Create DECISION_LOOP.md

Add symbolic memory buffer, rule inducer, and boredom detector

Add undo reasoning (ATT), symbolic memory buffer, boredom detector, and rule inducer

Add smart cell-select: affordance map, budget manager, conditional mechanics, irreversibility detection

Add v2 architecture document with competition strategy and roadmap

Add v2 action effects tracker and goal-directed selector

Add v2 CEM planner for model-based planning

Add v2 exploration module with systematic action scanning and state graph"

Add v2 world model with CNN encoder + RSSM dynamics (DreamerV3-style)

Add v2 agent: online world model, systematic exploration, CEM planning, action semantics

Add v2 agent: online world model, systematic exploration, CEM planning

Upload train_fast.py

Upload train_arc1.py

Upload train_arc1.py

Upload README.md

Upload src/models/ttt_adapter.py

Upload src/models/agent.py

Update architecture docs with state persistence and uncertainty

Update README with launch insights - state persistence and TTT critical

Add geometric augmentations

Add ARC dataset loader

Add integration test

Add ARC training script

Add grid masking strategy

Add full Grid-JEPA system

Add action-conditioned predictor module

Add RSSM dynamics model

Add Grid-JEPA encoder module

Add requirements

Add architecture documentation

Add references documentation

Initial project README

initial commit