Generating Visuals with gpt-image-2 in VS Code

Agent Skills in Visual Studio have captured my interest, particularly from a Data & AI perspective. Rather than repeating prompts or depending on always-on instructions, Agent Skills serve as reusable instruction sets that Copilot can automatically apply when relevant. For Data and AI work, this is significant: we can encode team playbooks such as: - Data quality checks - Pipeline patterns - Governance guardrails - Model validation steps These can be shared as “skills” in the repository, ensuring everyone achieves consistent outcomes. My takeaway is that this represents a move towards scaling engineering standards and domain knowledge, rather than just speeding up code generation. I’m curious to hear how others view this shift. #DataEngineering #AIEngineering #Copilot #AgenticAI #VisualStudio

Running LLMs locally has become surprisingly accessible with tools like Ollama. With Ollama, you can pull and run open models directly on your local machine. Pair that with Continue extension(https://continue.dev/) inside Visual Studio Code, and you get a pretty capable local AI coding setup. Using the Ollama provider in Continue, you can connect different models for local agentic development and experimentation. But there’s an important tradeoff: * Smaller models like Llama 3.2 can run reasonably well even on modest hardware. * Larger models like Qwen 3.6 are a different story. To get fast responses, you usually need expensive GPUs and a fairly powerful machine. That’s where OpenRouter(https://openrouter.ai/) becomes useful. Instead of running massive models locally, you can access models like Qwen through OpenRouter and still use them directly inside VS Code via Continue extension. Why would engineers want to test different models anyway? Curiosity. Every model feels a little different: reasoning style coding ability speed instruction following agentic behavior context handling Some models now support massive context windows too. For example, Qwen 3.6 Plus on OpenRouter supports a 1M token context window, which opens up interesting possibilities for large repositories, long-running agent workflows, and deep codebase analysis. And honestly, experimenting with different models is one of the fastest ways to understand where this space is heading.

I used to spend hours on Game server architecture — real-time multiplayer with WebSockets tasks. Then I tried vibe coding — letting AI handle the scaffolding while I focused on design. Result: 3x faster prototyping, same code quality. The workflow: 1. Describe the architecture in plain English 2. AI generates the boilerplate 3. I review, refactor, and optimize 4. Ship in days instead of weeks The developers who will thrive in the next 5 years aren't the ones who type the fastest. They're the ones who think the clearest. Have you tried AI-assisted development? What was your experience? #GameDev #IndieDev #GameDesign

I used to spend hours on Game server architecture — real-time multiplayer with WebSockets tasks. Then I tried vibe coding — letting AI handle the scaffolding while I focused on design. Result: 3x faster prototyping, same code quality. The workflow: 1. Describe the architecture in plain English 2. AI generates the boilerplate 3. I review, refactor, and optimize 4. Ship in days instead of weeks The developers who will thrive in the next 5 years aren't the ones who type the fastest. They're the ones who think the clearest. Have you tried AI-assisted development? What was your experience? #GameDev #IndieDev #GameDesign

🆕 ReSharper Is Bringing More AI Agents Into Visual Studio The Early Access Program for ReSharper 2026.2 is here, and this first preview is all about breaking the AI vendor lock-in in Microsoft Visual Studio. We’re introducing Junie – the LLM-agnostic AI agent from JetBrains – as our first step toward full ACP (Agent Client Protocol) support in ReSharper inside Visual Studio. It’s an early preview, but the direction is clear: more choice, less lock-in, and the freedom to choose the best AI agent for the job at hand. Try the EAP (it’s free!), explore Junie in ReSharper, and tell us which agents you’d like to see supported next. Learn more 👉 https://lnkd.in/dwikm2T2 #dotnet #csharp #visualstudio

On .NET Insights this week, David Ramel explores where developer tooling is headed, and how AI fits into the bigger picture. From early insights into Visual Studio 2027 to hands-on experiences with VS Code Agents, it's clear that Microsoft is pushing toward more agentic, action-driven development workflows. At the same time, updates to VS Code continue to expand AI capabilities, while .NET 11 previews emphasize core coding fundamentals, showing that strong engineering practices still matter alongside AI. And beyond the tools, we look at the human side of development, with practical guidance on debugging team challenges before they impact delivery. The takeaway: the future of development is a blend of AI, tooling and strong fundamentals, not one replacing the other. Catch the recap below and read the full stories on our site: https://lnkd.in/fHhXGtc #AIforDevelopers #VisualStudio #VSCode #DotNet #SoftwareDevelopment

✨ In my free time, I like turning simple ideas into something that feels like sci-fi Built Gesture Cube — a touchless 3D interaction where your hands become the controller. 🧊 👉 Point to rotate 🤏 Pinch to scale ✊ Fist to reset No mouse. No screen touch. Just gestures + computer vision. It genuinely feels like giving the browser a new way to understand humans — more natural, more intuitive. 🎥 Demo below — what should I build next with this? #CreativeCoding #ComputerVision #ThreeJS #JavaScript #AI #Innovation #SideProject

I have been testing the Chat Customizations Evaluations extension for Visual Studio Code and it has quickly become one of the more useful AI development tools in my workflow. The extension helps analyze, validate, and improve prompts directly inside Visual Studio Code. Features like contradiction detection, semantic ambiguity analysis, cognitive load assessment, and prompt coverage checks make it especially valuable when building reusable AI skills and agent workflows. I have been using it heavily while creating new skills for Claude, GitHub Copilot workflows, and VSCode automation projects. It does a good job catching weak spots in prompts before they become larger issues downstream. If you are building AI agents, structured prompt pipelines, or reusable automation skills, this extension is worth exploring. #AI #PromptEngineering #VSCode #ClaudeAI #GitHubCopilot #Automation #LLM #ArtificialIntelligence #DeveloperTools

In the next set of videos in the #VSCode Learn series (Episodes 7–14), we go deeper into agent‑first development in Visual Studio Code, building on the foundational concepts introduced earlier. These sessions focus on: - Applying agent-based workflows to more advanced development scenarios - Understanding how agents interact across tools, environments, and sessions - Improving control, observability, and confidence when working with AI agents - Moving from experimentation to real, production‑ready developer workflows If you’re exploring how AI agents can become a first‑class part of your daily development experience in VS Code, this part of the series helps bridge the gap between concepts and practical adoption. ▶️ Watch the full playlist here: https://msft.it/6046vygH6 #MSFTAdvocate #AIAgents #AgenticAI #DeveloperExperience #AIEngineering #Microsoft

In the next set of videos in the #VSCode Learn series (Episodes 7–14), we go deeper into agent‑first development in Visual Studio Code, building on the foundational concepts introduced earlier. These sessions focus on: - Applying agent-based workflows to more advanced development scenarios - Understanding how agents interact across tools, environments, and sessions - Improving control, observability, and confidence when working with AI agents - Moving from experimentation to real, production‑ready developer workflows If you’re exploring how AI agents can become a first‑class part of your daily development experience in VS Code, this part of the series helps bridge the gap between concepts and practical adoption. ▶️ Watch the full playlist here: https://msft.it/6045vKbKf #MSFTAdvocate #AIAgents #AgenticAI #DeveloperExperience #AIEngineering #Microsoft