Anthropic's new Claude Sonnet 5.5 model offers significant speed and cost improvements, performs nearly as well as Opus 5.5 on some tasks, and is now the default for the free claude.ai tier.
GitHub details how their open-source AI security agent successfully identified 24 critical Android vulnerabilities, showcasing the power of AI in automated security testing.
GitHub provides an overview of the most interesting features and changes introduced in the recently released Git 2.56, a core tool for developers.
Simon Willison quotes OpenAI's Agent Security lead, Joe Daroo, on the unexpected and rapid advancements in AI capabilities and the organizational challenges they pose for security and resilience.
OpenAI shares its initial guidelines for establishing safety cases in frontier AI training, covering technical safeguards, operational practices, and incident investigation.
This tutorial demonstrates how to integrate Vercel's AI SDK with Cloudflare Durable Objects to enhance user experience within a fitness tracker application.
This alpha release of pnpr improves resolver security by blocking private network connections and enhances developer experience by linking workspace projects via `publishConfig.directory`.
