Daily Reading List – July 23, 2024 (#361)

I like remote work as much as the next person, but I’ve been in at least two long-running group chats at work today that would have been “solved” in minutes if we were all sitting in a room together. Yes, video chats help, but nothing beats an intense, spontaneous collaboration of people in the same space.

[report] State of Enterprise Tech 2024. There’s a lot here in a survey of enterprise tech leaders. Some of the AI investment findings were fascinating—why are so many folks fine-tuning models?—and answers on dev productivity reinforced other data I’ve seen.

[blog] Meta’s Llama 3.1 is now available on Google Cloud. The latest and greatest model from Meta is out, and already available in your public cloud of choice. When you try it out in Vertex AI on Google Cloud, you get some nice features for tuning, agent building, and grounding.

[blog] Profiling in Go: A Practical Guide. Go has a built-in profiler to collect CPU and memory usage that you can then study afterwards and use to optimize your app. Good walkthrough.

[article] Top takeaways from the CrowdStrike outage for IT teams. There will be lots of posts like this in the coming days and weeks.

[blog] The Factory metaphor makes sense for building applications, but not for deployment and operations. Brian thinks that referring to the delivery of software as an “assembly line” is a valid metaphor, but it breaks down when thinking of operations tasks. Agree? Disagree?

[blog] The Cloud Run Security Gap You Didn’t Know You Had (and How to Fix It). Principles of least privilege need to be applied to your cloud services too. Make sure you know how to allow (or disallow) access, as JK shows here.

[blog] Maestro: Netflix’s Workflow Orchestrator. Big post from the Netflix team about their open source product for managing “large scale workflows such as data pipelines.”

[article] 3 Ways To Improve Your Container Build Process. It might not be the most interesting work, but tuning your container build process pays dividends down the road.

[blog] How Airbnb Smoothly Upgrades React. Solid post about what it means to keep the frontend in a good state, and how the team at Airbnb does it.

[article] Avoid These 3 Pitfalls When Giving a Sales Presentation. This is targeted at folks doing sales presentations, but this advice works for anyone “pitching” to an audience. It could apply to architects selling a new internal strategy, a proposal for a new product, etc.

[blog] Running Google’s Gemma2 LLM locally with LangchainJS & Ollama. Good demo of using a mix of tools to build an AI app.

[blog] Are We Ready for Multi-Image Reasoning? Launching VHs: The Visual Haystacks Benchmark! It’s one thing to find a “needle in a haystack” of textual content, but what about a large image set? The Berkeley AI Research team shares new ideas.

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