Designing an AI Native System from Scratch
Instead of retrofitting agents into existing systems, I discuss what it means to design a system for agents from the ground up and how agent organization should be structured.
Agent
The Missing Piece of AI Agents: Knowledge Infrastructure
The main reason AI agents fail is the lack of knowledge infrastructure in existing systems.
A Conceptual Framework for LLM-Based Agent Systems
I attempt to build a framework-agnostic conceptual model that covers the core components of LLM-based agent systems and the relationships between them.