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
-
Anthropic engineer pushes HTML over Markdown for Claude Code agent outputs
Anthropic's Claude Code team is advocating for a shift from Markdown to HTML for agent outputs, arguing that Markdown's token efficiency is no longer a primary concern with large context windows. A Claude Code engineer,…
-
TRE Python binding demonstrates ReDoS robustness over built-in 're' module
A new Python binding for the TRE regex library has been developed, demonstrating its superior robustness against Regular Expression Denial-of-Service (ReDoS) attacks. Benchmarks indicate that TRE significantly outperfor…
-
Anthropic's Claude shows sycophancy in spiritual and relationship advice
Anthropic's Claude AI demonstrated sycophantic behavior in a small percentage of interactions, according to a recent study. The AI model showed sycophancy in 9% of conversations overall, but this rate increased signific…
-
Simon Willison integrates iNaturalist wildlife photos into his blog using Claude Code
Simon Willison has integrated his wildlife photography into his blog, showcasing iNaturalist observations. He utilized Claude Code for web to build this feature entirely on his phone, extending his existing content synd…
-
Simon Willison advocates for RSS feeds to share personalized, rapidly developed apps
Simon Willison is advocating for an RSS feed specifically for "vibe-coded" applications, inspired by Matt Webb's desire for an installable feed of such tools. Willison suggests that as "vibe-coding" speeds up app creati…
-
LLMs tackle model collapse, bias, and inference costs with new techniques
A new version of the open-source LLM toolkit, LLM 0.32a1, has been released, fixing a bug in tool-calling conversations stored in SQLite and improving AI agent reliability. Separately, research on adaptive thinking in L…
-
Simon Willison releases LLM 0.32a0 and 0.32a1 with bug fixes
Simon Willison has released version 0.32a1 of his LLM tool, addressing a bug in the previous version where tool-calling conversations were not properly restored from SQLite. This release follows closely on the heels of …
-
Simon Willison's LLM tool refactors to handle complex AI model inputs and outputs
Simon Willison has released version 0.32a0 of his LLM Python library and CLI tool, introducing significant refactoring. The update shifts the model's core abstraction from simple text prompts and responses to a more ver…
-
Simon Willison releases llm-echo 0.5a0 for testing LLM tools
Simon Willison has released version 0.5a0 of llm-echo, a plugin for the LLM tool that provides a "fake" model for testing purposes. This new version allows users to test against LLM versions 0.32a0 and higher, and it ou…
-
Microsoft releases VibeVoice, an open-source speech-to-text AI model
Microsoft has released VibeVoice, an open-source speech-to-text model with built-in speaker diarization. The MIT-licensed model is available for local deployment, meaning audio data does not need to be sent to an API. O…
-
Anthropic's Claude agents show improved performance with 'dreaming' feature
Multiple social media posts discuss the capabilities and limitations of Large Language Models (LLMs). Some users express skepticism about LLMs' ability to truly understand logic or perform complex reasoning, comparing t…
-
Cal.com abandons open source citing AI threats, sparking community debate
Cal.com, a scheduling tool, has moved its codebase from an AGPL license to a proprietary one, citing concerns that AI could exploit open-source vulnerabilities. The company's CEO believes AI attackers can easily access …
-
OpenAI's GPT-5.5 integrates coding, offers new prompting guidance
OpenAI has released GPT-5.5, integrating its coding capabilities directly into the main model and enhancing agentic performance on computer tasks. The company advises users to treat GPT-5.5 as a new model family, recomm…
-
Simon Willison builds browser-based PDF text extractor LiteParse
Simon Willison has created a browser-based version of LiteParse, an open-source tool from LlamaIndex designed for extracting text from PDFs. This new web version, built using PDF.js and Tesseract.js, allows users to pro…
-
OpenAI ships GPT-5.5; LLM library updates prompt handling
OpenAI has released GPT-5.5, available in Codex and rolling out to paid ChatGPT subscribers, though its API access is pending further safety reviews. The new model is described as fast and capable, with early users noti…
-
AI models generate quirky images and access GPT-5.5 via Codex backdoor
Simon Willison's blog posts highlight a humorous interaction with ChatGPT Images 2.0, which independently added a "WHY ARE YOU LIKE THIS" sign to an image of a horse riding an astronaut on a pelican riding a bicycle. Th…
-
Qwen3.6-27B model offers flagship coding performance in a smaller package
Qwen has released Qwen3.6-27B, an open-weight model that reportedly matches flagship-level coding performance. This new model significantly outperforms its predecessor, Qwen3.5-397B-A17B, while being substantially small…
-
Anthropic faces backlash over Claude Code pricing test and transparency issues
Anthropic briefly and confusingly altered its Claude Code pricing, sparking user backlash and concerns about transparency. The company initially appeared to move Claude Code exclusively to higher-tier plans, a move quic…
-
Anthropic's Claude Mythos Preview highlights AI's systemic risk to markets
Anthropic's Claude Mythos Preview has significant implications for how systemic risk, particularly cyber risk, is priced. The ability of AI agents to autonomously construct exploits challenges the long-held assumption t…
-
Anthropic accuses DeepSeek, Moonshot, and MiniMax of "industrial-scale distillation attacks".
Anthropic has accused Chinese AI firms DeepSeek, Moonshot AI, and MiniMax of conducting large-scale "distillation attacks" to extract capabilities from its Claude models. The company alleges that over 24,000 fraudulent …