Week 05: Agency — Autonomous Agents, Tool Use, and Multi-Agent Systems

Description

Designing autonomous AI agents capable of reasoning, reflection, tool execution, and interacting with external environments.

Key Topics

• Agent Architecture: Perception $\to$ Memory $\to$ Planning $\to$ Tool Execution $\to$ Evaluation.
• Model Context Protocol (MCP): Connecting LLMs to filesystems, databases, browsers, and creative software.
• Participatory Design: Structuring multi-agent debates and autonomous collaborative creative systems.

References & Wiki Links

• Human-Computer Team Automation
• Participatory Design
• Recursive Self-Improvement