# News: April 9, 2026

> 13 curated links on AI agents, Python, security, Google Cloud, and engineering, selected by Médéric Hurier.

- Author: [Médéric Hurier (Fmind)](https://www.fmind.dev/)
- Published: 2026-04-09
- URL: https://www.fmind.dev/news/2026-04-09/

- [Project Glasswing](<https://www.anthropic.com/glasswing>): Anthropic announces a major initiative to secure critical software using its new, unreleased Claude Mythos Preview model, capable of finding and exploiting complex vulnerabilities.
- [Claude Mythos is real and it scares me...](<https://www.youtube.com/watch?v=SQhfkWdxVvE>): Matthew Berman discusses the implications of Anthropic's new Mythos model and its unprecedented cybersecurity capabilities.
- [Meta is back! Muse Spark independent benchmarks](<https://x.com/ArtificialAnlys/status/2041913043379220801>): Meta's new Muse Spark model scores 52 on the Intelligence Index, marking a significant leap as their first non-open-weights frontier model.
- [Google prepares Jules V2 agent capable of taking bigger tasks](<https://www.testingcatalog.com/google-prepares-jules-v2-agent-capable-of-taking-bigger-tasks/>): Google is reportedly building a next-generation "Jitro" coding agent focused on high-level goal-setting and KPI-driven autonomous development.
- [MemPalace: High-scoring AI memory system](<https://github.com/milla-jovovich/mempalace>): An open-source, local AI memory system using a "palace" architecture and AAAK compression to provide persistent context without cloud dependencies.
- [Introducing Claude Managed Agents](<https://www.youtube.com/watch?v=I1BvAHOsjBU>): Anthropic introduces a new offering for businesses to deploy and manage AI agents at scale with built-in governance.
- [Envoy: A future-ready foundation for agentic AI networking](<https://cloud.google.com/blog/products/networking/the-case-for-envoy-networking-in-the-agentic-ai-era/>): Google Cloud explains how the Envoy proxy is evolving with protocol deframing to handle agentic traffic like MCP and A2A for governance and security.
- [Cloud Infrastructure in the Agent-Native Era](<https://services.google.com/fh/files/misc/cloud_infrastructure_in_the_agent_native_era.pdf>): A white paper outlining the shift in cloud infrastructure requirements to support the unique networking and security needs of autonomous AI agents.
- [Anthropic races to contain Claude Code leak](<https://futurism.com/artificial-intelligence/anthropic-suddenly-cares-about-intellectual-property-claude-leak>): Anthropic issues copyright takedowns after accidentally leaking the source code for its Claude Code agent through a misconfigured npm package.
- [How to add persistent memory to your AI agent](<https://www.youtube.com/watch?v=HDqzJJhZsxw>): Google Cloud Tech provides a guide on implementing persistent memory for AI agents using their managed infrastructure.
- [Google just casually disrupted the open-source AI narrative...](<https://www.youtube.com/watch?v=-01ZCTt-CJw>): Fireship breaks down how recent moves from major labs, including Meta's shift from open weights, are changing the developer ecosystem.
- [Google Cloud named a Leader in Sovereign Cloud Platforms](<https://cloud.google.com/blog/products/identity-security/a-leader-in-forrester-wave-sovereign-cloud-platform-2026/>): Forrester names Google Cloud a leader in sovereign cloud platforms, highlighting its "sovereignty-by-design" approach and AI capabilities.
- [Waymo’s skyrocketing ridership in one chart](<https://techcrunch.com/2026/03/27/waymo-skyrocketing-ridership-in-one-chart/>): Waymo reaches 500,000 paid robotaxi rides per week across 10 U.S. cities, showing a tenfold growth in ridership over two years.