Daily Reading List – October 22, 2024 (#424)

I had a fun day with a couple of customer presentations and live demos. My real job is leading a terrific team here, but it’s great to spend some time each week with real, live users.

[blog] Highlights from the 10th DORA report. People are going to be talking about this. The biggest research program into software delivery performance shipped its 10th edition. And it’s full of surprising results about AI, platform engineering, and sustained performance. Download here.

[article] Taking Advantage of Cell-Based Architectures to Build Resilient and Fault-Tolerant Systems. This is an article that’s part of a series and promotes the idea of “cells” of infrastructure that are domain based and isolate failures to a fraction of the architecture.

[blog] Using AI Generated Code Will Make You a Bad Programmer. Here we go, a frisky counter argument for AI assisted coding! I don’t agree with much of it, but I see the points.

[blog] How We Generated Millions of Content Annotations. Even in a generative AI world, there’s plenty of need for classification models. Spotify talks about how they scaled human expertise to add annotations to digital content at scale.

[blog] Unraveling Cinematic Connections: Analyzing Movie People Relationships 🤫 with Spanner Graph! Fun post that shows how a built-in graph capability for Cloud Spanner makes it easier to discover nested relationships.

[blog] Introducing computer use, a new Claude 3.5 Sonnet, and Claude 3.5 Haiku. Very impressive work from Anthropic. The “computer use” feature is kinda bonkers. Simon played around with it. Try it now on Google Cloud.

[article] Strategies for making doc bugs actionable. Bugs without a clear response are frustrating. Tom looks at how to help engineers submit actionable bugs for docs, and how to make the bugs themselves for actionable.

[paper] The Ultimate Guide to Fine-Tuning LLMs from Basics to Breakthroughs: An Exhaustive Review of Technologies, Research, Best Practices, Applied Research Challenges and Opportunities. That’s a crazy-long title. But also a very long, interesting paper that will teach you a lot about fine-tuning an LLM.

[blog] Develop Serverless Applications on Cloud Run: Google Cloud Challenge Lab Walkthrough. This is a good step-by-step for getting an app built and deployed to one of my favorite cloud services.

[blog] The Narrows Bridge: From Open Source to AI. This might be too inside-baseball for some folks, but I like Stephen’s thoughtful look at how open source does (or really, doesn’t) apply to AI.

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