Simon Willison
Technical BlogDeveloper and open-source maintainer focused on practical LLM tooling, datasets, and experimentation.
Simon Willison is a prolific developer-writer who documents how to actually build with large language models using small, composable tools. He maintains projects that make it easier to run local models, explore SQLite at scale, and chain prompts without unnecessary ceremony, often releasing a tight library alongside a blog post that explains the design trade-offs.
Willison’s essays are grounded in hands-on notes: rate limits encountered in APIs, surprising tokenizer edge cases, prompt patterns that degrade with tool calling, and the maintenance burden of supporting multiple model providers. That honesty about friction makes his writing unusually actionable for engineers shipping features rather than chasing leaderboard scores.
He is worth following if you want a steady stream of pattern libraries for LLM-powered apps—especially around data journalism-style analysis, CLI utilities, and open datasets. His blend of curiosity and craft is a reliable antidote to hype-driven “just add AI” product advice.