Simon Willison comments on Anthropic's Project Glasswing, an initiative to limit access to their powerful Claude Mythos model exclusively to security researchers, emphasizing its necessity for AI safety.
Simon Willison discusses GLM-5.1, a new development in AI research focused on improving models' ability to handle complex, multi-step tasks over extended periods.
This JavaScript Weekly issue highlights Google's new JSIR for advanced tooling, the release of browser performance benchmark JetStream 3, and provides a comprehensive overview of the current JavaScript ecosystem.
Discover how `container-name` and `@container` queries provide a robust and intuitive way to scope styles directly to individual components, simplifying CSS management.
Explore various modern CSS techniques, including cascade layers and specificity strategies, to replace the `!important` keyword with more predictable and maintainable styling solutions.
Learn the best practices for implementing and utilizing subagents within Claude Code, enhancing the capabilities and complexity of your AI agent workflows.
Explore the year-long process of designing and coding R—K ’26, a portfolio that seamlessly blends identity, rhythm, and interaction to create a strong digital presence.
Victor Yocco explores the nuanced approach to designing transparency in agentic AI, focusing on identifying key decision points to build user trust through clarity rather than overwhelming data.
This minor release of Vitest brings new experimental features like `preParse` flag and custom snapshot matchers, alongside numerous bug fixes to improve stability and developer experience.
