Author: Richard Seroter

There’s some good content today on strategic planning, managing our workloads, and what we should be learning next.

[article] AI is the future, but most companies’ plans are short-term. Nothing shocking or brand new here, though still a useful reminder to plan for more than short-term wins.

[article] What Agentic Workflows Mean to Microservices Developers. A smart question for everyone to ask is “how does my work with [X] now support AI and agents?” This post looks at your work building services and using API gateways.

[blog] MCP Client and Server with the Java MCP SDK and LangChain4j. Guillaume wrote a terrific post that shows you exactly how to build an Model Context Protocol server and client that provides tools that support an LLM.

[blog] Software Architecture for Developers. Every developer is going to need become (partially) an architect. I don’t see another path now that LLMs are such competent coders. So what does that look like? This post sheds light on what you’d need to consider.

[article] Meta releases Llama 4, a new crop of flagship AI models. Two of these new models were released last Saturday and looked impressive. There are now questions about the benchmarks and overall vibes of the models, but a strong achievement nonetheless.

[blog] Google announces Sec-Gemini v1, a new experimental cybersecurity model. This looks very cool. For it to remain helpful for digesting the latest threats, I’d imagine that we’d need to ship new versions often.

[article] When You’re Overloaded—and Delegating Isn’t an Option. Good advice here. Are you swamped but can’t delegate the work? Here are a few viable paths.

[blog] Is the .NET Ecosystem in Crisis? Probably not crisis, although there are changes to popular OSS projects (now going commercial) that has .NET devs uneasy.

[article] Claude Code and the Art of Test-Driven Development. I liked the walkthrough here of using an AI assistant to help you do TDD.

[article] In their own words: Microsoft leaders, past and present, on the company’s legacy and impact. Fifty years old! Microsoft stands as an massive success story, and this post gets perspective from those who participated along the way.

[article] Where and how developers learn in 2025? Books, forums, newsletters, and LLMs are a good tools to leverage this year.

[blog] Evaluating progress of LLMs on scientific problem-solving. Our research team shared a new benchmark to measure the potential of LLMs in helping scientists.

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

It was an awesome day in Vegas as pieces of Google Cloud Next ’25 took shape. The event starts on Wednesday, and I hope you’ll tune into parts of it. Today’s reading list is relatively light on AI, but that’s just to ease you into a week that should be VERY heavy on AI content.

[blog] Making Sense of the AI Developer Tools Ecosystem. The items named here change often, but the categories are sound.

[blog] Managing Applications across Fleets of Kubernetes Clusters. You’ve got a lot of good choices here. There are open products, vendor-specific ones, and a mix of both. Either way, consider something for keeping a handle on your set of Kubernetes clusters.

[article] Research: Hiring a Superstar Can Backfire. My hiring manager didn’t have to worry here, but if you are actually looking to hire an industry superstar, read this first.

[blog] The good, the bad, and the ugly of managing Sphinx projects with Bazel. Use a software build tool (Bazel) to transform tech docs to HTML? Here’s a candid look at what you should know.

[article] Enterprises have high ROI hopes for agentic AI. Reasonable excitement. For many companies, agents have a more tangible value proposition than general purpose generative AI.

[article] The 13 software engineering laws. Conway’s Law, Brooks’ Law, Price’s Law, and ten others are explained in this post. Good read.

[blog] SQL Pipe Syntax Style Guide & Best Practices. You’ll find many examples here and get a sense for how useful this SQL alternative is.

[blog] Building AI Agents — When To Use Frameworks Vs. SDKs. Frameworks like LangChain aren’t the same as using SDKs from the likes of Vercel. Get an assessment of the differences in this post.

[article] Sometimes never compete on price. Great post, as usual, from Jason. This one encourages companies, especially startups, to not default to “low price” as the strategy.

[blog] Mechanical Orchard ignites major shift in enterprise IT transformation with Imogen. I’m very bullish on what the MO folks have cooking. It’s impressive tech, and may help companies FINALLY get traction on their mainframe modernizations.

[blog] Accelerate Mainframe Modernization with gen AI from Google Cloud and its partners. Speaking of mainframes, we’ve also got some great tech, with a shoutout to MO in this post.

[article] The 8 best Go web frameworks for 2025: Updated list. Upleveling your web game with Go? Or looking to make a change? This is a well-done post that explains the best options.

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

It’s time for a mini-vacation before next week’s festivities. See you back here on Monday!

[blog] No elephants: Breakthroughs in image generation. All of us have access now to tools that help us do creative imagery. Where does that lead us?

[article] Gartner forecasts gen AI spending to hit $644B in 2025: What it means for enterprise IT leaders. It’s interesting to see where the spending comes from. Most in hardware?

[article] What barriers prevent engineers from achieving “flow”? If you’re a tech leader or architect, you should be obsessed with ways to keep your developers in a flow state.

[article] AI training lags despite increased use at work, survey says. Lots of AI getting deployed, but not enough training of staff. Let’s fix that!

[blog] In defense of ruthless managers. Competent, frank-talking, and outcome-oriented managers are valuable, even if misunderstood.

[blog] The Strategy Behind MCP. Thought-provoking! Fintan looks at competing models of LLM as a layer, or LLMs at the core. And which approach Anthropic is betting on.

[blog] New in NotebookLM: Discover sources from around the web. Such a great idea! Now tap into Google Search smarts to find additional sources of insight for your research projects in NotebookLM.

[blog] Meta Open Source: 2024 by the numbers. There’s a lot going on at Meta, and they still seem to prioritize open source investments.

[blog] The Augmented Architect: Real-Time Enterprise Architecture In The Age Of AI. Here’s an extended writeup from folks at Forrester Research. Good read for architects!

[blog] GKE at 65,000 nodes: Evaluating performance for simulated mixed AI workloads. That’s a lot of nodes. Nobody can build a bigger cluster than us, and here’s some great details on the tooling, test cases, and metrics used for a benchmark.

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

Today’s my last workday of the week (officially) before I take a couple of days off to recharge before Google Cloud Next. I’ll likely still have a reading list tomorrow, and then back on Monday!

[blog] Taking a responsible path to AGI. Very important topic, and the Google DeepMind team has a new paper with fresh perspectives on getting to artificial general intelligence.

[blog] How we built the new family of Gemini Robotics models. Relatedly, we now have robots that can make salads and dunk basketballs. Truly, we’re in the golden age.

[blog] Introducing Multi-Cluster Orchestrator: Scale your Kubernetes workloads across regions. This week is KubeCon, so expect a flood of Kubernetes-related news. For this one, having centralized orchestration makes it easier to manage fleets and place workloads.

[blog] Demystifying the Dreaded Garage Quote: How Gemini Became My Co-Pilot in Car Repair. I loved this. How often do we get documents—legal docs, healthcare forms, work quotes—and Alexis shows us how AI takes away our helplessness.

[article] How to Prepare for a Meeting Where Emotions Will Run High. Got a spicy meeting coming up? Here’s some help getting ready for it.

[blog] Why Go’s Tooling Feels Alien (but Powerful) to C# Developers. This won’t be the case in all situations, but it’s a reasonable look at the more IDE-centric C# toolset versus CLI-driven approach for Go devs.

[article] Developers feel the heat as AI coding tools shorten deadlines. Not surprising, as role expectations will absolutely change. But tech leaders also need to be well-versed in what’s feasible.

[article] Analysis: Google-Wiz deal sparks questions about cloud security strategy at AWS and Microsoft. This piece looks at how the hyperscalers are competing (or not) in the cloud security space.

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

Today’s reading list has lots of interesting advice. Whether it’s how to become an expert, build high performing apps, or motivate yourself through a career slump, there’s something for everyone.

[article] How To Become An Expert At Anything: 5 Scientific Steps To Mastery. Some of this might be intuitive to you, but I liked the reminders of what it takes to become an expert.

[article] Azul Report Finds 83% of CIOs are Spending More on their Cloud Infrastructure and Applications than Anticipated, While Only 2% are Under Budget. There’s no wholesale repatriation going on, and folks are going headfirst into the cloud.

[site] Patterns and Techniques for Writing High-Performance Applications with Go. This site has some awesome content for devs who want to build better (Go) apps.

[blog] Accelerate analytics with AI-assisted data preparation in BigQuery, now GA. AI-assisted cloud is the way to go. And I’m not talking chatbots. I mean deeply embedded capabilities directly into services. Here’s a great example.

[article] JFrog Survey Surfaces Limited DevSecOps Gains. Oof, our industry still seems pretty terrible at application security. What makes this better?

[article] Google’s Gemini 2.5 Pro is the smartest model you’re not using – and 4 reasons it matters for enterprise AI. We caught up with Gemini 2.0. And we raced ahead with Gemini 2.5. Give it a whirl at no cost.

[article] How Airbnb Used LLMs to Accelerate Test Migration. Interesting look at how Airbnb used LLMs to update their code base test files.

[blog] The evolution of graph learning. Good history lesson here, with lots of links to research that’s gotten us to where we are today.

[article] In a Slump at Work? Here’s How to Motivate Yourself. Nobody is constantly “on” during a long career. Even for long stretches. We hit slumps and downturns. I liked this advice to get out of it.

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

It’s a short week with my trip to Las Vegas on Thursday ahead of Google Cloud Next. I don’t like the widget in Google Calendar that shows me that 18.6 of my Mon-Wed hours are spent in meetings.

[article] MCP: What It Is and Why It Matters. You’ll have no questions left after reading Addy’s post about this protocol and system that helps AI agents access external resources in a consistent way.

[blog] Meet MCP: Your LLM’s Super-Helpful Assistant! More on MCP, with an actual demo of how you’d build the components yourself.

[report] 2025 Emerging Technology Trends. I don’t know of another company that spends more on technology than JPMC. Here’s their annual tech report that’s worth a look.

[blog] Gemini 2.5 Pro vs. Claude 3.7 Sonnet: Coding Comparison. it’s borderline blasphemy to say anything but Claude—an outstanding model—is the best at coding. But Gemini is coming in strong.

[article] Top 10 App Developer Skills and Trends to Learn in 2025. Solid list. Would your learning goals look different today compared to what you planned on January 1st?

[article] Is this strategy any good? How do you evaluate a strategy? When should you end a strategy? Will explores these questions.

[article] Google Just Showed How To Sell Agentic AI — Will Rivals Pay Attention? Honestly, I rarely hear people give credit to Google for customer-oriented selling (versus hyper-focusing on the engineering). But we’re getting it.

[article] 5 Levels of Agentic AI Systems. Speaking of agents, here’s a short, visual look at different agent patterns.

[article] Temporal lands $146 million at a flat valuation, eyes agentic AI expansion. I know folks who’ve recently joined, and it’s not surprising to see a fresh AI-centric view from them.

[article] How I Spent 17,784 Hours in 5 Years as a Startup Founder. Don’t copy+paste someone else’s journey. Your’s will differ. But it’s great to observe what others have learned.

[article] How engineering managers can use GenAI. This list applies to most technical leaders. Figure out where you can start making AI do some of the work!

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

I went and saw a comedy show last night, and doing a baseball game tomorrow night. So, I’m getting some of my favorite types of entertainment in batches. I hope you’re doing the same!

[article] The role of developer skills in agentic coding. Truly excellent post from Birgitta. She’s been heavily using AI coding agents, and sees a clear role for the developer. Keep learning to code.

[blog] Learn to code, ignore AI, then use AI to code even better. Very much agree. Don’t buy into the AI-only abstraction.

[article] Tracing the thoughts of a large language model. How to models like Anthropic’s Claude “think”? Here’s a fantastic piece from Anthropic that studies how LLMs plan and respond.

[blog] Taming the Complexity of Cloud-Native Development with Internal Developer Platforms. The industry is agreeing on the pieces of an IDP. Delivering one that feels integrated is no small feat.

[docs] AlloyDB AI use cases. Looking for inspiration for how to use your relational database with AI systems? This new landing page in our docs calls out some scenarios and tutorials.

[blog] What is an AI Agent? Strong piece and a good all-up look at agent definitions, technologies, and patterns.

[blog] An inside look into Google’s AI innovations: AI Luminaries at Cloud Next. We’ve got a special program at Cloud Next for leaders looking to get the big picture on AI.

[blog] Gemini 2.0 vs Gemini 2.5. It’s important to build a muscle (and system) to eval prompts, models, and agents. Here, we see that Gemini 2.5 is darn good at code generation.

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

Great day. Learned a few things, had some productive team meetings, and caught up after a half-week of traveling. Good items below!

[docs] Function Calling with the Gemini API. Great new documentation for a very important feature if you care about building AI agents, real-time AI systems, or action-driven AI apps.

[article] What’s keeping technology leaders up at night? This looks like an accurate list of things that tech leaders should be thinking about.

[article] Product, Design and AI. What’s the mix of skills you need on an empowered product team? This is a great list of skills to develop, acquire, and value.

[blog] Introducing the new Vertex AI Model Garden CLI and SDK. This is low-key awesome. Three lines of code, or a single CLI call to deploy virtually any open LLM.

[article] Escaping POC Purgatory: Evaluation-Driven Development for AI Systems. Stuck with cool prototypes with no path to making them “real”? This post lays out “evaluation driven development” where you test and assess your AI app at each stage.

[blog] 7 tips for effective system prompting: A developer’s guide to building better AI applications. Good, simple tips for getting more out of your prompts. I need to apply a few of these lessons.

[blog] Colossus under the hood: How we deliver SSD performance at HDD prices. Some of the most critical and well-engineered systems in the world are at Google. Here’s a look at how we’ve built parts of our storage subsystem.

[blog] Getting started with Rust on Google Cloud. We like Rust within Google, and Karl has a great post that shows you how to get an app built with Rust up and running in the cloud.

[article] Top 11 technical writing skills for product marketers. I’m fortunate to watch technical writers in action, and it’s a developed skill. This is great for anyone writing to technical audiences.

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

Home after my last trip before Google Cloud Next. Which isn’t much time. But, plenty is going on around our industry, and it’s important to observe even while our own work demands our attention.

[blog] Another observability 3.0 appears on the horizon. Sheesh, I thought we just “got” 2.0? Charity reframes the differences between “classIc” observability and the new style.

[article] How vibe coding will affect Engineering Managers. Very accurate looking list. What’s going to happen when engineering is no longer the bottleneck?

[blog] Putting Gemini 2.5 Pro through its paces. Forget all these snazzy AI benchmarks; can your model draw a pelican on a bicycle? If not, just get outta here.

[blog] Mastering Complex Workloads with Kubernetes JobSet and GKE metrics. I mentioned this new Kubernetes API a few days ago, and here’s an example of how to use it.

[paper] A-MEM: Agentic Memory for LLM Agents. Here’s a research paper on issues with current approaches to storage (or “memory”) for AI agents, and a proposed solution.

[article] All pretty models are wrong, but some ugly models are useful. good points from Jason. Don’t over-invest in perfectly structured frameworks or models that claim to describe human behavior or the world around us. But, they can be useful input to other ideas.

[blog] Leveraging GCP Model Armor for Robust LLM and Agentic AI Security. I didn’t know much about this cloud service, now I do. Whoever you get this capability from, just get it from someone.

[article] How to Build Career Resilience in Uncertain Times. Strong advice. I still try to operate every day as if I’ll get fired. That motivates me to keep my skills sharp, network healthy, and finances in decent shape.

[paper] Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations. NIST shares this new paper on AI-driven attacks and how you might mitigate the threats.

[article] Cloud veterans launch ConfigHub to fix ‘configuration hell’. Friend of the blog (and former colleague) Brian Grant built a company with smart co-founders and is now shipping something that may solve a problem for you.

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

Whew. Wild day with team offsites, Cloud Next keynote rehearsals, and some important Gemini model launches. The reading list below has its share of AI, but many other topics too!

[blog] Gemini 2.5: Our most intelligent AI model. Big release! Baked in “thinking” and some outstanding performance with coding.

[blog] Here’s how I use LLMs to help me write code. Just a terrific piece from Simon that has pragmatic wisdom, and a great example of a real-world situation where LLM helped him leverage his own expertise.

[article] Valkey 8.1’s Performance Gains Disrupt In-Memory Databases. This Redis fork isn’t going away. Rather, it’s becoming a legit top-tier option for those looking for a blazing fast cache.

[blog] How to Choose the Right LLM. The JetBrains IDEs let you choose which LLM to use to help you code. This piece outlines your options.

[article] When You Oversell Your Abilities and It Backfires. We all do it at one time or another. Some of you do it hardcore on LinkedIn. How do you regain credibility once you’re found out?

[article] 20 Business terms every Engineering Manager should know. Really, really useful list. I wouldn’t have recognized most of these until halfway through my career.

[blog] Considering Data Science Users in Your API Design. I hadn’t thought about this. Yes, now that data science teams (and agent builders) will be using your APIs, what changes should you consider?

[blog] Build gen AI agents using Google Cloud databases. If you’re going to build agents, you’re going to need a strategy for interacting with data services.

[blog] Building Vertex AI RAG Engine with Gemini 2 Flash LLM. Speaking of data services, you might want an easier option for processing and caching your data. Our RAG Engine is one such option.

[blog] Interoperability Is Key to Unlocking Agentic AI’s Future. This seems right, and I liked the nine dimensions of agentic interoperability called out here.

[article] Future Self. I’ll often joke about Past Richard screwing Future Richard with bad travel choices. Find the balance between the cost you take today to help tomorrow, and not over-doing it.

[blog] Figma to Flutter: Convert designs to clean Flutter code. As much as you can use AI to build interactive, working prototypes, there’s still obviously use cases for systems like Figma. Here’s a new tool for converting those designs to Flutter widgets.

[article] Getting Started with TDD: A Practical Guide to Beginning a Lasting Practice. This is the rare piece that could have actually used the term “comprehensive guide” and been accurate.

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