Simon Willison — Independent developer and writer documenting practical experiments, tools, and deep analysis around large language models, generative AI, web development, security, and emerging programming workflows through detailed posts and daily TILs.
381 articles from this blog
Release notes for llm 0.31.1, a command-line tool for accessing large language models, with a bug fix for OpenAI Chat Completion endpoints.
Jarred Sumner details the rewrite of Bun from Zig to Rust using AI coding agents, highlighting memory safety and test-driven porting.
OpenAI upgrades ChatGPT voice mode with GPT-Live, using GPT-5.5 for complex tasks and improved conversation flow.
Kenton Varda criticizes AI-written change descriptions as worse than useless, lacking high-level context for PR reviews.
sqlite-utils 4.0 release introduces database schema migrations, nested transactions, and compound foreign keys for SQLite.
Announcement of sqlite-migrate 0.2, a simple database migration system for SQLite based on sqlite-utils.
An experimental Web Component for embedding code from GitHub using GPT-5.5, fetching raw files and displaying line ranges.
Announcement of sqlite-utils 4.0, a Python CLI and library for SQLite databases with new schema migration features.
Release announcement for sqlite-utils 4.0rc4, a Python CLI and library for manipulating SQLite databases, with feedback from Claude Fable 5.
Tencent releases Hy3, a 295B-parameter MoE model outperforming larger open-source models, available for free on OpenRouter.
sqlite-utils 4.0rc3 release adds compound foreign key support and case-insensitive column name handling.
Release of micropython-wasm 0.1a2, a Python library for running a MicroPython sandbox using WebAssembly, with a new CLI.
Using MicroPython compiled to WebAssembly to run Python code in a secure sandbox for plugins.
OpenAI introduces Lockdown Mode to prevent data exfiltration from prompt injection attacks in ChatGPT.
Andreas Kling explains Ladybird browser's policy change to stop accepting public pull requests for quality and accountability.
Explores the tension between AI enthusiasts racing to adopt AI and skeptics warning about reliability and knowledge loss.
A quote from Emanuel Maiberg about Google employees sharing memes criticizing its AI, with a note on Google's statement change.
Uber caps employee AI coding tool spending at $1,500/month per tool to manage costs after overspending its 2026 AI budget.
Microsoft announces MAI-Thinking-1 and MAI-Code-1-Flash LLMs, with details on parameters, licensing, and training data.
Announcing datasette-agent-micropython 0.1a0, a MicroPython WASM sandbox for safe code execution in Datasette Agent.
