01 / CELLULAR MEMORYTRAINED CHECKPOINT
Memory, cell by cell.
Write a fresh association. Follow a query around a ring. Discover what disappears when private memory is erased.
Explore cellular memoryTHE RESEARCH PROGRAMME
Small experiments.
Explicit questions.
Two independent models explore memory and language. The method is the same: expose the mechanism, keep the controls, and document the boundary.
02 / CAUSAL LANGUAGESEPARATE MODEL
A local rule for the next byte.
An attention-free model reuses a small neural update to generate template text, one byte at a time.
Explore causal languageFIELD GUIDE / 01
What are neural cellular automata?
A large system. A small shared rule. A different way to think about computation.
Read the guide →EVIDENCE / RECORDED BASELINES
Results, with the conditions attached
Useful results are more than a headline number. They include the task, controls and what was not tested.
Read the guide →METHOD / REPRODUCIBILITY
Make the mechanism inspectable
Separate what changes, test what matters, and keep a trail back to the source.
Read the guide →RESEARCH / OPEN QUESTIONS
Questions worth testing next
The next step is a better experiment—not a bigger claim.
Read the guide →IMAGE RESEARCH / 03
NCA image generation: LATTICE and a trained browser proof
Train a shared local rule. Follow an image from noise. Measure the limits before scaling.
Read the guide →