Stories about LLM and Agents.
What makes up an agent, and what makes it a single individual? Starting with context and runtime, this article explores agent implementation, state management, and lifecycle, as well as the tension between freedom and manageability.
and many of my thoughts on general-purpose agents
Context = Grounding + Getter
What kind of API do agents like?
broadly injected memory for global context, and agent-driven retrieval for specific goals.
Engineering is still about solving today's problems under today's constraints.
Tools, MCP, Bash, Filesystem, OS, and Agent Skills
There is no silver bullet for LLM structured output.
If we want a problem-solving 'industrial robot' agent, this kind of memory is neither consistent with the software philosophy we appreciate nor technically sound. What the 'industrial robot' agent needs to design is state.
Trying to build AGI with only two methods!