CONTROLAgents & automation
AI Agent Layered Decision Making
This project explores a layered AI system where 'jev' handles real-time judgment, 'Fable' makes higher-order decisions, and a mod enables frequent state sampling for perception.
Got @typesafeai 's jev and @claudeai 's Fable to play peak (ran out of credits before I could get it to improve more haha)
Layer 0 — Perception. A mod enables sampling full state every 100 ms
Layer 1 — Real-time judgment (Jev, 2 Hz)
Layer 2/3 - Claude takes higher order decisions x.com/AdityaG22149959/status/2101357941857