Flight Lab / v0.1 / in progress

FLIGHT LAB

Paper airplane physics · ages 4–8

A learning experience for kids 4–8 about paper airplane physics.

The conviction behind it: when AI is used in education, it should enforce desirable difficulty, not eliminate it — and stay in the background while it does.

60-second walkthrough
Loom embed coming soon
01 — Why this exists

The conviction

Most AI-in-education tools today are aimed at making learning easier — answer this, summarize that, solve for me. The result: less friction, less thinking, less learning.

The research is clear that difficulty — the productive kind — is the mechanism of learning. Retrieval, spacing, struggle before instruction. AI is uniquely positioned to preserve that difficulty rather than erase it.

Flight Lab is one small test of what that looks like with a real topic, a real kid, and a real artifact they made themselves.

  • Production before instruction — the kid makes their plane before anything is explained.
  • Concrete before abstract — start from the object in their hand, not a physics diagram.
  • The subject is the subject — AI stays in the background; physics is what's on stage.
  • Trust the learner — 4-year-olds are capable. No babying, no gamified dopamine.
02 — How it works

The five-beat loop

  1. 01Beat

    MAKE

    Cold open. No instructions. The app says: make one yourself. The kid folds a plane however they know how.

  2. 02Beat

    TEST

    Fly it. Watch what it does. Notice the behavior before naming anything — a barrel roll, a nosedive, a graceful arc.

  3. 03Beat

    REVIEW

    Photograph the plane. Claude looks at the actual object and asks Socratic questions grounded in what it sees — no lectures, no generic advice.

  4. 04Beat

    EXPLAIN

    One concept per session. For v0.1: asymmetry and lift. Annotated on the kid's own photo, not a stock diagram.

  5. 05Beat

    ITERATE

    Fold again. This time with a hypothesis. Fly again. The loop is the learning.

03 — What makes it different

Against the grain

Most AI-in-education tools
Flight Lab
Chatbot tutor that answers the kid's questions.
Socratic prompts grounded in a photo of the kid's actual plane.
Animated mascot with a personality.
One calm voice. No character. Physics is the protagonist.
Points, badges, streaks.
No gamification. Intrinsic motivation: the plane flies or it doesn't.
Pre-written curriculum, advance by tapping Next.
One concept per session, surfaced only when the kid's artifact warrants it.
Text- and screen-heavy for ages 4–8.
Voice-first with pictogram support. The phone is a tool, not the experience.
04 — What's next

Open threads

  • Near term

    Video upload + flight-path judging

    Instead of a still photo, the kid records the flight. Claude identifies failure modes from motion — barrel roll, stall, tumble.

  • Near term

    Session 2+ of the curriculum

    One concept per session. After asymmetry: weight distribution, wing area, surface disruption. Spacing handled by the app.

  • Research

    User testing with real 4–8 year olds

    The hypothesis lives or dies in a living room, not in Figma. Looking for failure modes around frustration, attention, the voice.

  • Design

    Caregiver view

    A companion surface for the adult in the room — what the kid just tried, what concept surfaced, what to ask later.

05 — How it's built

Stack

Framework
Next.js 15 (App Router, Turbopack)
Language
TypeScript
Styling
Tailwind CSS v4 (CSS-first @theme)
AI
Anthropic SDK — Claude vision model for photo review
Voice
ElevenLabs TTS (planned)
Motion
Framer Motion — stepped, pixelated animations
Fonts
IBM Plex Mono + VT323
Deploy
Vercel — paperairplane.vincelaw.co
06 — About this work

Context

Flight Lab is a portfolio piece built for an application to Senior Product Manager, Education Labs at Anthropic. It is the smallest possible artifact that can carry a real conviction about what AI-in-education should be.

The thinking behind it — the vision, the principles, the design decisions — is published here. Nothing else to dig through.