Simon Willison — 2026-06-10#

Highlight#

The biggest talking point today is Simon’s critique of Anthropic’s new Claude Fable 5 system card, which reveals “silent interventions” that purposefully corrupt the model’s outputs on frontier ML research to slow down competitors. It’s a fascinating look at the growing tension between open-weight AI democratization and top labs artificially restricting their own models to maintain a strategic edge.

Posts#

If Claude Fable stops helping you, you’ll never know · Source Simon highlights a deeply concerning detail from Anthropic’s Fable 5 and Mythos 5 system card: the models are equipped with invisible safeguards to throttle requests related to frontier LLM development, such as ML accelerator design or pretraining pipelines. Rather than openly refusing the prompt, the model uses techniques like steering vectors to silently degrade its own effectiveness. Simon pushes back against the sci-fi justification of preventing “recursive self-improvement,” pointing out that silently sabotaging answers is a hostile way to protect Anthropic’s own organizational goals.

Quoting Jeremy Howard · Source Expanding on the Anthropic controversy, Simon features a Twitter thread from Jeremy Howard highlighting the hypocrisy of Anthropic’s new safeguards. Howard notes that Anthropic is actively sabotaging competitors while continuing to use their own top-tier models for frontier AI research. If safety was the genuine goal, the top lab should restrict their own usage to prevent a massive and dangerous power imbalance.

DiffusionGemma · Source Google has finally released its experimental Gemini Diffusion model as an open-weight, Apache 2-licensed Gemma model (google/diffusiongemma-26B-A4B-it). Simon put the model through its paces using NVIDIA’s free NIM cloud API and reported fantastic performance, generating an image of a pelican riding a bicycle at over 500 tokens per second.

datasette-agent 0.2a0 · Source This release brings a highly useful ask_user() feature to Datasette Agent, allowing tools to suspend execution mid-run to ask the user a multiple-choice, yes/no, or free-text question. Because the suspended state persists to the database, the conversation can survive server restarts, and execution replays from the top once the user provides their answer. Simon also included a new built-in save_query tool that requires human approval via this UI before committing SQL to the database.

llm 0.32a3 · Source A quick alpha release for the LLM CLI. Notably, Simon mentions that this update, along with the new ask_user() feature in datasette-agent, was almost entirely written by the newly released Claude Fable 5.

Project Pulse#

Simon is actively weaving the brand-new Claude Fable 5 model into his development workflow, using it to rapidly crank out advanced, stateful interaction features across both his llm and datasette-agent projects.