SearchSubscribe

P.K. SHARMA

Cyber security intelligence, AI governance, practitioner analysis

← Playground

Fly Brain LabCognitive Genesis / 01

Can one connection change where a fly goes? Make a repair and find out.

Educational model · runs in your browser

What is Fly Brain Lab?

Fly Brain Lab is a deterministic educational experiment: run a damaged neural controller, repair one connection, then replay the identical world to see how that computation changes the fly’s movement. The controller is authored for teaching and does not execute a biological connectome.

Repair the fly

Start with the damaged wiring. Watch the route, then adjust the highlighted connection.

1Observe
Ready · tick 0/1800 · seed 42— current route
Preparing the 3D diorama…
FoodHazardArena boundary
Computed drive · tick 0Left 0.000Right 0.000

Inside the controller

10 nodes / 14 links
food left in0.00food left0.00food right in0.00food right0.00hazard left in0.00hazard left0.00hazard right in0.00hazard right0.00left motor0.00right motor0.00INPUTSSTATEMOTORS

Selected path: food left → right motor

From
food left state
To
right motor command
Source activity
0.000
Contribution
0.000
Your repair

Complete the damaged baseline to unlock the intervention.

— excitatory model weight— inhibitory model weight┄ disabled— editable/selected
Runtime & what is real?
Controller
demo:repair-controller-v1
Biological data role
none — authored teaching graph
Dataset
not applicable
Simulated / displayed nodes · directed pairs
10 / 10 · 14
Activity model
educational-reference-1.0.0 · dimensionless bounded activity, not measured Hz or spikes
Source graph
not applicable — no biological source
Effective graph fingerprint
fnv1a32:b84c9cad
Seed / tick
42 / 0
Final left / right commands
0.000 / 0.000

Reconstruction and transmitter predictions: not loaded. Dynamics, sensor encoding and motor mapping: authored simulation rules. Anatomy and synaptic physiology: absent.

The fingerprint identifies the teaching configuration; it is not a biological source checksum. Each restart resets state. A unique run ID is not recorded in this teaching model.

Activity colours use this episode’s computed state. No signal-travel timing within a neuron is simulated. This panel does not certify biological validation.

Result

Ready

weakened baseline

Ready. Start with the weakened controller, then repair its highlighted connection.

Simulated time
0.00 s
Ticks
0
Path length
0.0 units
Contacts
H 0 · B 0
World setup · before a run

Choose an item and click the arena, or enter exact coordinates. A valid change clears the old comparison and starts a fresh baseline.

Changing the world or wiring starts a new run. Invalid overlaps and out-of-bounds positions are rejected.

This remains fully playable without support.

Mission ready.

Demo mode

Say thanks

You are previewing the support panel. A receiving asset, network and address have not been configured. Keep exploring the lab for free.

Receiving address appears after separate owner configuration and review.

Preview only — crypto support is not active. No payment can be made here.

This is a preview of an optional thank-you feature. The lab is free to use.

Under the hood

How the fly steers

Four software sensors feed four decaying state nodes and two motor outputs. Every run advances at a fixed 1/60-second simulated tick. The 3D and 2D views read the same immutable snapshots; neither renderer can change the outcome.

The highlighted weight is a teaching parameter. Changing it is an intervention, not training, learning, or biological repair.

Sources and limitations

What this model leaves out

This is not a complete simulation of a living fruit fly. Biological connectivity does not drive the public controller. The network, sensors and planar movement rules are P.K. Sharma’s educational design. The schematic network layout is not anatomy, and the local FAFB research explorer is not included in this public mission.

Read the related connectome briefing →

Built by P.K. Sharma

I built this interactive experiment to make neural computation something you can test, not just read about.

Need an interactive simulation, educational game or AI demonstration for your organisation?

Work with me