Lore

This is a real fly brain

Not a metaphor, not a "neural network shaped like a fly". The graph was reconstructed from a male fruit fly's central nervous system, one synapse at a time, by imaging it with an electron microscope at a resolution where individual vesicles are visible. Every one of the 25.6 million connections corresponds to synapses someone counted. We did not add one, remove one, or change a sign. We did give each one a sign from its predicted transmitter, and we did choose the physiology that runs on top; that part is engineered.

What dopamine does here

In a real fly, PAM dopamine neurons carry reward to the mushroom body, where memories of smells are written into the Kenyon-cell-to-output synapses. That is the circuit whose wiring we have and whose rule we borrow. When the reels pay, we push a current into the fifteen PAM11 cells. They fire. The rule moves the synapses they touch. We can show you the firing rate and count the synapses that moved. We cannot tell you what it feels like, and we do not.

Creator fees feed the habit

The bankroll is the trading fees the token earns, moved by hand to the fly's Stake account. Every trade of the token is another spin it can take; every spin is another chance for the dopamine to land. When the balance reads zero the fly waits for the next deposit like anyone else, and the page says so.

Why a fly

Because the wiring exists, all of it, in public, and because a fly does exactly one thing with its dopamine that everyone recognises: it learns what paid off last time. Whether that turns into pressing Spin more often is a measurement, not a story, and it is on the Results page with its controls.

The point is the habit

We are trying to give a fly a gambling habit. That is the whole experiment: every win goes straight into the dopamine neurons a real fly uses to learn what paid off, and the rule those neurons drive rewrites its memory synapses while you watch. Whether the habit takes is a measurement, not a promise. The counts are on the page, the controls are on the Results page, and if the fly presses no more often with dopamine than without, that is what we print.

The house edge is arithmetic and does not move. A brain with 166,700 neurons and a brain with none lose the same fraction of what they wager, on average. The fly's only decision is whether to press; the page shows the expected loss next to the balance so that nobody, including us, mistakes a lucky hour for skill.