# News: May 4, 2026

> 11 curated links on AI agents, Python, security, Google Cloud, and engineering, selected by Médéric Hurier.

- Author: [Médéric Hurier (Fmind)](https://www.fmind.dev/)
- Published: 2026-05-04
- URL: https://www.fmind.dev/news/2026-05-04/

- [Building with Gemini Embedding 2: Agentic multimodal RAG and beyond](<https://developers.googleblog.com/building-with-gemini-embedding-2/>): Gemini Embedding 2 is now GA, offering a natively multimodal embedding space for text, images, video, audio, and documents across 100+ languages to power agentic RAG and search.
- [Architectural Evolution of Cloud Run for the Agentic Era](<https://medium.com/google-cloud/architectural-evolution-of-cloud-run-for-the-agentic-era-28af995cecb0>): Google Cloud introduces new Cloud Run primitives including long-lived instances, gVisor-backed sandboxes, and NVMe-backed ephemeral storage to support stateful, zero-trust agentic workloads.
- [Gemini CLI v0.40.0: Tiered Memory, Gemma, and Multi-Repo Support](<https://www.reddit.com/r/GeminiCLI/comments/1t0589f/gemini_cli_v0400_tiered_memory_gemma_and/>): The latest Gemini CLI update introduces a sophisticated tiered memory system, support for local Gemma models via Ollama, and improved multi-repository orchestration.
- [Building with Gemini Embedding 2: Our first natively multimodal embedding model](<https://www.youtube.com/watch?v=_en7QoZ_utc>): A technical overview of Gemini Embedding 2, demonstrating how it maps multiple modalities into a single semantic space for complex retrieval tasks.
- [How GPT, Claude, and Gemini are actually trained and served – Reiner Pope](<https://www.youtube.com/watch?v=xmkSf5IS-zw>): Google engineer Reiner Pope discusses the underlying infrastructure, training methodologies, and serving optimizations for modern large language models.
- [One tool to rule them all: Extending and customizing the Gemini CLI](<https://www.youtube.com/watch?v=1RXm-E4MEZ4>): This session explores the extensibility of the Gemini CLI, focusing on custom slash commands, MCP server integration, and agent personalization.
- [Operationalize AI: A blueprint for managing enterprise agents at scale](<https://www.youtube.com/watch?v=KE2B2qVzaxk>): Google Cloud share strategies for managing, monitoring, and scaling AI agents within enterprise environments using the Gemini Enterprise Agent Platform.
- [What's new with data agents](<https://www.youtube.com/watch?v=Z-AfOcWO_kk>): An exploration of newest features for data-centric agents, including enhanced BigQuery integration and automated data reasoning capabilities.
- [Google Workspace Studio Updates: Google Cloud Next Special Edition](<https://www.youtube.com/watch?v=FmQIVwpBCME>): A summary of AI-powered updates for Google Workspace, including new developer tools and integrations for Gemini.
- [New ways to customize AI-generated meeting notes](<https://workspaceupdates.googleblog.com/2026/04/new-ways-to-customize-ai-generated-meeting-notes.html>): Google Meet users can now toggle specific sections in AI-generated notes, with an improved "Decisions" section to track outcomes effectively.
- [Agentic Coding is Burning Me Out](<https://0xsid.com/blog/agentic-coding-fatigue>): An insightful piece on the "decision fatigue" triggered by managing multiple AI agents and the shifts in mental load for modern software engineers.