Simon Willison ships LLM 0.37, using frontier models to triage 12 contributor PRs
Tactic · Blog · stat: 12 fixes Simon Willison released version 0.37 of his open-source LLM CLI tool, adding default support for gpt-6-luna and multimodal embeddings. To manage a wave of AI-assisted…
Tactic · Blog · stat: 12 fixes
Simon Willison released version 0.37 of his open-source LLM CLI tool, adding default support for gpt-6-luna and multimodal embeddings. To manage a wave of AI-assisted pull requests without maintainer burnout, Willison used Opus 5.5 and GPT-6 Astra to triage incoming code, ultimately merging contributions from 12 authors.
Using LLMs as triage agents makes AI-generated PR floods manageable for solo maintainers. Maintainers can counter incoming AI-generated contribution noise by using frontier models as automated PR triage screeners.
Every claim ties to a primary source. See our methodology.