Search for a command to run...
Program.cs looks harmless because it usually starts as a few lines of setup code. Create the builder. Register some services. Add authentication. Map the endpoints. Run the app. That makes it easy to
Kestrel is one of the reasons modern ASP.NET Core feels so different from the old .NET web stack. You can put a small Minimal API in front of it, run a load test, and get numbers that would have sound
When Allocations Become Your Real Bottleneck
Cursor Composer 2.5 is worth looking at because it pushes AI coding further away from autocomplete and closer to a real development loop. For .NET engineers, thats where things start looking interesti
SignalR feels simple when you have a chat window, a live dashboard, or a small notification feature. Add a hub, connect from the browser, call a method, broadcast to a group, job done. That simplicity
TurboVec is interesting because it attacks one of the problems that appears after a RAG system starts to grow. Embeddings are easy to talk about when you have a few thousand chunks. They become much h
There is a trap in this question. When someone asks whether a .NET endpoint can handle a million requests per second, the instinct is to jump straight into Minimal APIs, Kestrel tuning, JSON serialisa
A lot of .NET AI code starts the same way. You install a provider SDK, create a client, pass in a prompt and get a response back. For a prototype, thats fine. For a production application, it can turn
When observability comes up you might ask whether you should use OpenTelemetry or Application Insights. That sounds reasonable, but it puts two different things in the same box. OpenTelemetry is a sta
