Category: Daily Reading List

Many folks were working differently with AI long before anyone documented. But now folks are capturing some of these changes to the dev workflow. I wrote up a post yesterday on it, and there are complementary points in today’s reading list.

[blog] AI-Driven Prototyping: v0, Bolt, and Lovable Compared. Excellent post from Addy about some fascinating tools for building apps.

[blog] Why asking for the ROI of AI in software development is the wrong question. It’s past time to question whether AI makes software development better. It does. This post looks at measuring the impact of integrating it into the workflow.

[blog] Prophecies of the Flood. Ethan talks about the conviction that AI researchers inside AI labs have about “the emergence of something unprecedented.”

[blog] My AI/LLM predictions for the next 1, 3 and 6 years, for Oxide and Friends. Another smart person looks at the future. Nobody has any real idea of what’s going to happen, but even the thought-exercise is useful.

[article] Open source licenses: Everything you need to know. Is this the most thrilling topic. No. But it does matter as you plan your long-term bets.

[article] What Effective CEOs Do After Their First 90 Days. Any leaders should have a plan when taking over a new team or organization. This offers a decent blueprint!

[blog] Understanding Private Cloud, Hybrid Infrastructure, Multi-Cloud, and Distributed Cloud: A Comprehensive Framework. Keith goes through each deployment model and how they interact.

[blog] How to build dynamic web experiences with Conversational Agents. I hadn’t thought of making web pages evolve in reaction to what’s happening in the chat agent. Wild.

[blog] How to be a Ghostbuster. Are nearly 10% of developers doing virtually no work in their job? This is on the offending dev, but also on the team leaders who aren’t close enough to the work.

[blog] 8 coding lessons from almost three decades of software development. These are great. Hopefully your team discusses the ideas and their merits.

[blog] Cursor vs Windsurf vs GitHub Copilot. You might not be downloading and using every AI tool right now. I’d have a lot of question is you WERE. Analysis like this helps you get the gist. See this related one on Windsurf vs Cursor from the same folks.

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

It was a good day of starting up my weekly (internal) newsletter for the year, catching up on doc reviews and email requests, and scoping out some coding projects for the weekend.

[blog] Continuous Delivery for GKE with Cloud Deploy and Argo CD. Great example of combining the control flow of a continuous delivery tool, with the scaled and structured process of GitOps.

[blog] Enterprise Architects: From Cartographers to Scouts. What does it mean for architects to switch their mindset and work from being cartographers (mapping out the space) to scouts (finding paths)?

[article] Generative AI in Government: What to Expect in 2025. Good piece that stays grounded on what matters most: better service for constituents.

[article] Are You Checked Out at Work? I like that this article isn’t about why you’re checked out. It’s about how to get yourself re-engaged.

[article] Tech Outage. I’m with Bob here. It’s easy to throw screens at our kids, especially in public settings. But we’re better off resisting the urge. I’m taking my kids out for dinner tonight, and they bring coloring books and action figures for us to play with.

[article] Rumors of SaaS’s death may be exaggerated. Are AI agents going to make database-driven SaaS (Salesforce, etc) unnecessary? Color me skeptical.

[blog] How to work with open source formats on BigQuery. A good data warehouse will play with a variety of data and table formats. I thought this was a good look at what the means in practice.

[blog] Design Patterns for Building Resilient Systems. Nothing Earth-shattering here, but that doesn’t mean this isn’t important info to refresh yourself with.

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

My day started at 2:30am when I got up to deliver a presentation to a company in Spain. I couldn’t fall back asleep after, so Google got some bonus hours of work out of me today. In today’s reading list, you’ll find some smart content on strategy and getting better results from AI models.

[article] Refining strategy with Wardley Mapping. I’ve tried to get into this, and never could. Probably not smart enough. But, some folks swear by this technique of understanding your strategic landscape.

[blog] A Good Life is Active not Passive. Important point by Brad here, and one that we should all take to heart as we resist becoming passive consumers of entertainment.

[blog] Supervised Fine Tuning for Gemini: A best practices guide. Get some good insights into the process for adding precision to existing models like Gemini.

[article] What’s Ahead for AI-Assisted Coding, Open Source and More. Everyone’s guessing, but observing lots of guesses can still help you feel more aware of the opportunities out there.

[blog] Monitor your Google Gemini apps with Datadog LLM Observability. As expected, software vendors are adding support for the LLMs that the market is embracing.

[blog] What makes strong engineers strong? Solid list of four traits. Many of these would apply to most professionals in technology.

[blog] The Last Strategy Framework You’ll Ever Need. There are plenty of frameworks for strategic planning (see the the first entry in my reading list today!). John explains a few of them, and explains why strategy is hard work.

[blog] Introducing Vertex AI RAG Engine: Scale your Vertex AI RAG pipeline with confidence. Getting more customized and contextual responses from an LLM is valuable. Retrieval augmented generation is a pattern, but involves some machinery. We’ve just made it much easier.

[blog] Evaluating RAG pipelines. Did you follow the previous entry here and create a RAG pipeline? Cool. How do you evaluate whether it’s the most effective combination and overall system? Mete helps you identify success measurements.

[blog] When to Consider Building a Private Cloud: A Pragmatic Perspective. Basically, they’re a fit for certain use cases, but shouldn’t a go-to for many companies.

[blog] Announcing Supporters of Chromium-based Browsers. I didn’t know the scale, effort, or cost of supporting a project like Chromium. Now I do, and am glad to see more folks pitching into this browser.

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

There are still plenty of “what’s coming in 2025” pieces coming out, and I’ll continue including the ones that catch my eye.

[article] Agents. You want eight thousand words about AI agents? Today’s your lucky day. Chip wrote a fantastic piece that goes into helpful depth on the topic.

[blog] Ultimate guide to CI/CD: Fundamentals to advanced implementation. This doesn’t break new ground, but it’s good to till the current ground. That’s the extent of my farming metaphors.

[article] 5 Mistakes Managers Make When Giving Negative Feedback. Few folks like giving corrective feedback, but it’s a necessary task. This article offers mistakes we make, and how to fix them.

[article] Explore vs Execute. Great essay from Jason here. What happens after you have the “fit” for a given product? What does an operational switch look like?

[article] Developer Productivity in 2025: More AI, but Mixed Results. A batch of experts provide input into what this year may hold for developer experience and productivity.

[blog] Mistakes engineers make in large established codebases. The main mistakes come from “lack of consistency” according to this interesting write up.

[article] How to Overcome Common LLMs Pitfalls and Build Smarter AI Systems. My colleague Mete does a good job outlining a few challenges and opportunities for those working with LLMs.

[youtube-video] The amazing, but unsettling future of technology… Entertaining video, as always from this source. But also a good overview of the hot tech (AI, agents, AR/VR, robotics, etc) and questions surrounding them.

[article] Engineers hit AI development roadblocks: IBM. Fifteen tools to build an AI app seems like … a lot. That statistic, and more in these findings.

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

Today’s reading list happens to have a lot of good material for those planning ahead. How do you avoid overcommitting? Or do a competitive analysis? Refine a strategy? Create OKRs?

[blog] How I built an agent with Pydantic AI and Google Gemini. Karl might have just saved you a few million dollars in consultant fees! He built a SWOT analysis tool using a variety of technologies (including the Gemini API).

[blog] Open Source Discussion Archetypes. You may not personally care about the OSS licensing fracas that’s going on, but somebody does. Steve looks at the groups that are intentionally or unintentionally invested.

[article] My 2025 trends predictions. Tom’s a tech writer at Google, and shares some useful thoughts on AI and how it’s going to impact his field.

[blog] How to effectively refine engineering strategy. Are you doing big strategic exercises, or doing small refinements until a strategy is “ready” for mass deployment? Will has a good post that can help you understand the problems being solved, and which tools to use.

[article] Meta is ditching third-party fact checkers. Good. However well intentioned, centralized moderators inevitably apply unhelpful bias. Community Notes is a useful approach.

[page] 2024 AI timeline. This is a cool way to view all the major LLM releases from last year. It shows both open and AI only (closed) models.

[article] Why You’re Chronically Overcommitted. This will probably resonate with most of you. I’ve fought back against this over the past few years to good effect.

[article] Google maps the future of AI agents: Five lessons for businesses. Are you clear on what agents are, and what you should be doing with them? I’m not sure I am, so you’re better off than me. But this VentureBeat story looks at a recent paper and offers direction.

[blog] Hitting OKRs vs Doing Your Job. OKRs (or however you’re capturing major goals) are for special focus, not the day to day work.

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

Happy New Year! I’m back with the first edition of the Reading List for 2025. I trust you took some time over this holiday break to recharge. Let’s have some fun together this year.

[article] Our Favorite Management Tips of 2024. Very good categories here, including leadership skills to develop, building executive presence, writing skills, and asking smarter questions.

[article] How AI-assisted coding will change software engineering: hard truths. Sheesh, this is awesome. My colleague Addy takes over the Pragmatic Engineer newsletter for this edition to give some straight talk on AI assisted dev tools.

[blog] Cognitive load is what matters. Excellent post that we should all take to heart. What are we doing to reduce the extraneous, unnecessary cognitive load on our peers or customers?

[docs] Choose Cloud Managed Service for Apache Kafka or Pub/Sub. I like docs like this which outline the benefits and tradeoffs between a set of managed services.

[blog] Go Developer Survey 2024 H2 Results. Check out the results from 4000+ Go developers. You’ll find good insights into how they’re using AI, challenges within teams, and more.

[article] Bing is doing its best Google Search impression anytime you look up the word Google. Yeesh. I tried it out, which probably gives Bing more traffic than it’s used to getting!

[article] What engineering managers need to know for 2025. Who knows what this year holds, but these represent some useful categories to consider.

[blog] The 12 Anti-factors of Infrastructure as Code. You might be doing some of these practices and thinking you’re following the best ideas. Brian pushes back on some mainstream practices.

[blog] The PyTorch developer’s guide to JAX fundamentals. If you love PyTorch, use PyTorch. If you’re intrigued by the performance and ecosystem of JAX, this post will help you map some of the familiar concepts.

[blog] Open ML News – Holidays Edition. My daily summaries are often fed by other people’s aggregated summaries. Omar recaps some of the most interesting open model news.

[article] From AI agents to enterprise budgets, 20 VCs share their predictions on enterprise tech in 2025. Where do VCs want to invest in 2025, what’s hot besides AI, and what’s going to happen with enterprise IT budgets? Good questions and answers here.

[article] 25 experts predict how AI will change business and life in 2025. Is the time horizon off? Or the predictions too unrealistic? I don’t know, but it’s useful to see where smart folks see things going.

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

That’s a wrap on 2024 for me. I cleared out the reading queue, which should hold you over for a few days. I’ll keep reading things and occasionally posting on X/Twitter, but I’m generally taking off the next two weeks. See you back on the Reading List on January 6th, and THANK YOU for reading.

[blog] Google Research 2024: Breakthroughs for impact at every scale. What a year. Look at the advances in SO MANY areas of life.

[blog] Platform Engineering Evolution: Trends and Future Impact on Software Delivery. Excellent content here from Paula. She does a deft job exploring the history of platform engineering and where it’s heading.

[blog] Spanner in 2024: A year of innovation. Spanner might be the best database in the cloud, and with its multi-model nature, you can now do so much—full text search! vector search! graph!—at once.

[blog] What just happened. Many of us are asking the same question. December was a breakthrough month in AI with the entire landscape shifting. We shipped an absolute ton.

[blog] December in LLMs has been a lot. It has. Across a wide range of dimensions, AI looks drastically different in the market than just a few months ago.

[article] Workers fear their skills will be obsolete this decade, report finds. Some will be. Just like some skills I had ten years ago that don’t matter today!

[blog] Migrate your applications to Kubernetes using Istio. Sometimes your migration assistance comes from places you didn’t look into.

[blog] AlloyDB Omni and local models on GKE. Great walkthrough of co-locating a model and a database in the same Kubernetes cluster, without sacrificing much.

[article] Year in Review: Containers Get Smaller, Faster, More Secure. Sure, lots of AI things happened this year. But plenty moved forward in container-land as well.

[blog] Calling all devs: Code the future of baseball with Google Cloud and MLB. Baseball, AI, and cash. I’m a fan of all those things. And now you can use the second to improve the first and win the last.

[article] Comparing PyTorch and JAX. Both are great. JAX has a lot of fans, and many more added themselves to the list in 2024. Learn more about both here.

[blog] Ends & Means. Bob explores this “ends justify the means” philosophy, and reminds us to consider if our well-intentioned means are actually causing bad outcomes.

[blog] Build Go applications using Project IDX and the Gemini API. I walked through this exact tutorial myself to see if it’s as easy as they said. It was. It’s free to use this charged up dev environment!

[blog] Do Not Fix Bugs Reported in Your Open Source Projects. The premise feels right to me. “Train” the community to document, test, and propose fixes to open source software.

[blog] Firebase App Hosting: Emulators and a look forward to 2025. Firebase experienced a bit of a renaissance this year with new features, more cloud service integrations, and updated interfaces.

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

We had another AI surprise in store today with the new Gemini 2.0 Flash Thinking model. It’s very cool to see the model’s reasoning front and center. Next year, I don’t see how (or why) you’d avoid putting AI into your personal workflow. I even wrote a post today about it!

[article] Google releases its own ‘reasoning’ AI model. Super cool, and it gives very impressive results. You can see its transparent “thinking” right now for free in Google AI Studio. Here’s Simon’s quick take.

[article] InfoQ Java Trends Report – December 2024. Python and JavaScript get the accolades, but Java remains a lowkey hero in this space. It’s widely used and consistently improved.

[blog] Apigee API hub is now generally available. This API leader has been around a while, yet it stays on top in analyst ratings! This looks like a valuable new capability to centralize your APIs.

[article] Salesforce introduces Agentforce 2.0, unveils pre-built use cases. You will not escape agents next year. Unless you go live in the woods and cut yourself off from society. Even then.

[blog] Cloud Pub/Sub 2024 highlights: Native integrations, sharing and more. The Google Cloud messaging and app integration stack is solid. This post has a good recap on some new, interesting capabilities. Plus a sneak peek into what’s coming.

[blog] Dive Deep into Gemini: Explore Starter Apps in AI Studio. I really love when tools and platforms give me a way to immediately see the value of their tech. That’s what these starter apps do for Gemini 2.0.

[article] How To Add Tool Support to AI Agents for Performing Actions. What are tools in relation to LLMs? How do you use them, and why would you? Jani goes deep.

[article] 6 Questions to Find Out How Your Employees Are Really Doing. These are good questions to ask to get beyond the perfunctory “how you doing?”

[article] Quantum Error Correction Update 2024. Do you understand what quantum computers really do? I do not. But I understand more after reading this.

[blog] Applying a Cloud Deploy Policy to an Existing Pipeline. I like this feature. If you have an existing pipeline for shipping your code, you can apply these policies—such as not deploying updates during peak hours—without changing the pipeline itself.

[blog] The status of LLM-based Terraform Generation at the end of 2024. It’s getting better, but still have in-house expertise to understand (and fix) sub-optimal output.

[blog] 321 real-world gen AI use cases from the world’s leading organizations. Once again, we’ve updated this post of real-world use cases for AI. Now up to 321, it’s a great source of inspiration.

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

There’s a lot of programming language content today. I’ve got something for fans of Go, JavaScript, and Java. And still stuff for those who don’t code at all right now!

[blog] “Founder Mode” and the Art of Mythmaking. Charity weighs in on this topic in typical form: thoughtful, long-form, and super honest.

[blog] Detecting objects with Gemini 2.0 and LangChain4j. Guillaume tries out a new model version to see if it does a task better than an old version. This makes me think that if you don’t have a ongoing (not bursty, project style) team doing regular evals of new AI capabilities, you’re going to miss out.

[blog] The Evolution of Infrastructure in the Cloud Era. Will folks finally embrace higher order abstractions or use AI to manage cloud resource lifecycles? Keith explores.

[blog] Celebrating Flutter’s “Production Era.” Flutter has a vibrant, broad community and it’s rad to see how widely it’s adopted. The new Flutter AI toolkit looks good too.

[blog] Go Protobuf: The new Opaque API. This post stayed on the front page of Hacker News for a couple of days, with some spirited discussion. The motivation for this new API is spelled out in the post.

[blog] Tailor your search engine with AI-powered hybrid search in Spanner. Having full-text and vector search in the same engine opens up some simpler possibilities.

[site] 2024 State of JavaScript. I don’t think I’ve seen survey results like this before. Both the types of details (which array features you use) or the way it’s presented. Neat. Just jarring.

[blog] NotebookLM gets a new look, audio interactivity and a premium version. This is looking so good, and will likely establish a new interaction model for LLMs.

[blog] Translating Java to Kotlin at Scale. Read about the “why” and “how” from Meta’s engineering team.

[blog] Google Cloud’s commitment to responsible AI is now ISO/IEC certified. Looks like a good certification to have, and one to ask your provider for.

[blog] Java in the Small. Java isn’t a tiny scripting language, but it’s also not as heavyweight as it used to be. Read this to reset your expectations.

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

I’m sensing some senioritis among co-workers who are getting ready to punch out for the holiday break. Today was a packed day, and tomorrow is too. But it’s clear that work is winding down. For you too?

[article] 4 Listening Skills Leaders Need to Master. I liked this list. A few of these felt new, or at least phrased differently than I’ve heard before.

[blog] FACTS Grounding: A new benchmark for evaluating the factuality of large language models. This appears to be an important release from the Google DeepMind team. Can we find a common way to measure an LLM’s ability to return factual data? Now we can.

[blog] Measure What Matters. Good observability perspectives from Honeycomb here. With bonus YouTube video.

[article] Introducing the DX Core 4. Speaking of measurement, how should you measure dev experience? This post shows off a new unified framework, and this related post digs into benchmark values.

[blog] Reach beyond the IDE with tools for Gemini Code Assist. There’s been a rolling thunder of good releases for this AI-assisted dev tool from Google Cloud. Now, stay in flow by bringing in data from GitLab, GitHub, Sentry, Atlassian, and Google Docs. News here, here, and here.

[article] The 51 most disruptive startups of 2024. Who’s the next big thing? Nobody knows. But there lurking out there, doing cool stuff. Here’s a list of those doing disruptive things.

[blog] Preferring throwaway code over design docs. The idea of using a draft pull request to propose software design ideas is cool. Maybe quit with the giant upfront design in 2025, and get to work prototyping faster?

[blog] BazelCon 2024: A celebration of community and the launch of Bazel 8. There’s a passionate community around this multi-platform, multi-language build system.

[blog] The 3 top mistakes of managers under 30. These sound familiar, and can still be issues as you get past 30!

[blog] Migrating Chainguard’s Serving Infrastructure to Cloud Run. Where you start running your software may not be where it ends up. That’s ok. Chainguard started with Kubernetes, and moved to our serverless container service for a simple setup.

[blog] Your Perfect Infrastructure May Not Be So Perfect. Back to upfront design. It’s hard! This post on the Pulumi blog looks at change-ready architecture.

[article] Is Your Internal Developer Platform Missing Orchestration? Daniel does a good job exploring the approaches one might take in building out a platform.

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