Category: Daily Reading List

Took a day-trip up to Sunnyvale and back to meet with a customer, but mixed in some reading along the way. Check it out below.

[article] 10 more bad programming habits we secretly love. I was hacking around this past weekend and likely did all of these. I apologize for nothing. But, it’s good to be aware of the bad behavior!

[paper] Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context. Here’s the direct link to the Google DeepMind paper that’s been updated to include some meaty details about Gemini Flash.

[article] Is Your Mindset About Generative AI Limiting Your Professional Growth? What are five limiting mindsets that might be holding you back from embracing and exploiting AI in your career? This is a good list.

[blog] Avoid the Long Parameter List. Do you love methods that have a dozen input parameters? I do not. This posts advises us to encapsulate groups of parameters into objects.

[blog] Doing Math with Large Language Models. I like how these examples get increasingly more complex, but along the way, the LLM shows its work.

[article] A Comprehensive Guide to Function Calling in LLMs. This is a good overview, although there’s a lot more to dig into for each individual model and use case.

[article] The science behind DORA. Here’s a thorough look at what it takes to properly survey people and analyze the results to arrive at meaningful software delivery performance metrics.

[blog] Generative Q&A with LangChain, Gemini and Chroma. As folks discover Gemini and what it can do, they’re writing up more organic walkthroughs of using it with their tools of choice. Cool to see.

[article] Your Social Media Presence Can Help You Land (or Lose) a Job Opportunity. I could argue that I got my last four jobs because of my writing and social presence. It’s a job in itself, but how you present yourself online will likely have career implications.

##

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

Happy Monday. I took time away from work this weekend which was much needed. I still did some side coding, but I spent the majority of the time reading and messing around with the kids. I had some reading to catch up on, and enjoyed the items I found below.

[article] The dilemma of quality versus speed is false. Good point that “quality is the prerequisite for speed” ad that you won’t have speed without investing in ongoing quality.

[article] How Technical Debt Can Impact Innovation and How to Fix It. We could argue over some of the conclusions here, but the fundamental point seems sound: architectural debt is expensive and limits your progress.

[blog] Deploying GitHub Action Runners on GKE with dind-rootless. If you want to run self-hosted GitHub Action runners on Kubernetes, this is a good walkthrough. Check out the companion piece with troubleshooting tips.

[blog] What makes a good REST API? Whether you’re building your first REST-style API or your hundredth, this seems like a useful guide for what to consider.

[blog] Architectural Blueprints for RAG Automation: Advanced Document Understanding using Vertex AI Search. This is a detailed look at a managed service that helps with grounded your LLM in (unstructured) data.

[blog] Ahrefs Joins Others in Suggesting That On-Premises Hosting Can Be More Cost Effective than Cloud. I can’t speak to everyone’s unique situation, but it’ll be rare when cloud is more expensive than the loaded cost of on-premises infrastructure.

[article] Slack privacy furore: Company updates legals, says no customer data used to train genAI. Yikes. I’m glad they’re doing the right thing now. Systems like Slack store a lot of sensitive data!

[article] Maker vs. Manager: How Your Schedule Can Make or Break You. Some powerful points in here, especially if your job requires periods where you’re either a “maker” or “manager.” How are you balancing those duties?

[blog] Collaborative ML research projects within a single cloud environment. Will we see more of these cross-team or cross-org projects? I hope so. It requires the right sort of platform, however.

[article] Harness Survey Surfaces Raft of DevOps Challenges. Don’t read this if you’ve had a tough day. These results are depressing. We’ve got to help software teams get a handle on their complexity.

[blog] Hugging Face + Google Visual Blocks. This is a very cool way to see ML pipelines in action, and build browser-based pipelines yourself.

##

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

Today was a day to catch up a bit. Are you going into the weekend feeling behind, or relatively on top of things? I’ll gift you a shorter reading list today so that I don’t mess up your todo list.

[blog] Test Failures Should Be Actionable. Short post, but an important point. Do you write tests this way?

[blog] How Google’s AI model Gemini got its name. I like this story, as it’s fun to see the background behind a name that will last.

[blog] The #1 Platform Engineering Problem You’ve Never Heard Of: Platform Decay. There is a lot written about setting up a platform, but much less about keeping it healthy over time. This post tries to address the latter.

[article] Use Active Listening to Help a Colleague Make a Hard Decision. Many of us like to jump in and solve problems in front of us, but it’s often better to be a better listener. Here are four types of active listening support.

[blog] 100 things we announced at I/O 2024. Here’s a good list to skim to see what’s available now, and coming soon, in the consumer and builder space.

[blog] To tune or not to tune? A guide to leveraging your data with LLMs. Today, there are a handful of valid ways to get better responses from an LLM. There might be more tomorrow! But for now, how do you decide among current options? This post has a useful breakdown.

[article] AI coding tools are your interns, not your replacement. I think AI assistants for devs are more helpful to junior folks than other might believe. But I get the various perspectives shared in this article.

##

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

Heading home today after a few days away from home, and I learned a lot this week. I enjoyed the items in my reading queue today, and I think you’ll like them too.

[article] Use Engineering Strategy to Reduce Friction and Improve Developer Experience. Can it be “simple” to create an engineering strategy that aligns the team, improves focus, and helps you make decisions faster? Will tries to prove that in this talk.

[blog] Optimize Your Athletic Performance with Gemini and Vertex AI. You should read this. We built this amazing physical demo that folks love, but it also shows all sorts of ways to use AI and cloud services to get an outcome that non-techies understand and enjoy.

[blog] Automatically disabling leaked service account keys: What you need to know. It’s opt-in, but I really like the idea of automatically disabling account keys found in public repos within GitHub or Gitlab.

[blog] The Cost Crisis in Metrics Tooling. Charity says that folks are collecting so many metrics that cost us a fortune, and don’t tell us much.

[blog] Introducing Aparna Sinha, Capital One’s Head of AI Product. My former colleague has an exciting new role, and it makes me wonder if we’ll see this position created at many different enterprise companies.

[article] Google Cloud’s Matt Renner: “Half of all AI PoCs are just a bad idea.” I don’t mind if the POCs are bad ideas as long as folks can run quick experiments. But Matt’s point is sound, and Google has staffed up to reduce your rate of failure.

[article] Research: What Companies Don’t Know About How Workers Use AI. Trust is a big angle in this story, whether that’s between management and employees, or models and companies.

[blog] Why do only a small percentage of GenAI projects actually make it into production? Continuing a theme from the last few articles, this one explores why companies can struggle to convert their AI dreams into monetizable products.

[docs] LlamaIndex on Vertex AI for RAG overview. AI experiments can go faster when you use frameworks that make it simpler. Just 30 lines of code to ingest data, index it, augment LLM retrieval, and generate results?

[article] How does developer satisfaction lead to improved productivity? Lizzie looks at the SPACE framework and some self-reported data that can help you make better decisions about your dev experience.

[article] Kora: A cloud-native redesign of the Apache Kafka engine. This looks like a major update, and one that’s heavily influenced by Confluent’s work running a global, multi-cloud managed service.

[blog] What’s new in Flutter 3.22. WebAssembly, and more in this latest release of the popular framework for mobile and web apps. More coverage here.

##

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

After yesterday at Google I/O, I was in Vegas today for a Gartner conference. It’s been a wonderful chance to learn from “regular” folks with normal tech challenges.

[blog] Google I/O 2024: An I/O for a new generation. Here’s a good overview of I/O based on Sundar’s keynote.

[blog] Sociotechnical Maestros with Gene Kim. Listen to the podcast, or read the transcript here. Product wizard John talks to DevOps and culture guru Gene about empowerment, vision, and other considerations for high performing teams.

[paper] Global deployment with Compute Engine and Spanner. What does it look like to build a globally available system with VMs and databases? Here’s a complete reference guide.

[blog] Introducing PaliGemma, Gemma 2, and an Upgraded Responsible AI Toolkit. New, and better open AI models now ready to use.

[article] Why and How Teams Are Replacing External Database Caches. Are separate caches uncool now? Are databases performant enough? This post looks at some examples.

[blog] What’s new in Firebase at I/O 2024. Firebase came out STRONG at Google I/O. You’ve got Firebase App Hosting for serverless compute beneath your Next.js or Angular app, a nifty-looking relational database offering, new integrations with the Vertex AI platform, and the new Genkit for building generative AI apps.

[blog] Done is Better Than Perfect. On one hand, this is how you get technical debt. But on the other, sometimes those efforts at perfection create more debt and negative business impact.

[blog] Introducing VideoFX, plus new features for ImageFX and MusicFX. These are wild tools for creative folks to use when building what they love.

[article] “Developers should own security. And test in production.” I saw a lot of “shift left” content today. Observability apparently needs to shift left too. Or, let’s have platforms do more, by shifting down.

[blog] Vertex AI at I/O: Bringing new Gemini and Gemma models to Google Cloud customers. Context caching is a big deal. And the updated tools for those building agents are great.

[article] Three Reasons DevOps Should Consider Rocky Linux 9.4. I like what this team is doing, and you should definitely consider Rocky Linux as a distribution.

##

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

Whew. Today was an excellent day at Google I/O and I left inspired and exhausted. Between Cloud Next and now I/O, I need a conference break. But I’ll tell you, I wouldn’t exchange platforms or potential with any other cloud right now. We’re in a terrific spot.

[blog] Google I/O 2024 recap: Making AI accessible and helpful for every developer. We announced and/or released many new things today, and this is worth a skim if you care about modern apps.

[article] Amazon Web Services CEO Adam Selipsky steps down; senior exec Matt Garman named AWS CEO. Wow, big change for our friends at AWS. Corey thinks its a good move. Maybe! I sure like having a product-thinker and engineering mind leading Google Cloud. Also, OpenAI’s co-founder and chief scientist moves on. Lots of musical chairs today!

[blog] Machine Learning on GCP: From Notebooks to Pipelines. This author says that too many folks think AI can be purely addressed with notebooks instead of proper real ML pipelines.

[blog] 4 Rituals To Keep You Happy All The Time. I’m pretty chipper, but even I’m not happy all the time. Nonetheless, Eric offers a few ways to re-program yourself to bias towards happiness.

[blog] Strategy as a Series of Beliefs. The answer to ‘what is strategy?” won’t be the same if you ask two “experts” in the domain. I like how Dave broke down this way of thinking.

[article] How Do You Measure Developer Experience? Generic guidance on dev productivity or experience doesn’t really land with me. Rather, I want to see it grounded in a real-life scenario. That’s what you’ll find here.

[article] 100X Scaling: How Figma Scaled its Databases. Learn how Figma sharded their database and the benefits they saw as a result.

##

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

I had my rehearsal at Google I/O today, and then attended a half-day offsite with my team. Before that, I read a bunch of things you’ll find below.

[article] AI-assisted code authoring at scale. When Meta rolled out AI assisted tools to their engineers, what were the results? This post reviews a paper with the details.

[blog] Keeping Up with Tech. This post was mainly about being a t-shaped developer, with a mention of how to find the “next thing” to learn about.

[blog] I built the same app with Flutter, React Native, and Ionic. You can learn a lot by using different languages or frameworks to solves the same problem. Here’s a look at three popular app frameworks.

[article] How to Present to an Audience That Knows More Than You. Many folks dislike public speaking. It can be even more nerve-wracking if you know the audience is smarter than you on a given topic. That’s often the case! This article has some advice for those situations.

[blog] Unlocking Codebase Insights with Vertex AI Gemini 1.5 Pro: A Comprehensive Guide. What questions would you ask of an inherited codebase? Where are the security issues? How should I troubleshoot? What does each component do? This post shows examples for AI-based interrogation.

[article] Do Enormous LLM Context Windows Spell the End of RAG? Somewhat related to the previous piece. If I can send in all my code to an LLM, I may not need it chunked up and vector-stored prior.

[blog] Good Suffering. Bob has a good reminder that “resilience” is a developed characteristic, and talks about having “high standards and low expectations.”

[blog] Using Google’s Gemini Function Calling to Fetch Real-Time Company News and Insights from External APIs. This is a good example of “function calling” as a technique for getting real-time data into your generative AI apps.

[blog] How to (and how not to) design REST APIs. Derek affirms and rejects some “conventional wisdom” around HTTP (or REST) APIs.

[article] IT services spend will reach $2 trillion by 2028: Forrester. Spending is up, especially on cloud. But this notes that cloud spending is still less than 10% of the IT services spend!

[blog] Billed for unauthorized requests? Google Cloud Storage vs. AWS S3. This topic came up a week or so ago when a user got a massive bill for unauthorized 403 requests to their Amazon S3 bucket. AWS is fixing that, and this post looks at whether you’re at the same risk on Google Cloud. Hint: you’re not.

[blog] Podless Kubernetes. As an academic exercise, could you deploy a full Kubernetes cluster and mesh in a single node binary, without any pods running? John did just that.

##

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

It’s the end of a solid workweek. Next week, I’ll be at Google I/O delivering the cloud keynote, and then jetting off to Vegas to talk to analysts at the Gartner event going on there. Those events have wildly different audiences, and I deeply enjoy both.

[article] Gatekeepers Limit Continuous Delivery’s Benefits. The theory of constraints says you have to optimize the limiting factor. If you improve how fast folks can write code, but can’t ship it any faster, what’s the point?

[blog] Building AI-Powered Apps with LangChain: A 2024 Guide. Good details about this framework that has quickly become a go-to for folks building generative AI apps.

[blog] AlloyDB vs. self-managed PostgreSQL: a price-performance comparison. There are, of course, cases where self-managing hardware or software is a better move for you. But the number of cases is shrinking.

[blog] Paramount+: A streaming powerhouse with limitless entertainment. Few folks need a zero-downtime, globally available service. But when you do, or if you want to apply the lessons learned from such an approach, see this post.

[blog] 6 Best Practices for Hosting Developer-Focused Events. The in-person meetup scene seems hot right now, with lots of tech folks getting together to learn. This is a good post for those putting on events.

[blog] Kubernetes 1.30 is now available in GKE in record time. Staying up to date in your Kubernetes cluster isn’t about features; it’s about security and stability updates. Nobody keeps you up to date like our GKE team.

[blog] What’s new with Active Assist: New Hub UI and four new recommendations. Even if the AI hype train has taken focus off cost and security optimization, those things still matter a ton. I like the new security and audit-related “recommendations” we’re offering Cloud customers now.

[blog] The DevRel Guide to Business Jargon. Get definitions of common sales and marketing terms like TAM, ARR, conversion rate, and more.

[blog] How chaos testing adds extra reliability to Spanner’s fault-tolerant design. I was impressed by the extent to which we try to “break” Spanner to test its reliability. You might find inspiration for tests you want to run against your own systems.

##

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

Each day, I just read whatever pops into my feeds and newsletters. I’m not looking for a theme, but sometimes one pops out at me. Today? It seemed like a lot of content focused on optimization and doing things the right way. For example, check out items below about improving dev experience, efficient hosting of streaming platforms, doing CI well, controlling ops metrics volume, and scaling Kubernetes.

[article] Boost Developer Productivity by Reducing Their ‘Paper Cuts’. If you’re a small or mid-size tech team, what can you do to build a better dev experience for your team? This article offers an outstanding blueprint.

[article] 20 Lessons From 20 Different Paths to Product-Market Fit — Advice for Founders, From Founders. This is useful for founders, yes, but also for anyone building products that aim to solve a problem.

[blog] Three Laws of Software Complexity (or: why software engineers are always grumpy). Do well-designed systems always degrade into badly designed ones? Mahesh makes that argument, and others.

[paper] Capabilities of Gemini Models in Medicine. There’s 30+ pages of description and data in this new paper, and it may inspire you for use cases outside of medicine.

[blog] Yahoo Benchmarks Dataflow vs Self-managed Flink Efficiency for two Streaming use-cases– Which is More Cost-Effective? Some Yahoo! engineers wanted to compare the cost and performance of running their own data processing stack (on Kubernetes) or using a managed service.

[article] How is Flutter Platform-Agnostic? This framework renders interfaces across desktop, web, and mobile. how does it do that? Good deep dive here.

[article] Generative AI interest now shapes talent strategy, employers say. Makes sense. This likely impacts who you hire, how you train, and what you build.

[blog] Optimizing CI in Google Cloud Build. Darren wrote a fantastic post that’s helpful whether you’re using the Google Cloud services he mentions, or not.

[article] Microsoft and OpenAI’s increasingly complicated relationship. Partnerships are critical, and Microsoft does them well overall. But I’m glad Google has its own foundation models to bet on. A related perspective here.

[blog] Controlling metric ingestion with Google Cloud Managed Service for Prometheus. If you’re drowning in metrics, and paying a lot for the privilege, consider some of the steps called out here.

[blog] The Streaming War Is Over and All It Cost Was the Entertainment Industry. Bundles versus individual best-of-breed is a pendulum. For streaming, we’re inevitably back at bundles.

[blog] The surprising economics of Horizontal Pod Autoscaling tuning. Some good advice here on autoscaling, with a clear view of the tradeoffs of setting different resource targets.

##

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

Weirdly, many of the things I read today seemed to relate to other things I read today. I’ll chalk it up to a weird wrinkle in the universe.

[blog] The Ultimate Guide to building Developer Tool Websites. This post justifies the title of “ultimate guide.” It’s a really good look at how to build websites that developers find useful.

[blog] A tour of Gemini 1.5 Pro samples. Speaking of useful things offered on dev websites, Mete looks at code samples here. He looks at how to use Gemini to process audio, video, and multi-modalities at the same time.

[blog] Humility: The Secret Ingredient to Modern Tech Agility. We don’t celebrate humility (which I guess would be ironic in some way), but it’s a key ingredient to agile leaders and teams.

[blog] AlphaFold 3 predicts the structure and interactions of all of life’s molecules. Talk about humbling! We’re now learning more about the biological world than ever before.

[[blog] The biggest effect on code quality. This author says its not about tools or skills; code quality is impacted by teams working in “crunch mode” and making mistakes.

[blog] What causes new engineers to “sink or swim”? If you want new developers on your team to be successful, check out this guidance for what determines that.

[article] What Can Incident Teams Learn From Crisis Management? My friend Dormain wrote up an insightful piece on the need for strong communication during incidents.

[blog] Maintain business continuity across regions with BigQuery managed disaster recovery. I just mentioned incident management, and here’s a piece on a new feature of BigQuery that offers automatic failover in the unlikely case of a regional outage.

[blog] Simplifying and standardizing software at scale. Here’s how the McDonalds engineering team is building golden paths to help teams build, ship, and run software better.

[blog] What is Istio? The Kubernetes service mesh explained. I don’t know if Istio helps you simplify things, but it’s definitely a powerful mesh that’s getting easier to use.

[blog] Observability, Telemetry, and Monitoring: Learn About the Differences. Stop using these words interchangeably! I’m mostly saying that to myself. But also, you might find this clarifying.

##

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