I had four spontaneous meetings today, all on different topics. It’s important to me to have flex in my calendar so I can tackle things that pop up on a given day. But, it does contribute to a more chaotic schedule. Do you keep slack in the day, or schedule people out on days where you’re officially “open”?
[blog] Beyond one-on-one: Authoring, simulating, and testing dynamic human-AI group conversations. This is bonkers and super cool. Google did research into group conversations and open sourced a framework and tool to create and run these workflows.
[blog] Ship types, not docs. Good message. This is one reason our Google Cloud reference API docs are auto-generated when upstream protobuf definitions change. Need to keep this all in sync!
[blog] An AI Agent Published a Hit Piece on Me. Yeah, this is a pretty crazy story. An agent gets a PR refused, and it writes a takedown post about the maintainer!
[blog] Five key recommendations for platform teams in 2026. Except for #5, these have been oft-repeated recommendations for years now.
[blog] Don’t Make Them Wait: Improving AI UX with Streaming Thoughts. We’re all impatient, even when waiting for a magic LLM to deliver magic answers. This is a good post about maintaining engagement by streaming thoughts back to the user.
[blog] 7 Technical Takeaways from Using Gemini to Generate Code Samples at Scale. A year ago, we started on an effort to responsibly use AI to help build and maintain our code sample portfolio. Here’s a check-in.
[blog] Agents Can Either Be Useful or Secure. It’s time to rethink authorization approaches and how you secure your systems.
[artice] Your Strategy Needs a Visual Metaphor. Break up your dense, complex, and boring strategies with a compelling visual. AI tools are genuinely helpful here.
[blog] I Built an Agent Skill for Google’s ADK — Here’s Why Your Coding Agent Needs One Too. I’d like to see frameworks come with Skills or whatever can help agentic tools use those frameworks more effectively.
[article] The death of reactive IT: How predictive engineering will redefine cloud performance in 10 years. Our platforms will absolutely become more autonomous and proactive in handling operational tasks.
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