#Announcements

• .NET 10 Preview 2 is now available! - .NET Blog - We’re seeing a some new minor functionality in C# 14 with partial constructors and events and more performance improvements to arrays, so some minor wins.
• Announcing Rust 1.85.1 | Rust Blog - We get a fixed doctest, where rust will run and test your documented code, to ensure your code comments are still valid, I ❤️ this feature,z + other minor bug fixes.
• Google Online Security Blog: Announcing OSV-Scanner V2: Vulnerability scanner and remediation tool for open source - Google release v2 of their open sourced and add layer-aware container scanning for Debian, Ubuntu, and Alpine container images. This tool is incredibly powerful for finding all listed vulnerabilities in your container with history behind changes.
• MSTest 3.8: Top 10 features to supercharge your .NET tests! - .NET Blog (& Microsoft.Testing.Platform: Now Supported by All Major .NET Test Frameworks + Introducing the MSTest Runner) - MSTest, a replacement to VSTest which is built into dotnet CLI, has made some fantastic improvements and added features. It may be time to start assess if you should be changing over from xUnit or NUnit, just being able to build your tests to an executable your can run on any platform or instruction set nearly warrants it.
• Copilot in Bing Webmaster Tools is Now Available to All Users - Bing webmaster GA’s co-pilot, providing AI driven analysis and responses for how to optimise your website, and will also help you identify issues on your website. Lightning house with AI, I’m sure google will follow this up with Gemini 2.0 in your chrome DevTools performance tab.
• Upcoming Updates for Azure Pipelines Agents Images - Azure DevOps Blog - Updates to Azure DevOps pipeline agents, mainly around Ubuntu 20 & 24 updates which will require rolling brownouts, so you may run into pipeline failures on these images.

#Articles

• So, What’s the Difference Between Observability and Monitoring? | Honeycomb - Monitoring is notification based, think dashboards, alerts, etc. Whereas, observability is the data we log and how we enrich it with data for debugging purposes, such as correlation.
• Make Ubuntu packages 90% faster by rebuilding them - Recompiling linux packages, the example being jq, with TCMalloc over the default Ubuntu C allocator and using mimalloc, for some drastically improved speed over the OS defaults.
• Karatsuba Matrix Multiplication and its Efficient Custom Hardware Implementations - taking a well known algorithm and inspecting the usage and performance on custom circuitry to run it.
• Container Network Interface (CNI) in Kubernetes: An Introduction | by Homayoon (Hue) Alimohammadi | Mar, 2025 | ITNEXT - How CNI and CNI plugins work with explanations of how it sets up your kube networks for your pods for internal and external traffic.
• Next.js vs TanStack | Kyle Gill - A love for TanStack and replacing the Next internals of a react app with TanStack.
• How Netflix Stores 140 Million Hours of Viewing Data Per Day - A rehash of Netflix Tech Blog’s article, which covers how they started on Apache Cassandra and evolved from compressing, to leaving viewing data uncompressed for hot data, chunking their data based on customer to scale horizontally, chunking also on preview vs full show and archiving + compressing of infrequently accessed data.
• Implementing Generic Types in C - by Sir Whinesalot - Interesting ways of implementing generic concepts into C.
• Acquiring Tokens For Non-Graph APIs In Azure Functions · Den Delimarsky - Interesting you need to jump through some hoops to retrieve Azure hosted function app tokens, I’m sure Az CLI or powershell would be able to get this.
• GDC 2025: How Build Insights Reduced Call of Duty: Modern Warfare II’s Build Times by 50% - How Microsoft helped their newly acquired company, Activision, decrease their C++ build times significantly through some deep analysis and funny optimisations such as removing forces in-lining. They’re key takeaways are insightful:

• Use Build Insights to identify bottlenecks. Windows Performance Analyzer’s timeline view is incredibly useful for narrowing down slowdowns.
• Be mindful of force inlining. Excessive inlining can significantly increase compile times, especially when applied deep in the call stack, where it can be hidden and lead to a combinatoric explosion of calls. Also, it can limit the compiler’s ability to make optimization choices, leading to worse runtime performance.
• Leverage Unity/Jumbo builds. Combining multiple source files into a single unit helped Activision shave one minute off their build time. Checkout our blogpost on Support for Unity(Jumbo) files in Visual Studio, to learn how you can take advantage of this for your builds!
• Continuously measure and refine. Optimization is an ongoing process, not a one-time fix.

#.NET

• .NET Aspirations - Use ASP.NET Core HTTPS Development Certificate · Alexandre Nédélec - Spinning up HTTPS in Aspire, replicating a more prod like environment locally.
• Adding Cloudflare Turnstile to an ASP.NET Core website - A guid on how to put in a CloudFlare turnstile (CF’s replacement to CAPTCHA), into your HTML page, and check from your server if it was valid.
• .NET Aspire Dashboard - The mystery of the hidden endpoint - A bit of a gotcha moment in Aspire, as it only uses the first launch settings to get the application URL.

#Infrastructure

• Dynamically Scaling Containers With KEDA and IBM App Connect - How to scale out you pods based on the number of messages in your IBM message queue.

#Architecture

• Screaming Architecture: Not Driven By Entities - CodeOpinion - Explaining differences between vertical slice, event driven, CRUD & streaming code architecture and pros/cons of how it reflects, or doesn’t reflect, your business’s domains and workflows.

#AI

• Microsoft and NVIDIA accelerate AI development and performance   | Microsoft Azure Blog (+ Accelerating Agentic Workflows with NVIDIA AgentIQ, Azure AI Foundry and Semantic Kernel | Semantic Kernel) - Microsoft unveil NVIDIA AgentIQ in Azure AI Foundry, allowing easy development from VS Code for NVidia’s model.
• What Are AI Agents Really About? + Agentic AI: Workflows vs. agents - Explains AI agents and how they work.
• Agentic AI - What and How! - A great video on explaining what is Agentic AI’s are and in essence they’re AI agents which are more autonomous, responding to events and automatically coming up with an action plan using internal/external resources (RAG) and running those tasks. And the video goes over building one.
• Building Agentic Flows with LangGraph & Model Context Protocol - A guide on building your own Agentic AI flows, utilising MCP (Model Context Protocol) to pull data into your AI researcher for decision making.
• EP154: What is MCP? - ByteByteGo Newsletter - Explains in what Anthropic’s open protocol Model Context Protocol (MCP) is and how the MCP client transforms the model’s want to a call to the MCP server, which acts as the middleman to the external resource (e.g. database, an API, etc).

There is so much chatter and content coming out for Agentic AI and MCP, it’s a bit overwhelming, but Google’s examples have a practical example which makes it a bit easier to understand:

• generative-ai/gemini/agents/research-multi-agents/intro_research_multi_agents_gemini_2_0.ipynb at main · GoogleCloudPlatform/generative-ai
• GoogleCloudPlatform/agent-starter-pack: A collection of production-ready Generative AI Agent templates built for Google Cloud. It accelerates development by providing a holistic, production-ready solution, addressing common challenges (Deployment & Operations, Evaluation, Customization, Observability) in building and deploying GenAI agents.
• generative-ai/gemini/agent-engine/langgraph_human_in_the_loop.ipynb at main · GoogleCloudPlatform/generative-ai

#Other Interesting Reads

• Debug Visualizer - Visual Studio Marketplace - Found a really nice debugging extension which allows you to visualise your data.
• Nvidia announces two ‘personal AI supercomputers’ | TechCrunch - NVidia announce 2 new PC’s claiming their “AI supercomputers”, DGX Spark and DGX Station.