It was a good day, and I’m doing a quick trip to Sunnyvale tonight to talk to a customer tomorrow about how Google uses AI to improve our own software delivery. Should be fun.
[blog] Updated production-ready Gemini models, reduced 1.5 Pro pricing, increased rate limits, and more. Wow. You’re getting premium models at the lowest price. I like the faster output and lower latency too.
[blog] Using Generative AI to Automatically Create a Video Talk from an Article. The concept of NotebookLM has fired folks up—generate consumable media from a pile of written content—and Lak wanted to see if he could create slides and videos from a blog post.
[article] How to Keep Learning at Work — Even When You Feel Fried. This one spoke to me. There’s good advice here on changing your mindset and goals.
[blog] Part1: Switching from Terraform to OpenTofu and Automating your infrastructure on Google Cloud. I haven’t noticed a ton of OpenTofu content, so this one caught my attention. Learn how to use this Terraform fork.
[blog] Customers are putting Gemini to work. Are companies really doing meaningful, business-critical things with generative AI? Some are, yes. This post names them. Related.
[blog] Beyond Backend: Honeycomb for Frontend Observability is Now GA. It’s hard to troubleshoot issues or find opportunities if you aren’t looking at all the data.
[blog] Evaluate open LLMs with Vertex AI and Gemini. Which prompts or models work best for what you’re trying to do? Philipp tries out our managed evaluation service to see what prompts work best on Llama 3.1.
[blog] Everything you need to know about the Gemini API as a developer in less than 5 minutes. Logan gets you up and running with the Gemini API, even if you don’t use Google Cloud.
[article] OpenTelemetry Isn’t the Hero We Need: Here’s Why it’s Failing our Stack. The writer comes from a “competing” product in eBPF, but that doesn’t mean the argument lacks merit.
[blog] Generative AI Cost Optimization Strategies. This AWS post offers up some solid considerations for a cost-aware AI strategy.
[blog] What drives users to Infrastructure as Code? Brian says that GUIs don’t scale, and chaining together CLI commands is brittle. IaC offers a better way to establish repeatable and maintainable infrastructure setups.
[blog] Apply to secure your company’s spot for the Project Starline x HP product. This was probably the most inspiring physical demonstration I participated in this year. Get your company in the queue.
[article] Navigating LLM Deployment: Tips, Tricks, and Techniques. Check this out for further advice on optimizing your LLM deployments.
Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:
Richard Seroter's Architecture Musings 
Leave a comment