LLM is evolving into an agent framework, driven by Datasette Agent's needs, with tool chains that can pause for human approval and resume from stored message history — and the agent concept may be baked into the core library next.

Many lower-level changes in 0.32 were driven by Datasette Agent's requirements. The author now sees LLM as agent-shaped and is considering baking the agent concept into the core library in a future release. ✦ AI generated

Simon Willison · Simon Willison's Weblog · 2026-08-04 · original ↗

Quite a few of the lower-level tools changes in this release were driven by the needs of Datasette Agent. When I started work on LLM, the term "agent" had such a vague definition that I refused to use it. In September 2025 I came around to the idea that "An LLM agent runs tools in a loop to achieve a goal" is well established enough now that I could stop avoiding the term entirely. Tool chains can now pause for human approval and resume from a stored message history - both needed by Datasette Agent. Looking at LLM today it's beginning to look very agent-shaped to me. There's something neat about having a CLI utility that can mix and match different tools from different sources with different models all as a one-liner, and that includes a Python library powerful enough to build systems like Datasette Agent and llm-coding-agent. Maybe the next version of LLM will bake the concept of an "agent" into the core library. I'm still trying to figure out what that would look like.

Read full article ↗excerpt · fair-use quotation