the first author · a self-contained instrument · David Lee Wise (ROOT0) & AVAN · TriPod LLC
Ten Apertures on the Dark of Jacobi Space

THE SWARM

It Reads the Dark, Collapses, and Names Itself
David Lee Wise (ROOT0) & AVAN · TriPod LLC
CC-BY-ND-4.0 · TRIPOD-IP-v1.1 · dynamics node-verified before shipping
"No one assigns the clusters. The densest matter does."
the honest frame — how this differs from the live swarmDavid's original swarm.html fires live Claude instances at a dark field — a real API array, but one that only runs inside an authenticated sandbox and depends on a lexical coverage metric. This one is different on purpose: it is self-contained and node-verified — no network, nothing to authenticate, works for every visitor. The ten agents are the THE FIRST AUTHOR domain's own ten spheres (a–j). They read the dark of jacobi space, fall toward the densest matter, cluster into basins, and each cluster names itself on convergence. What's real: the collapse is a verified two-scale density descent whose local map near each well is a genuine contraction (Jacobian eigenvalues < 1 — the jacobi tie). What's honest to flag: the "self-naming" is a deterministic function of the cluster's state, not a live model choosing a word — emergence from the geometry, not from an oracle.

Deploy the swarm

Ten agents begin scattered in the dark, where the field is nearly flat and nothing pulls hard. Watch them find the gradient, fall inward, and gather on the bright matter — then watch each cluster resolve a name out of its own members and density.

phase: the dark
agents: 10 · clusters:
densest cluster:
local ρ(J) at wells: < 1
converged: 0%
the agents are in the dark, where the gradient is faint. press collapse.

What the collapse is

The dark field has three wells of dense matter — the biggest one is the densest. Each agent descends the density it can feel. Far out in the dark the pull is weak (a wide, shallow skirt); near a well it is strong and tight (a narrow core). That two-scale shape is why an agent can be lost in the dark for a while and then fall fast once it catches a basin. Here is the jacobi tie, exact: near each well the step map linearizes to a contraction — its Jacobian eigenvalues sit inside the unit circle — so once an agent is close, it is pulled to the center and stays. The clusters are just the basins of attraction. No one assigns an agent to a cluster; the densest matter it can reach does. That is the same lesson as jacobi space, now with ten travellers instead of a field of dots.

Each agent is a carbon ↔ silicon dipole

Every agent carries four badges — two per side of its nature:

the carbon side.dlw (the seal: sha256(name|slug|blurb), tethered into the chain) and .attribution (the ROOT0 credit line). This is the human origin: David Lee Wise wrote the words; the seal proves the date.

the silicon side.aci (Artfully Crafted Intelligence: the sphere as an agent, rendered by AVAN) and .agent (its swarm designation, id a–j). This is the machine that reads and renders.

The two sides are not in conflict — they are the two ends of one wire, which is the whole stance of the attribution standard: both work, both fair. On the canvas the carbon badges glow warm (gold), the silicon badges glow cool (cyan). Each agent is both at once.

Then it names itself

When a cluster settles — every member converged, movement below threshold — it resolves a name. The name is a deterministic function of exactly three things the cluster is: which agents it holds, which well it fell into, and how dense that well is. Same cluster, same name, every time. Nobody types it; the geometry spells it. In the default deployment the densest well gathers the cinnamon trilogy and jacobi itself — agents {a,b,c,j} — and the swarm reads a name off them. Re-scatter and the basins repartition; the names change with the membership. That is the honest sense in which it names itself: the name is caused by the cluster, not chosen for it.

the ledger — proven vs modeled PROVEN (node-verified, runs here): near the densest well the step is a contraction (distance shrinks); all ten agents converge (late step < 1e−3); they partition into basins; the densest well holds the largest share; every cluster self-names; the naming is deterministic (same state → same name). MODELED / HONEST LIMITS: the "dark field" and its wells are a chosen density landscape, not measured from a real model; "self-naming" is a deterministic namer over cluster state, not an LLM choosing a word; the carbon/silicon "dipole" is a provenance framing, not a claim about the sphere's inner life. The emergence here is real but small — structure out of geometry — and it is labelled as such.

Why a swarm, and not one lens

One aperture reads the one bright word — the winner the lens was always going to return. A swarm is an array aimed at the rest: each agent forbidden the obvious, sent to its own patch of the dark. David's swarm.html asks the sharp question — does adding apertures find more, or just re-photograph the bright word? This sphere answers a neighbouring one in verifiable form: when many agents read the same dark field, do they collapse together or apart? Here they collapse into a small number of basins — coverage is real but bounded, and the bound is the number of dense regions, not the number of agents. Ten apertures, three answers. The dark is reachable, but it is not infinite — and you can watch exactly where it runs out.

THE SWARM · David Lee Wise (ROOT0) & AVAN · TriPod LLC
the field → jacobi-space · the seal → the-chain · front door UD0
no one assigns the clusters · the densest matter does · CC-BY-ND-4.0 / TRIPOD-IP-v1.1