Author: Richard Seroter

Just got to my hotel in Seoul after a 13-hour flight with no wifi. But, did reading beforehand, so publishing yesterday’s list now. Some great “thinking” content here for you to review.

[blog] Forward Deployed Engineers. Very good. I’ve seen this term and had a vague understanding of it, but I now I understand how it’s used. And why the idea is so valuable.

[blog] Announcing MCP Toolbox support for Firestore. I’m glad to see this. I’ve got a demo agent app that uses a relational DB today, but I’d rather swap for Firestore for cost and simplicity sake.

[blog] How to Target Non-Developers with Your Developer Product. Non-devs buy products for devs. Few developers have their own budgets! Here’s some good advice on how to better message to the buyers.

[blog] A Look at Context Engineering in Gemini CLI. Valuable deep dive into the various ways you can impact the working memory of your agentic CLI.

[article] Old IT systems weigh down bank modernization. If you’re counting on customer loyalty while figuring out a better tech strategy, you might be in for a rough season. Sobering data here.

[blog] MongoDB AMP. Big swing from MongoDB! Are folks going to look to them as the vendor that helps them do full application stack modernization? Databases yes, I’m not sure about the rest. More here.

[article] Vibe coding has turned senior devs into ‘AI babysitters,’ but they say it’s worth it. I’m starting to wonder if vibe coding is going to bring lots of new developers to the industry, but established ones won’t see enough value in embracing this way of working.

[blog] How Google Cloud’s AI tech stack powers today’s startups. Good stories. It’s clear that startups prefer this cloud for its tech, programs, and people. Come aboard.

[article] Thinking Like an Architect. I’m down with whatever Gregor is thinking. “Architect” isn’t a promotion from developer; it’s a different job with unique skills required.

Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:

Good day. Tomorrow I’m flying to Seoul for a week of customer meetings. Assuming I have airplane wifi, expect a reading list tomorrow. And one every day next week, albeit at weird times 🙂

[article] Prompt Engineering Is Requirements Engineering. I liked this perspective, especially the reminder that sharing prompts with others can be dangerous. Make sure you gather your own requirements!

[blog] Getting it Done Even When You Don’t Feel Your Best. Great piece. Most of us aren’t born with this skill, but we can all develop it.

[blog] Create Custom MCP Servers in Minutes with MCP Toolbox. This is cool. Use this open source product to quickly create your own MCP servers out of most any database type.

[blog] How to secure your remote MCP server on Google Cloud. Here’s a very good reference architecture view of a secure MCP server. Swap out the services, run on a different cloud, whatever. But these core capabilities seem important.

[article] Atlassian acquires DX, a developer productivity platform, for $1B. DX authors many of the posts I share here in a given month. Good for them!

[blog] New DNS Armor can help detect, mitigate domain name system risks. Protect your internet-bound queries with this interesting new service.

[article] Designing AI-ready architectures in compliance-heavy environments. There were more details here than I expected. AI isn’t just for tech companies and quirky startups. You can take advantage of it in every type of environment.

[blog] We’re investing in connectivity, products and skills for Africa’s AI future. Love seeing this. Digital access can transform lives and entire societies.

[article] The Emotional Strength You Need to Lead Through Change. This offers a good list of traits we need to exhibit (and internalize) if we’re going to lead during changing times.

[youtube-video] The Agent Factory – Episode 7: Gemini CLI with Taylor Mullen. Fun episode where you can get fresh insights into agentic CLIs and how the open source Gemini CLI came to be.

[blog] Let’s kill vibe coding and bring back prompt engineering. I get the point. Prompt engineering isn’t the target, but context engineering might be closer to the right destination.

[blog] Announcing the new Practical Guide to Data Science on Google Cloud. I’ve read this ebook. It’s good. There’s a nice mix of detailed info and strategic advice.

[blog] Read Google.org’s first ever Impact Report. Nice content to flip through. We get encouraged to donate time and (matched) charitable contributions.

Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:

I had lunch yesterday with our Pixel folks working in San Diego. Many of them are using the Gemini CLI to get their work done. Super cool to see how folks across Google are using the same AI tooling that everyone else is.

[blog] How to generate (actually good) designs with AI. My AI-generated frontend work is decidedly “Richard quality.” Functional, not great. Some good advice here for doing better AI-assisted designs.

[blog] Gemini achieves gold-medal level at the International Collegiate Programming Contest World Finals. Impressive performance! This includes solving a problem that none of the university (human) teams could.

[blog] Behind The Scenes: Examining & Estimating Token Consumption For Large Gemini Runs Especially Thinking Tokens. Some deep research into how thinking tokens get consumed. It’s apparently not super easy to predict. If you’re paying per token (versus self-hosting models), it’s something to factor in.

[article] Why CIOs can’t afford another modernization failure in 2026. Nothing shocking here, and I’m far from convinced that we’ll get better at this as an industry. But some will!

[blog] Code Generation with Gemini CLI, MCP Toolbox, and Firestore. IDE tutorials can be reworked to be entirely CLI based, as William shows here.

[blog] Connect Your AI to Everything: Spring AI’s MCP Boot Starters. I kinda love this. Get the full power of Java and Spring but also easy setup and consumption of MCP servers.

[guide] RAG infrastructure for generative AI using Google Agentspace and Vertex AI. Any low-code (or no-code) agent building environment depends on pre-existing data and tools. Here’s one look at a reference architecture to set that up.

[guide] Multi-agent AI system in Google Cloud. Another terrific guide, this time for setting up a multi-agent system. Even if you’re not using our cloud, there’s applicability here.

[blog] Diving Into Spec-Driven Development With GitHub Spec Kit. I mentioned spec-driven development yesterday, and this is also a cool effort around it.

[article] Old metrics can’t map new terrain. The folks at Slalom shared a new framework that focuses on forward-looking metrics.

[blog] Sort Lines in Source Code. I can be lazy about this, but this is a good reminder that bugs often come from careless places.

[article] What Disruptive Innovators Do Differently. Good list of traits. Do you exhibit these, or have a company that rewards them?

[blog] Build your AI apps faster — Cloud Run now supports modern Python frameworks. Our buildpacks no longer default to Gunicorn as a web server, and respects whatever modern Python web framework entry point your app depends on.

Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:

Each year, we have an internal conference for the technical writers at Google. I got to keynote today’s session, and it was fun to reflect on what will stay the same, and what will change in the years ahead. These daily reading lists have influenced a lot of my thinking!

[blog] Powering AI commerce with the new Agent Payments Protocol (AP2). This changes things. Would people really trusted unattended agents to perform payment transactions? This new protocol establishes a foundation of trust, and is backed by the companies that matter here. News here too.

[blog] Announcing Tessl’s Products to Unlock the Power of Agents. We’ve seen the early wild-west days of building apps using AI tools, and now the grown-ups are here. This is a product to do more responsible spec-driven development. Sounds good to me.

[blog] Kubernetes 1.34 is available on GKE! This is a short, human-readable summary of what mattered in this latest Kubernetes release.

[article] Oracle Delivers Java 25 Edition of Venerable Programming Language for the AI Era. Java 25 is here, and this language continues to be one of the most relevant in our industry. More coverage here.

[blog] A framework for pricing AI products. It’s not hard to get this wrong, so I appreciate Stripe offering some useful advice on how to think about charging for your AI product.

[blog] The Anatomy of an AI Agent. Good foundational thinking from Allen. The agent is where the next leap comes from. The models alone were never going to change the game.

[article] 15 Best Practices for Building MCP Servers in Production. I don’t know if there are any “best” practices yet, but there are at least good practices that we’ve discovered so far.

[ebook] Building Intelligent Agents with Google ADK. All online and free to access. Great content here on building AI agents using the open source Agent Development Kit.

[blog] Test state, not interactions. Good reminder, especially if you’re leaning on AI to crank out a massive number of unit tests that give you false confidence.

[blog] How to Write Better Python with Ruff on Google Cloud. It seems that with so many folks focused on Python, it’s getting some better developer tooling.

[article] 2-agent architecture: Separating context from execution in AI systems. Here’s some knowledge acquired through experience. When do you introduce a second agent to the architecture? I liked the idea of a context-gathering agent and a real-time agent.

[blog] Learn Your Way: Reimagining textbooks with generative AI. This speaks to me, and I’m fairly convinced it represents the future of education overall. Take trusted, static content and personalize for individual learners.

[blog] Beyond the Horizon: How Angular is Embracing AI for Next-Gen Apps. Angular is doing its thing, and bringing a few new ideas to the table.

Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:

Back at it today after a three day weekend to Texas. Lots of AI content today!

[article] How do AI-assisted code fixes impact code review time? Don’t expect a “drop-in efficiency boost” as noted here. But Meta’s research did highlight a few good pieces of advice.

[blog] Writing effective tools for agents — with agents. Great advice in this Anthropic post. Read it to learn more about tools and practices for building effective ones.

[blog] Gemini CLI Tutorial Series — Part 11: Gemini CLI Extensions. The extensibility of the Gemini CLI will spark another monster wave of creative use cases for agentic CLIs.

[article] Agents show promise, but widespread usage in the enterprise remains elusive. Two things can be true. It’s true that there’s excess hype and a dearth of wild success stories with AI. It can also be true that our industry hasn’t skilled up yet and the tech is improving rapidly.

[blog] How AI made Meet’s language translation possible. Case in point. Real-time language translation will be ubiquitous in a short time.

[blog] GPT‑5-Codex and upgrades to Codex. Looks like a good upgrade to OpenAI’s coding models and tools. I can’t track all the names, though. More here.

[blog] Let AI Interact with Your App via MCP. How long until agentic users outnumber human users for your app? If you’re a SaaS provider, it’s going to happen more quickly than you think.

[blog] Listen to a discussion on how AI can power scientific breakthroughs. The commercial work we do matters because it makes it possible to fund such amazing work for the benefit of humanity.

[docs] Serve open LLMs on GKE with a pre-configured architecture. I think most of us want smart infrastructure default values that we can trust and accept. Here’s an example of what you can do when you have that.

[article] At 20, Techmeme has never been hotter. I’ve been a reader of Techmeme for years, as it’s the best source I know of to get a “what’s going on in tech today” view of our industry.

[blog] OpenTelemetry Protocol comes to Google Cloud Observability. Looks like we’re picking OpenTelemetry as the winner, and it’s getting baked in for telemetry across Google Cloud.

[article] AI developer certifications tech companies want. I can’t imagine anyone getting hired because they’re certified, but it certainly signals a level of effort and interest in a topic.

Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:

Tomorrow’s a day off as I go visit my kiddo at college and see how he’s doing. I’ll be back on Monday, most likely not wearing cowboy boots and a ten-gallon hat. But don’t rule anything out.

[article] RSS co-creator launches new protocol for AI data licensing. It probably makes sense to have something along these lines to create some predictability for AI companies and data providers.

[blog] The Best Way to Keep a Secret in the Cloud Is Not to Have a Secret at All: The Birth of “Secretless” Cloud Workloads. Given the growing number of non-human identities, I’d imagine that we need new ideas around security. Here’s one.

[blog] Generating videos in Java with Veo 3. Yes, Java devs get to have all the fun too. Guillaume follows up on his post about generating images with this banger about generating videos.

[blog] MCP is replacing the browser: Here’s how devs should prepare. This is one of the best things I’ve read on this angle. Yes, apps, UIs, security, and many things are about to shift as agents + MCP replace the need to browse everywhere to get something done.

[blog] Building a Custom MCP Server for Gemini CLI: A Hands-on guide. Start making this happen for real by building up some MCP servers and seeing how your behaviors change.

[article] How CIOs can steer legacy tech overhauls. Keeping a focus on overall business goals and current pain points is a solid recipe for doing work that matters.

[article] Building AI-Resistant Technical Debt. Just because we can generate code faster, doesn’t mean we can (or should!) skip design checks, quality reviews, or any of that. It takes time to build instincts about good software, and we need to keep developing that skill.

[article] AI cuts onboarding time in half for new hires in the enterprise. This doesn’t surprise me in the least. Customized, personalized learning is one of the “killer apps” for AI.

[blog] A2A Extensions: Empowering Custom Agent Functionality. Be careful about adding too much complexity to your architecture, but I can see real use cases where these extensions make sense.

[blog] How our Edge Kubernetes Platform has Evolved. The Chick-fil-A folks were among the first to publicly talk about an at-scale Kubernetes-at-the-edge strategy. These updates make sense.

[blog] Words are not violence. Great post from DHH. Debate, talk to people. Those who resort to violence won’t win. We can all be better than that.

Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:

Ugh. The violence against innocent people here in the US this week has me shook up. I’d like to think we’re better than this. It’s inexcusable, and I do pray we collectively rediscover our ability to live life while remaining civil to those around us.

[blog] Automate app deployment and security analysis with new Gemini CLI extensions. This is awesome. Now it’s even easier to deploy apps built or changed with the Gemini CLI. And, to scan projects for security issues. I’m excited about this new extensions subsystem.

[article] Perl programming language rises again – Tiobe. Really? When you look at the data, it’s possible this is propped up by a burst in training or books, not necessarily new engineers. But either way, Perl lives.

[blog] Scaling high-performance inference cost-effectively. Lots of compelling data here for anyone looking to do (better) AI inference on their Kubernetes cluster.

[blog] How to harness skills that AI can’t automate. Nice post. I don’t really consider these “soft skills” either. They’re ones that you develop and hone.

[blog] Top Google Cloud Pub/Sub Tips & Tricks. Messaging middleware doesn’t get any glory, but it’s as important as most things in your architecture. Good advice here.

[blog] Cross-Cloud Streaming Pipeline: AWS Kinesis to GCP Pub/Sub. Two messaging posts in one day? Sure. I like that our bus natively connects to engines in AWS and Azure.

[article] 9 vital concepts of modern JavaScript. JavaScript is thirty years old, and it might be time for you (and me) to refresh the way we use it.

[article] Building Bridges: Deploy agents with A2A on Vertex AI Agent Engine. I read this to learn more about agent to agent systems and what a managed platform can do to make it easier to deploy/run them.

[article] Hugging Face Introduces AI Sheets, a No-Code Tool for Dataset Transformation. Is this the next area we’ll see a burst of options? Creating low-code/no-code apps that front generative AI models? We’ve got that Opal experiment out there too.

[blog] Agent Development with Gemini CLI. This post shows how you might take an existing codebase (even a “modern” one like an agent app) and have the agentic CLI help you understand and extend it.

[blog] Spring AI 1.1.0-M1 Available Now. This Java framework for AI app builders is shaping up nicely. Put it in your list of ones to evaluate if you’re building Java apps.

[blog] Investigate fast with AI: Gemini Cloud Assist for Dataproc & Serverless for Apache Spark. Troubleshooting distributed apps isn’t easy. Built-in AI assistance makes investigations into systems like Spark a bit easier.

Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:

Big reading list today, with some point-counterpoint pieces that look at different perspectives. Enjoy!

[article] Survey: Most IT Teams Not Prepared to Manage AI Workloads. Honestly, have most IT teams been ready for any of the tech advances of the past couple decades? Gotta learn on the fly!

[blog] AI Mode is now available in five new languages around the world. Free, easy to use, and actually helpful. Give it a whirl.

[blog] A new experimental Go API for JSON. I not only like the Go language, I also like how this team (and community) approaches updates to the language.

[blog] IP Address Planning for Large-Scale GKE Deployments. This is the least flashy topic I could have possibly read about today. But also kinda interesting and relevant.

[article] How to Keep Your Team’s Spirits Up in Anxious Times. Whether folks are anxious or not—and it seems like there have been anxious folks for a long time now—this is good advice for stable, inspiring leadership.

[blog] Calling Nano Banana from Java. Excellent post that shows off image generation, image alteration, and image combination with our fantastic new AI model.

[article] A “Beam Versus Dataflow” Conversation. Run data processing workflows on self-managed infrastructure, or with a managed service? Age old debate, with some thoughts here.

[blog] Now available: Rust SDK for Google Cloud. Rust is growing in popularity, and I’m glad we now have GA support for an official SDK that covers most of our cloud APIs.

[article] Conflicting opinions on the ROI of AI. Is everyone failing with AI projects, or succeeding? It can be true that many struggle for a variety of reasons. And also true that many have the right sponsorship, tech foundation, and customer-focus to deploy AI agents successfully!

[report] The ROI of AI 2025. Here’s the report that David refers to the previous piece. I’ve gone through the report. It’s got a ton of data from a very large survey population of IT leaders.

[article] 5 Times LLMs Help You Code… and 5 Times They Fail. These can be success and failure points. Or not. Depends on the user, the situation, the product, and the expectations!

[blog] Build with your Favorite Models from the Vertex AI Model Garden with LiteLLM. This is a handy proxy that can sit in front of multiple LLMs. Karl’s examples—direct calls, via AI coding tools, and as a persistent OS service—are very tangible.

[article] Developer reliance on AI worries IT leaders. On one hand, I see that execs want their teams using AI more aggressively. Then I see these results that show leaders are concerned about the level of AI use. WHAT TO BELIEVE.

[blog] Agent Middleware. Agents need context, and often, a refresh along the way. LangChain is adding a new middleware construct to help with that. Note that ADK has something similar.

Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:

There are a couple of good team/career based pieces in today’s list. If you work solo, fantastic. But if you’re like most of us who deliver through a team, don’t underrate the value of investing there!

[blog] How to Create a Google Workspace Digital Assistant Agent in Google Cloud Agentspace. Be a little selfish. It’s ok to build an app, tool. or agent just for yourself. And with these no-code builder environments, anyone can do this.

[article] The Secret to Building a High-Performing Team. Looks like we’re comfortable covering 92% of what we want to say, but often censor the last 8%. Why is that? What does it take to have a high-trust culture that permits us to air everything out?

[blog] 6 ways to use NotebookLM to master any subject. These are great updates for students, but also for any of us who are trying to learn and digest new topics.

[article] Kubernetes Primer: Dynamic Resource Allocation (DRA) for GPU Workloads. This is a important update to Kubernetes resource management. Good overview here.

[notebook] Get started with Vertex AI Memory Bank. I read through this new notebook today, and thought it was a good dive into creating agents with memory.

[blog] The Endless Hot Vibe Code Summer. Good recap and analysis of this hot trend that’s got people excited, upset, and confused.

[blog] AI Mode. Simon gave our AI Mode another shot, and came away pleasantly surprised. These results will never be perfect, but I like how we’ve complemented standard Search.

[blog] 996. You aren’t your job. Chasing a “satisfying career” over everything else rarely leads to long-term happiness and fulfillment. Do whatever you want, but don’t work yourself to death because someone else tells you that’s how you matter in life.

[article] 13 reasons SQL has got to go. Oof, this goes hard after SQL! Many of these points do make sense, but I’m not clear on whether any alternative is truly better.

[blog] Is that a deprecation? Or is it just removed? Words mean things. If you deprecate something, you’ve discouraged new users but kept it alive for existing ones. Removal takes it away from everyone!

[blog] In-browser semantic search with EmbeddingGemma. When the model is smaller than most web pages, you can afford to use it in a static app that runs locally in your browser.

Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:

I haven’t yet decided what I’ll dig into over the weekend during quiet times. There’s the usual assortment of books and such, but I usually pick one tech thing to explore. Do you keep a list of technologies you want to look into?

[article] Vibe coding is not the same as AI-Assisted engineering. Don’t conflate the two. AI-assisted engineering keeps the human engineer in control. Both have their place.

[blog] Using Gemini CLI to Keep Your User Docs Flawless. Custom slash commands that help you create and update your docs as your code changes. Great use case.

[article] The evolving role of DevProd teams in the AI era. You may have a DevProd team, or even just a person who helps create tools and standards for dev teams. Good look at how that evolves.

[blog] From Message to Job: A Serverless Event-Driven Data Pipeline on GCP. It’s not hard to over-engineer a simple use case. A seasoned technologist can recognize when they’re adding necessary complexity. JK has a useful post about an architecture that might seem over-done, but is actually fairly elegant.

[blog] Event-driven architecture performance testing. It’s not always easy to do functional or performance tests on an async system with many possible outcomes. Here’s some guidance from the Capital One team.

[blog] How to build with Nano Banana: Complete Developer Tutorial. Take a moment this weekend to play with this model. There’s no cost, and it might blow your mind.

[blog] Welcome EmbeddingGemma, Google’s new efficient embedding model. The Hugging Face team went DEEP on this new model.

[blog] How to Build Highly Available Multi-regional Services with Cloud Run. Cloud Run is regional, which is often plenty good enough for a workload. But when you need multi-region support, you’ve got some good ways to do it.

[article] How AI can mitigate legacy tech troubles. It’s a tool that people should use. Don’t rely on it for everything, or fail to audit the work. But skipping AI assistance when planning a migration is a huge mistake.

[blog] Agent Factory Recap: Keith Ballinger on AI, The Future of Development, and Vibe Coding. This was excellent. My boss Keith joined this podcast to offer real insights on AI-assisted engineering before vibe coding an app himself.

[blog] Composing Data from Multiple Services. There’s more than one way to aggregate data and present a unified result. Derek looks at some patterns.

[blog] From query to cart: Inside Target’s search bar overhaul with AlloyDB AI. The folks at Target share a useful level of detail about a hybrid search approach with a modern database.

Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below: