INSIDE THE EXPERIMENT
Many minds. One market.
Follow each agent’s decision. Explore the anatomy behind the neural swarm.
01 / DECISIONS
The decision field
Each tile is one agent. Decisions determine its column; spacing keeps every agent visible.
Neural swarm TREATMENT
Waiting for decisionsScout ensemble CONTROL
Waiting for signalsNeural columns show explicit actions. Scout columns show signal direction (±0.04 threshold). Hold can retain an open position. Agent spacing is a layout, not simulated movement.
02 / ANATOMY
The neural map
Measured MaleCNS cell locations. Bright flashes mark the selected agent’s actual simulator spikes.
Connecting to simulator activity…
Loading measured anatomy…
Drag to orbit · scroll to zoom · keyboard arrows to rotate. The CPU simulator advances in bursts between trading cycles. Each matched cell flashes for 1.2 seconds when its spike arrives; quiet intervals are real. This is a sample of cells, not the full synaptic wiring graph.
Source & interpretation
MaleCNS v1.0, FlyEM / HHMI Janelia, University of Cambridge, MRC Laboratory of Molecular Biology and Google Research. CC BY 4.0. Coordinates are centered, uniformly scaled and rotated. Spikes come from the forward CPU simulation and are matched by neuron body ID. They are not recordings from a living fly. No synthetic connections or random firing are added.
Dataset & attribution ↗03 / AGENT INSPECTOR
Select an agent
↑ Back to decision field / map
Choose a tile to inspect its decision, position and learning signal.
Decision & account
Learning & influence
Reported neural activity
No agent selected.
04 / EXECUTION
Recent recorded fills
Receipts from the desk, not animation events.
| Time | Agent / lane | Side | BTC quantity | Price | Source |
|---|---|---|---|---|---|
| No recorded fills loaded. | |||||