EXPERIMENT 03 / NON-NEURAL CONTROL

A rule without a neural net.

A familiar local rule makes the difference between cellular and neural visible.

See the execution path ↗

EXPERIMENT 03 / NON-NEURAL CONTROL

Game of Life control

FIXED RULE / NO TRAINING

Your experiment starts here.

Select a compute backend and start the engine.
No server, account or model API required.

Click to toggle a cell while pausedPopulation: —

Steps / bytes—

Completed graphs—

Last command—

Shared weightsNone

Ready when you are. Nothing is running yet.

Python / ZIPP WASM · compute backend not initialisedPrivate session · no prompt upload

What you are looking at

Cellular does not always mean neural.

Cells survive with two or three live neighbours and are born with exactly three. Python submits the fixed local-rule graph through ZIPP to the selected backend. There are no trained parameters and no neural learning in this control.

48 × 48 cellsToroidal boundariesB3 / S23 rule

SCOPE & LIMITATIONS

This is a classical cellular automaton, not a trained neural cellular automaton.

Read the research context ↗Reproduce & inspect the source ↗

Inspect the Python that runs in ZIPP

Actual bundled source. Matrix graphs use the selected backend; this viewer is read-only.

Download life.py

Start the engine to view the verified source, or use the download link above.

Explicit backend. No silent substitution.

WebGL2 needs supported hardware and floating-point render targets. CPU reference is an intentional fallback, not a GPU result.

Live state, not a benchmark claim.

Colours visualise actual state. Per-command timings include host overhead and readback; they are not hardware comparisons.

Original checkpoints. Bounded sessions.

Python runs in a worker with execution limits. Stop engine terminates the worker and discards the current state.