinkering at home, and come back for a second cohort in the next season. (b) A smaller group moves into the ongoing monthly studio, which is where returning students continue working on bigger projects. (c) A few kids don't continue, which is completely fine — they still leave with a shipped thing they made.
We use enterprise-tier AI accounts that don't retain children's inputs for training, route everything through a parent-visible project dashboard, and never ask kids for personal information in prompts. Live sessions are private, password-gated, and recorded only for the cohort. Shipped projects are hosted on URLs that can be public, family-only, or deleted at any time — parents decide.
Full data and privacy documentation goes out with the enrollment packet.
A short application — kid's name, age, and one thing they'd want to make. We place a $150 hold on your card to reserve the seat; we only charge it if we confirm the spot. Jim reviews personally within 3 business days.
Apply for a cohort → Book a 15-min call →
Or email jim.kernan@gmail.com — we read that inbox ourselves.
We review every application personally. A $150 refundable hold secures your kid's seat — not charged until we confirm.
Nothing is charged until we review and accept. You'll land on a confirmation page once your card is authorized.