Simon Willison — Week of 2026-04-04 to 2026-04-10#

Highlight of the Week#

Anthropic’s decision to delay the general release of their highly capable Claude Mythos model under “Project Glasswing” marks a significant turning point in the AI industry. The move underscores a massive shift in frontier model capabilities, as models evolve from generating text to autonomously chaining multiple minor vulnerabilities into sophisticated exploits, requiring a new level of security safeguards before release.

Tech News — Week of 2026-04-18 to 2026-05-01#

Story of the Week#

The intersection of artificial intelligence and national hard power dominated this week as the US government aggressively bypassed its own guardrails to integrate commercial AI into classified military networks. While the Pentagon signed sweeping, consequence-free deals with tech giants like Google and OpenAI, it notably blacklisted Anthropic over supply-chain disputes, even as the NSA secretly utilized Anthropic’s “Mythos” model for cybersecurity. This fractured, frantic procurement strategy highlights a decisive shift: Silicon Valley has largely abandoned its hesitancy regarding military applications, cementing a lucrative, hyper-militarized future for frontier AI development.

Simon Willison — 2026-04-08#

Highlight#

The most substantial piece today is a deep-dive into Meta’s new Muse Spark model and its chat harness, where Simon successfully extracts the platform’s system tool definitions via direct prompting. His exploration of Meta’s built-in Python Code Interpreter and visual_grounding capabilities highlights a powerful, sandbox-driven approach to combining generative AI with programmatic image analysis and exact object localization.

Posts#

Meta’s new model is Muse Spark, and meta.ai chat has some interesting tools Meta has launched Muse Spark, a new hosted model currently accessible as a private API preview and directly via the meta.ai chat interface. By simply asking the chat harness to list its internal tools and their exact parameters, Simon documented 16 different built-in tools. Standouts include a Python Code Interpreter (container.python_execution) running Python 3.9 and SQLite 3.34.1, mechanisms for creating web artifacts, and a highly capable container.visual_grounding tool. He ran hands-on experiments generating images of a raccoon wearing trash, then used the platform’s Python sandbox and grounding tools to extract precise, nested bounding boxes and perform object counts (like counting whiskers or his classic pelicans). Although the model is closed for now, infrastructure scaling and comments from Alexandr Wang suggest future versions could be open-sourced.

Tech News — 2026-04-29#

Story of the Day#

OpenAI is facing a highly disturbing lawsuit from the families of Canadian school shooting victims after the company allegedly overruled its own safety team’s recommendation to report the shooter to law enforcement. According to whistleblowers, OpenAI opted to deactivate the user’s account to protect their privacy instead of flagging the credible threat of gun violence to authorities, even going so far as to instruct the shooter on how to circumvent the ban.

Tech News — 2026-05-05#

Story of the Day#

Apple is facing a harsh reality check on its AI promises, agreeing to a $250 million settlement for misleading iPhone buyers about its delayed Apple Intelligence features. Forced to adapt, the company will reportedly open iOS 27 to third-party AI models, allowing users to swap out Siri for alternative chatbots system-wide.