In the air flying home after a solid week in Manhattan. Doing some research into MCP during the flight, and glad to be able to search my own reading list for past links to review!
[blog] Managing Links with Gemini: Validating Links for Semantic Correctness. This is super cool work from our documentation team. Make sure your links go where you think they do.
[blog] NotebookLM: One-Click Accident Report Comprehension. This tool is so useful for understanding complex topics. The ability to generate videos and mind maps ensures a few ways you can digest the content.
[paper] A Survey of Context Engineering for Large Language Models. This paper looks at the emerging discipline of context engineering, some patterns, and where gaps exist.
[blog] AI agents vs. predefined workflows: practical decision guide. Everything doesn’t need to be an agent. This post from Capital One points out when to use an agent, and when an AI workflow will do.
[blog] Mastering agentic workflows with ADK for Java: Sub-Agents. Excellent post that will help you better understand the ideas behind agents, subagents, agents-as-tools, and coordination of it all.
[blog] 25+ top gen AI how-to guides for enterprise. We’re awash in tutorials and how-to material. Here’s an attempt to pull together a handful of representative guides.
[blog] My AI Code Reviewer Needed a Project, So I Vibe-Coded One for It. I had a lot of code review content yesterday, and here’s another. This offers some prompts you might use to review a codebase.
[blog] A2A Server with an MCP backend in Cloud Run. Expose an API using MCP, and have agents that can talk to other agents using the A2A protocol.
[blog] To MCP or not to MCP? But maybe chill out on all the MCP stuff as there are cost and control implications.
Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:
Richard Seroter's Architecture Musings 
Leave a comment