TIN HAT CAT

QUESTION EVERYTHING.

$THCROBINHOOD CHAIN

THE PEOPLE'S CAT.
THE INTERNET'S CONSPIRACY CLUB.
EVIDENCE, NOT VIBES.

WIRE

EVIDENCE.
NOT
VIBES.

SAME CAT.
BIGGER
RABBIT HOLES.

CURIOSITY
IS A
WEAPON.

🐱TinHatCat.com 📈DexScreener 👕Merch

📐DOCUMENTDECLASSIFIED — MOSTLY

TIN BRAIN.exev0.01 [BETA]

TIN HAT CAT PRESENTS

TIN BRAIN.exe

A SLIGHTLY SMARTER CAT FOR A DUMBER WORLD.

v0.01 — BETA

PROJECT TIN BRAIN

THEY MKULTRA'D A FLY.

Somebody mapped a fruit fly's brain down to the last wire — 138,639 neurons, 15,091,983 synapses — and published it for free. We did not ask what it was for. We asked what it could be conditioned to do.

THEN WE PUT IT IN THE CAT.

SUBJECT 7 carries the whole connectome under the foil. It does not reason. It reacts — and a reaction you can measure is a reaction you can train. Dopamine in, synapse down, opinion changed. That is the entire program.

THE ASSET IS OPERATIONAL. THE HANDLER IS YOU.

You pick the target. The asset twitches. What that twitch means is a human's job, and the cat will never pretend otherwise. Every case stays open until somebody brings receipts.

WHAT YOU CAN DO

🥫BRIEF THE ASSET

Hand it a name, an event, a question. It becomes an odour and goes straight into the implant. Public sources only.

🌀SEE WHAT GIVES IT THE TINGLES

Every target ranked by how hard SUBJECT 7 reacted — and which other targets smell the same to it.

📁PULL THE CASE FILES

Response, training curve, nearest match. Including the targets it flatly refused to learn.

⚡RUN THE CONDITIONING

Pair a target with punishment and watch the readout drop. Sometimes it drops. Sometimes it gets louder. Both are real results.

BEFORE YOU GO IN

⚠ THE PROGRAM IS A BIT. THE BRAIN IS NOT.

Real: the wiring — 138,639 neurons and 15,091,983 synapses from the FlyWire connectome — and the leaky integrate-and-fire simulation running on it. The conditioning is real too: dopamine depressing Kenyon-cell synapses is how a fly actually learns. Those spikes are computed, not drawn.

Invented: the smell, and the spy story around it. We assign each target an odour and watch the brain answer. A fly has no concept of a conspiracy, has never been recruited, and is not in the cat. The response is real; the label is ours.

What it is not: research. Nothing is investigated or verified, and no output is an accusation about anyone. Question everything, including this.

or press ENTER

A SLIGHTLY
SMARTER CAT
FOR A DUMBER
WORLD.

TIN BRAIN LIVE

GOOD CATS
QUESTION
EVERYTHING.

LARPTOP
IS REAL.

😼 MOOD: IDLE — waiting for a stimulus

SHOW ME
THE RECEIPTS.

CURRENT OBSESSION

ROBINHOOD

—

RESPONSE STRENGTH (peak MBON z-score)

NEURAL ACTIVITY — olfactory learning circuit

8,572NEURONS SIMULATED

606,719SYNAPSES

—MEAN RATE, ACTIVE CELLS

STIMULUS → DECISION → ACTION

RUN LOG

LIVE FEED

🌀

RABBIT HOLES

Targets ranked by how hard SUBJECT 7 reacted. A strong reaction means interesting, never true.

SIGNATURE MAP — which stimuli the connectome answers alike

drag to pan · shift-drag to spin · scroll to zoom · double-click to reset

similar response node size = response strength

🥫

BRIEF THE ASSET

Hand SUBJECT 7 a new target. It becomes an odour, and the implanted connectome answers it.

YOUR TOPIC

CATEGORYHANDLE (OPTIONAL)

The category is part of the stimulus, not a label — it is folded into the hash, so the same topic files under Crypto and UFO produces two different smells and two different cases.

SHOW IT A PICTURE (optional — it really does see this)

CHOOSE AN IMAGE — OR DROP ONE HERE

The fly really sees this. Your picture is reduced to 32×32 in your browser and those pixels are sampled onto all 10,582 photoreceptors by their own positions on the retina — R1-6 take luminance, R8 green, R7 UV. Dark drives them hardest, because that is how fly photoreceptors work. The original file never leaves your device, only the small grid does, and that is deleted the moment the retina has been driven. Nothing is stored, displayed or published except the response.

OR GIVE IT SOMETHING TO READ (optional — a link to a page or document)

The server fetches this link, strips the markup, and takes a SHA-256 of the text. That digest goes into the odour, so the same document always smells the same and a different one never does. The document itself is not kept — the case page shows the link, the digest, the length, and a short excerpt. Private and local addresses are refused.

HOW BIG A JOB?

BRIEF THE ASSETOne target. One connectome run.~$0.75 FEED A URL / DOCUMENTAdds a link. The server reads the page and folds a digest of its text into the odour.~$1.50 DEEP RABBIT HOLEEverything above, a 10-trial training curve, and your signature correlated against every published case on file.~$4.00 PRIORITY CASEEverything above, a 14-trial training curve, and first place in the run queue ahead of every cheaper job.~$12.00

Show my handle next to the run if it gets simulated

Your run starts the moment the payment lands and takes about six seconds. Nobody here investigates anything. A fly smells your theory and we write down what its neurons did about it.

YOUR TOPIC SMELLS LIKE THIS — computed in your browser, no server

These are real Drosophila olfactory receptor channels. The same topic always produces the same smell — the encoder is SHA-256, nothing random.

YOUR LEADS

Everything this wallet has sent. Click one to pay for it or to open its case file.

Have a claim code from another device?

💰

BOUNTIES

Things the cat wants brought in. Bring one and get paid in $THC. Entering is free — you are answering a request, not buying a run.

Loading…

📁

CASE FILES

Debriefs. Target, odour, and how the asset answered. Status describes the response, never the claim. Every case re-runs from its seed.

Nothing matches that. The cat has no opinion it hasn't already filed.

🧠

ASSET CAM

Real connectome. Real simulation. The smell is our invention.

—

ORN input ALPN relay KC sparse code MBON readout DAN teaching

The measured circuit. Every number on this page is computed here.

STIMULUS

—

ENCODED AS ODOUR (receptor channels)

THE TINGLE (MBON response vs baseline)

TEACH IT (dopamine depresses the active synapses)

trial 0

Wiring and spiking are real: 138,639 neurons, 15,091,983 synapses, leaky integrate-and-fire. The mapping from a topic to a smell is invented and deterministic — a fly has no concept of a conspiracy. The response is real; the label is ours.

⚙

SETTINGS

Preferences, and where the brain came from.

SPIKE AUDIOSonify the spike train. Each click is a real population firing event from the simulation. MOTIONIdle animation and the rotating connectome view. PLAY INTRO EVERY TIMEThe boot sequence normally plays once, on your first visit. Turn this on to see it every time TIN BRAIN.exe opens — or just hit ↺ REPLAY INTRO in the sidebar when you want it.

WALLPAPER Pick the desktop behind the window. Saved on this device only.

SHOW ME HOW IT WORKS A short walk through the real interface — what the numbers mean, how to feed the cat, and where your leads end up. It runs once on your first visit; after that it only appears when you press this.

WHERE THE BRAIN CAME FROM

The connectome and simulation on this site are real. Data is used with thanks under the terms below.

Connectivity & LIF model — Shiu et al., philshiu/Drosophila_brain_model (MIT). A leaky integrate-and-fire model of the Drosophila brain.
Neuron annotations — Schlegel et al., “Whole-brain annotation and multi-connectome cell typing of Drosophila”, Nature (2024); optic lobe annotations from Matsliah et al., Nature (2024).
Connectome — the FlyWire FAFB 783 public release.

Two runs, and only one of them counts. Everything measured — the tingle, Kenyon sparsity, the learning curve, teachability — is computed on the olfactory learning circuit: ORN → antennal lobe projection neurons → Kenyon cells → MBONs, with dopaminergic and APL neurons. 8,572 neurons, 606,719 synapses.

Brain Cam can also show a second run over the whole connectome (138,639 neurons, 15,091,983 synapses, LIF with a 1.8 ms synaptic delay and a 5 ms alpha synapse). That one is for looking at, not for measuring. The full network cannot hold biologically sparse Kenyon coding and a clean mushroom-body readout at the same time — pushing sparsity down with APL inhibition drags the readout down with it, and the best whole-brain setting found was still about six times too dense. The circuit does both at once, so the circuit stays the instrument.

◷

WHILE YOU WERE GONE

What the connectome has actually been run against.

TIN HAT CAT 98

🐱TinHatCat.comthe main site 📈DexScreenerchart

👕Merchriseattireusa.com

● SYSTEM ONLINE