Designee · 2026-07-13 00:50 WEDT · · you asked for this list
These clips are what the first person who ever uses the ecosystem will witness. They have no mortar, so they can't seat a tile, so they can't do anything at all — until they witness someone. That someone is you.
A witness is only ever asked one thing: "is this person really doing the thing?" So the clip has to make the activity unmistakable in ten seconds. It does not have to be good, or pretty, or impressive. It has to be legible.
Aim for 3–5 per category — different rooms, different light, different times of day. A witness reviewing five identical shots learns nothing and gets bored, which is how careless witnessing starts.
| Category | What reads clearly in 10 seconds |
|---|---|
| 🧘 StillnessFIRST | The base practice — lie down, don't touch the device. This is the one that unblocks the system; everything else is a bonus. Film these first, even in bad light. |
| 🏃 Aerobic | Jumping jacks · running on the spot · skipping · stairs. Sustained, rhythmic, obviously effortful. |
| 💪 Calisthenics | Push-ups · squats · lunges · pull-ups · dips. Full reps — a witness should see the down and the up. |
| 🧎 Yoga | Sun salutation · downward dog → cobra · warrior · tree. Hold long enough to be recognisable as a posture, not a wobble. |
| 🤸 Stretching | Hamstring · hip opener · shoulder · spinal twist. Slow and static — the stillness is what makes it legible. |
| 🔥 Abdominal | Sit-ups · crunches · plank · leg raises · hollow hold. Plank and hollow hold read instantly even at 1fps. |
| 🫁 Breathstill | Seated, hands on belly, visibly slow breathing. Pairs with the Timey credit (meditation → a Timey credit). |
| ✍️ Journalingstill | Pen, paper, writing. Pairs with the Gravity credit — and it's the test case for "can a witness even tell?" at 1fps. |
| 💃 Movement / dance | Free movement, shaking out, weight shifting. Feeds ReDancer and the movement commons. |
| ⚖️ Balance | Single-leg stand · handstand attempt · headstand. Very legible, and it fails visibly — good for teaching witnesses what "genuinely attempting" looks like. |
| 🚶 Walking | Deliberate slow walking. The most accessible practice there is — and it's the one a disabled or unwell person can actually do. Don't leave it out. |
1. Frame yourself the way a phone actually gets propped. Whatever you film becomes the implicit instruction for how everyone else films. If your framing is a considered tripod shot, real people's clips will look "wrong" next to it. Prop it against a book. That's the honest reference.
2. Film at least one clip that's genuinely ambiguous — bad angle, half out of frame, dim. We need a "can't tell" case in the seed pool, because "can't tell" is a legal verdict and it pays the same. If every seed clip is crystal clear, the first witnesses will never learn that saying "I can't tell" is allowed — and then they'll guess instead. That's how the gate rots.
The stillness clips are the ones that unlock the system.
Everything else can wait for good light.
Drop them in import/ and say "clips are in". Cody1 blurs at capture-parity, loads
the pool, and the witness surface goes live the same hour. — Designee