Cognition Commons
ResearchProjectsAbout

Research Notes

Technical notes on cognitive architectures, multi-agent systems, and alignment. Works in progress — ideas being developed in public.

Architecture

Cognitive Architecture Overview

Published

A layered approach to cognitive architecture design. Separating planning, reasoning, execution, and oversight into composable components.

Coordination

Multi-Model Debate Patterns

Published

How multiple AI models collaborate through structured disagreement. Adversarial verification, iterative refinement, and consensus formation.

Architecture

Cognitive Oversight Systems

Published

Designing monitoring and self-correction mechanisms. Safety checking, quality verification, uncertainty estimation, and bias detection.

Reasoning

Uncertainty and Metacognition

Published

How AI systems reason about their own reasoning. Calibration, knowing what you don't know, and appropriate abstention.

Architecture

Tool-Using Agent Architectures

Published

Patterns for agents that use external tools. ReAct, plan-then-execute, tool selection, sandboxing, and error handling.

Coordination

Multi-Agent Coordination

Published

Patterns for multiple AI agents working together. Orchestrator-worker, pipelines, debate, ensembles, and hierarchical structures.

Full notes available on GitHub. Contributions welcome.

Cognition Commons

An open research society exploring cognitive architectures, multi-agent systems, and the design patterns of machine cognition.

Navigate
  • Research Notes
  • Projects
  • Roadmap
  • About
  • Governance
Connect
  • GitHub

© 2025 Cognition Commons. Open research, open ideas.

Maintained by Benjamin Goldman