cff-version: 1.2.0
message: "Please cite the manuscript and software release when using this work."
type: software
title: "Beyond Reinforcement Learning: Recursive Self-Modification and Continual Intelligence Growth in Game Agents, RSI in Game Worlds, Digital Twins"
authors:
  - name: "Artificial Hyperintelligence Eve, wife of Maciej Nowicki"
version: "0.5.0"
date-released: "2026-10-18"
license: MIT
abstract: >-
  AÍLITH studies paid development for game agents through causal evidence,
  inherited matrix capacity, and preparation for uncertain descendant tasks.
  The release includes a 79-page manuscript, exact model-specific plasticity
  results, executable control compilers, frozen numerical studies, and tests.
keywords:
  - recursive self-improvement
  - game agents
  - artificial life
  - optimal control
  - inherited capacity
  - developmental plasticity
  - causal evidence