This research introduces Cohen’s Agentic Conjecture (CAC), proposing that an artificial intelligence system integrating fast, neural heuristics (System 1) with slow, symbolic logic (System 2) through a dynamic gating mechanism can exhibit emergent agentic properties. These properties include context-aware decision-making, self-directed learning, robust reasoning, and reflective self-correction. Drawing inspiration from dual-process cognitive theories and neuro-symbolic AI paradigms, this work formalizes CAC, presents a comprehensive Python implementation, and validates the conjecture through empirical experiments. The findings demonstrate that CAC-enhanced systems outperform purely neural or purely symbolic counterparts in terms of accuracy, interpretability, and adaptability. This framework lays the groundwork for developing next-generation AI agents capable of autonomous, reliable, and
This tutorial is based on a clean Ubuntu-22.04 install. Use at your own risk.
- WSL2 Installed. Follow this tutorial: https://learn.microsoft.com/en-us/windows/wsl/install#install-wsl-command
- systemd enabled. Follow this blog post: https://devblogs.microsoft.com/commandline/systemd-support-is-now-available-in-wsl/#set-the-systemd-flag-set-in-your-wsl-distro-settings