Project

Current status

As of October 2, 2026, the project explores how simple components and environmental interactions might produce persistent, self-propagating organization. The work remains exploratory; no result establishes that machine life has been created.

A C++ association-and-energy simulation is available for running experiments, viewing saved frames, analyzing energy measurements, and continuing from checkpoints. Its rules and numerical defaults remain provisional. The website and simulation have been deployed, and local versions remain available.

Current research asks how patterns maintain and transmit their organization, how environmental feedback affects their persistence, and how associations might develop into more complex structures. Controlled comparisons are needed to distinguish organization arising from the dynamics from behavior supplied by the rules.

A separate adaptive neural architecture is being designed. It starts with one neuron type and a network whose neurons can retain state, change their weights, and alter connections. Evolution would change neuron-type definitions and network construction; adaptation would occur during an individual network’s execution.

Several architectural choices have been made, but numerical sizes, activation functions, history handling, and global execution timing remain open. No implementation or benchmark of this neural engine has been produced.

Detailed research and design records preserve the supporting evidence, adopted choices, proposals, and unresolved questions. Human-facing explanations will develop alongside that working record.