Copilot, Cursor, and Custom LLMs: Navigating the New .NET Developer Experience
summary
Compares real results: Copilot in Visual Studio gives deep MSBuild and debugger integration while Cursor Composer ships faster IDE-first features like workspace indexing and project rules; running Ollama local models plus a vector index yields high‑fidelity RAG on private code. Cites numbers: generated output leads to ~20% more code and ~40% more review time, an initial 1–2 month velocity spike then ~9% higher complexity and ~4.9x static warnings. Shows a working video‑splitter demo, Visual Studio agents, Copilot instructions.md and Cursor rules.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 3:17Productivity lie and AI hype
- 11:00LLMs and harnesses
- 18:02Context for AI tools
- 24:24Running LLMs with Ollama
- 26:10VS AI features demo
- 32:27Cursor and project rules
- 33:49Local LLM harnesses
- 36:16Debugging AI in VS Code
- 37:19New developer loop
- 41:10AI adoption study
- 43:37Verdict and AI future
- 45:03Adopting AI tools
- 47:52Q&A
related talks
39mPower of AI and Code Quality and Performance
19mMicrosoft Copilot Demystified: A 20-Minute Introduction
59mGetting the Most Out of .NET Development with Visual Studio
54mMaximize joy minimize toil with great developer experiences
42mWhat’s new in GitHub Copilot and Visual Studio
01h 02mVisual Studio Toolbox Live
06h 22mUsing AI to Enhance Developer Productivity
01h 05mThe Next Generation of Developer Productivity with GitHub Copilot & Visual Studio