OZ

🚨 Calling all .NET Developers! 🚨 Need robust semantic search and vector capabilities but don't want the headache of managing a dedicated vector database server? Introducing Build5Nines.SharpVector: the lightweight, in-memory vector database built purely for C#/.NET. 💡 It enables crucial AI features—like RAG—with lightning speed and zero external dependencies needed for basic use! Key Perks: ⚡ In-Memory Speed (Think instant search). 🧠 Pluggable Embeddings (Local, OpenAI, Ollama support). 🛠️ Built-in for .NET (Fits right into your existing architecture). If your project involves text similarity, semantic search, or LLM context retrieval, this saves you time and complexity. Read the full breakdown and start building: https://lnkd.in/gjwXFGem #dotnet #csharp #softwaredevelopment #AI #dotnetdeveloper #vectordatabase

Is the Azure Portal overdue for a redesign? The Azure Portal has improved over the years, but its core experience still feels like it belongs to an older era of the web. A lot has changed in the last decade: design systems, developer tools, dashboards, AI-assisted workflows, and what we expect from cloud management interfaces. So I asked ChatGPT to help reimagine what a more modern, readable, and useful Azure Portal could look like. The goal wasn’t just a fresh coat of paint. It was better usability: Clearer navigation. More relevant data surfaced upfront. Faster access to recent resources. Azure Copilot more prominent to help you. A layout that feels more like the tools developers use every day. Here’s a concept of what the Azure Portal experience could become. What would you improve first? #microsoft #azure #cloud #azureportal #cloudmanagement

Will GitHub Copilot cost too much with usage billing after June 1st?! I've done some further testing of using GitHub Copilot with other models that cost less than "the best" frontier models like Opus 4.6/4.7 or GPT-5.5. I've found that using a model like GPT-5.4 mini actually performs well enough on many tasks that you wont really tell a difference. I've actually started using one of the smaller models either first, or for the initial "research", and saving the more expensive models for the tasks that really need them. #github #copilot #opus #gpt #openai #microsoft

A few weeks ago, I wrote this article about something we’re already starting to see across the AI industry: the end of the “all you can eat” era. For a long time, many of us assumed unlimited AI access would become the standard. But April 2026 made it clear that growing demand, infrastructure pressure, and compute costs are changing the game. In this article, I share my thoughts on: • why the flat-rate AI model is starting to break • what this means for developers and companies • how AI pricing models may evolve • and the broader impact on the tech ecosystem I’d love to hear your thoughts 👇 https://lnkd.in/gNWMeUD8

What happens when software delivery starts to look more like an automated factory? Not a “developers are going away” factory. More like a delivery system where the repetitive, policy-driven, verification-heavy work becomes automated: CI/CD, testing, security checks, infrastructure provisioning, observability, rollback, compliance evidence, and now AI-assisted engineering. That’s the idea behind a Software Dark Factory (aka AI-automated SDLC) The core argument: The future of software development is not about removing humans from the process. It is about moving humans out of the repetitive factory-floor work and into the control room, where architecture, product judgment, security, ethics, governance, and systems thinking matter most. AI makes this conversation especially interesting because it pushes automation beyond scripts and pipelines into generation, summarization, diagnosis, and adaptation. But the key is bounded autonomy: AI can help create, analyze, and propose, but it should not bypass policy, traceability, review, or accountability. The factory floor may get darker. But the control room needs to get brighter. I’d love to hear your take: Are we heading toward highly automated “Software Dark Factories,” or is this framing too optimistic about where AI, Software Development, DevOps, and platform engineering are going? Read the full article here: https://lnkd.in/eCYjEp2f What do you think: does this future excite you, concern you, or both? #SoftwareDevelopment #ArtificialIntelligence #DevOps #PlatformEngineering #SoftwareEngineering #AIAssistedDevelopment #Automation #CICD #CloudComputing #FutureOfWork #DeveloperProductivity #EngineeringLeadership

🎓 End of an era. I just renewed my AI-102 for the last time. The certification I have carried for 6 years is being replaced by: Microsoft Certified: Azure AI Apps and Agents Developer Associate (AI-103, in beta). The shift in the title says it all. "AI Engineer" → "AI Apps and Agents Developer." The future of this role is building agentic, multimodal, Foundry-native applications. But here is something worth knowing: The new exam does not start from scratch. It assumes you already understand: ✅ Microsoft Foundry ✅ Model deployment ✅ Agent tools and orchestration ✅ Responsible GenAI ✅ Vision and Language All foundations that the current AI-102 already tests. That is exactly why the book I wrote in 2025 is still relevant for anyone preparing for AI-103: 📘 "Ultimate Microsoft Certified Azure AI Engineer Associate (AI-102) Certification Guide" It is the AI-102 guide on the cover, and the AI-103 prerequisites between the lines. Thank you to everyone who studied with me, reviewed the book, or shared feedback along the way. See you on the AI-103 side. 👇 Book link in the first comment. Quick question: are you renewing AI-102 one last time, or jumping straight to AI-103? #AI102 #AI103 #AzureAI #MicrosoftLearn #MicrosoftFoundry #AI #MVPBuzz