Anytime a new version of something comes out in the tech industry—LLMs, mobile phones, services—people start asking “what’s next” a moment later. I’m quite excited by all the stuff in our Google queue, but I’m more interested in what’s already shipped. Let’s use what we got before obsessing over things we don’t have.
[blog] How AI Is Changing Open Source. More (random) projects, overwhelmed maintainers, and declining motivation to go to the effort of open sourcing? Sounds dire. Still reason for optimism.
[blog] Rich UI in Gemini Enterprise: A Field Report on A2A Agents and A2UI Primitives. Is the future of software distributed agents that use dynamic UIs for user interaction? Possibly. Pay attention to generative UI developments.
[article] MCP’s biggest update removes the machinery many servers were built around. A remote MCP server can now act more like a stateless HTTP service. This will require some changes, but should simplify things for many. More here, and here.
[blog] Don’t Let Your Agent Mess Up Your Database. Speaking of MCP, this is a great solution to a missing feature of MCP.
[blog] 3 things top 1% teams do differently. Your mileage may vary, but this post says that improving hiring, embracing parallelization, and adapting your code review flow are keys to success.
[blog] Detect early and enforce firmly with Google Cloud’s enhanced cost controls for AI spend. We’re doing some good work to try and get you ahead of any runaway costs.
[blog] How I set up containerized dev environments you can spin up in seconds for Antigravity CLI, Neat project. Isolate an agentic CLI into its own sandbox so that you can give it a limited set of rights before letting it go wild.
[article] Friday Forward – Be Bored. Big fan of forced boredom. Especially for adults. Sit still. Just be. Give your brain a chance to make novel connections.
[blog] Gemini API Managed Agents: 3.6 Flash, hooks, and more. Sweet update. Define your custom agent (prompt, tools, whatever) and have it run on a schedule in the cloud.
[article] Our Favorite Management Tips on Setting Strategy When the Path Is Unclear. I liked these. Few of us work in boring, stable places. Are you effectively navigating the constant change?
[blog] Merge Conflicts Are a Process Problem, Not a Git Problem. A merge conflict happens when Git doesn’t know which change should win. That’s a team problem to fix.
[article] Why I write skills instead of agents for knowledge work. Don’t customize an agent or LLM when a skill will do. Karl outlines his current thinking.
[article] Enterprises contend with mounting AI costs as tools sprawl. Not surprising. We’re in the experimentation stage where you’ve got to try a mix of tools before you consolidate. This post links to a series of articles on the topic.
[blog] Vibe code your first Android app with Antigravity 2.0 & Android Studio. Wow, Alicia made this workflow actually feel approachable. I may try it myself.
[blog] Google Cloud Skills Tutorial — Part 4 : Google Cloud Solution Architecture Skill. This scenario is exactly why agent skills are useful. Instead of relying solely on what an LLM knows (or what a search agent can scrape), feed your harness a skill that steers it to the most relevant cloud architecture.
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