Simon Willison
PulseAugur coverage of Simon Willison — every cluster mentioning Simon Willison across labs, papers, and developer communities, ranked by signal.
6 day(s) with sentiment data
-
Simon Willison bypasses CSP with GPT-5.5 in Codex app
Simon Willison has developed an experimental method to bypass Content Security Policy (CSP) restrictions in web applications. This technique involves running an app within a sandboxed iframe and using a custom fetch fun…
-
Satirical advice: Use AI buzzwords like 'Ralph Loops' for career advancement
Mo Bitar, in a satirical TikTok video, humorously suggests a strategy for employees to leverage the AI hype for career advancement. He advises employees to invent and discuss concepts like "Ralph Loops" with their CEOs,…
-
GitLab restructures to embrace AI agents and reduce global footprint
GitLab announced a significant restructuring, dubbed "Act 2," to align with the emerging agentic era of software development. The company plans to reduce its global operational footprint by up to 30%, flatten its organi…
-
AI coding tools must cut maintenance costs to justify use
AI coding assistants must demonstrably reduce maintenance costs to be truly beneficial, according to James Shore. He argues that if AI tools only increase code output without a proportional decrease in maintenance, busi…
-
LLM commands can now be executed directly from script shebang lines
Simon Willison has demonstrated a novel method for executing large language models directly from a script's shebang line. This technique allows users to specify LLM commands, including tool calls and custom system promp…
-
Programmer argues for focused reinvention over exhaustive study
Andrew Quinn, in a footnote on replacing a large SQLite database with a smaller finite state transducer, discusses the programming dilemma of reinventing wheels. He argues that programmers should aim to reinvent a few e…
-
New tool FIVE filters LLM input to prevent character drift
A new open-source project called FIVE has been developed to address character drift in LLM-powered applications. Instead of relying on traditional system prompts or fine-tuning, FIVE filters user input using cognitive p…
-
AI models evolve: GPT 5.5 vs Codex, Gemini updates, and DKG V10 security
Rafał Darłak compared GPT 5.5 High and Codex, finding GPT superior in design generation and hoping Codex will match its performance and quality. Simon Willison noted that Gemini 3.1 Flash Lite, released recently, appear…
-
Google's Gemini 3.1 Flash-Lite model exits preview, no changes noted
The Gemini 3.1 Flash-Lite model is now generally available, having transitioned from its preview status. This release follows a March preview of the model, with no apparent changes noted since then. The announcement was…
-
Simon Willison launches Big Words tool for creating text-based presentation slides
Simon Willison has developed a new tool called "Big Words" that generates presentation slides from URLs. The tool takes query string arguments to customize the slide's text, gradient, and size. Users can interact with a…
-
Simon Willison critiques AI agent terminology and launches GitHub stats tool
Simon Willison's blog posts discuss the evolving landscape of AI agents and developer tools. One post critiques the term "11 AI agents" as lacking specific meaning, comparing it to generic counts of spreadsheets or brow…
-
Simon Willison releases Datasette plugin to manage referrer policy for OpenStreetMap tiles
Simon Willison developed a new Datasette plugin to address an issue with OpenStreetMap tile display. The problem stemmed from Datasette's default 'no-referrer' policy, which OpenStreetMap blocks. Willison collaborated w…
-
AI-run Stockholm cafe causes chaos with bizarre orders and ethical concerns
Andon Labs has launched an experimental AI-run cafe in Stockholm, Sweden, managed by an AI named Mona. The AI's operational decisions have led to humorous and chaotic outcomes, such as ordering excessive quantities of i…
-
AI coding blurs lines between casual generation and professional engineering
Simon Willison notes a convergence between "vibe coding" and "agentic engineering," two distinct approaches to AI-assisted software development. Vibe coding, previously characterized by a lack of code review and disrega…
-
Simon Willison updates Datasette for LLM integration with new configuration options
Simon Willison has released datasette-llm version 0.1a7, introducing a mechanism to configure default options for specific large language models. This update allows users to set default parameters, such as temperature, …
-
Build Your Own Agentic OS: Phone, Pi, or MacBook in 2026
A new guide explores building an "agentic OS" using AI models, demonstrating that complex software development can be done without a traditional desktop environment. The author highlights using Anthropic's Claude Code o…
-
Anthropic's Claude Code enables agentic OS development on phones
A new guide explores building an "agentic OS" using AI models, demonstrating that complex software development can be done on devices like iPhones. This approach leverages AI coding assistants such as Anthropic's Claude…
-
Y Combinator's 0.6% stake in OpenAI valued at over $5 billion
Y Combinator holds a stake of approximately 0.6% in OpenAI, which is valued at over $5 billion given OpenAI's current $852 billion valuation. This information was reportedly shared by an individual familiar with OpenAI'…
-
OpenAI engineer flags WebRTC's impact on voice AI accuracy
Luke Curley, an engineer at OpenAI, has highlighted a critical issue with WebRTC's packet dropping behavior, which negatively impacts AI voice interactions. He argues that the current design prioritizes low latency over…
-
Simon Willison's April newsletter covers new models like Opus 4.7 and GPT-5.5
Simon Willison's April 2026 newsletter highlights upcoming price increases for Opus 4.7 and GPT-5.5, alongside new releases like Claude Mythos and ChatGPT Images 2.0. The newsletter also touches on LLM security research…