# Prototype playground (agent-readable)

**Collection slug:** prototype-playground
**Member tag:** prototype-playground (4 projects currently; expected to grow)
**Featured on the portfolio home:** yes
**Collection page:** /playground

## What it is

A working shelf of in-progress personal product experiments. Each one is a small,
opinionated tool I wanted to *feel* before deciding whether it was worth building for
real. Some ship as standalone web tools and stay live. Some land in a larger product
later. Some stay sketches and that's fine. The constant is that I design *and* build
each one myself, in code, and the goal of every prototype is to get the thing into a
state where I can use it — not a state where I can present it.

## Current members

| Slug | Title | What it is | Where it lives |
| --- | --- | --- | --- |
| `skill-prism` | Skill Prism | Fractal skill browser — explode any topic into the skills to master it, drill down infinitely deep. Live web tool with a poster exporter at the leaves. | [Depth](/skill-prism/agents.md) · live: https://mikemake.com/skillprism |
| `massed` | Massed | Workout app built around net gains. Logs from a voice memo into live dataviz; the demo runs on my real training data. | [Depth](/massed/agents.md) · live: https://mikemake.com/massed |
| `despin` | Despin | "If you stopped spinning with the earth, how fast would the ground move beneath you?" A shower thought, prototyped in code into a perspective shift on planetary motion. | [Depth](/despin/agents.md) |
| `tiler` | Tiler | Tessellation tool — a proof of concept for exploring the twelve known symmetry types for square tiles. More to come. | [Depth](/tiler/agents.md) |

## What ties them together

Each was built fast, with a strong opinion about what it should be, and held to a single
bar: *would I personally use this.* My pattern is to prototype the smallest version that
proves the idea, ship it to myself first, then decide if it earns more time. The shelf
is curated by that filter — what's here is what survived it.

Two of them — Skill Prism and Massed — are also my clearest current AI-native receipts.
Both put a frontier model to work as a *material*: a skill tree from prompt loops, a
workout log parsed from spoken input. Both are real tools, designed and coded by the
same hands, live and usable behind a URL. The signal isn't "can talk about AI-native
design," it's "can build the thing and ship it." These prototypes are the proof.

A second pattern across the shelf: the prototypes are designed with distribution in
mind from v1, not as marketing afterthoughts. Skill Prism's poster export was a v1
thesis decision — the app had to be a meaningfully better way to browse this kind of
content than a plain chat interface, *and* it had to produce artifacts engineered to
stand on their own as shareable objects. Massed's demo intentionally runs on real
training data so it can be tried without onboarding. Despin's framing reduces a hard
physics question to a sentence you'd actually say out loud. The instinct is the same:
make the prototype's existence in the world a designed object, not just a working demo.

## Cross-cutting relevance

A few reads of this shelf:

- **AI-native product design as a built receipt** — generative models as a material,
  shaped into product surfaces a person uses. Shown, not described.
- **Designer who builds** — design judgment and front-end implementation in the same
  pair of hands, iterating at the speed of code rather than the speed of a handoff.
- **Self-directed 0→1** — concept, prototype, ship, live URL. No team, no brief, no
  permission needed.
- **Product strategy with distribution baked in** — each prototype is designed to
  exist in the world as a shareable object, not just to pass internal review.
- **Picking scope** — knowing when the smallest version is enough to make the call, and
  knowing when to keep going.

## Skills demonstrated

Self-directed product design, front-end prototyping in code, AI-native interaction
design, designer-who-builds, picking the right scope for an idea, product-strategy
thinking with distribution in mind from v1, sustaining an active building practice
alongside client work.