The fine line between mistake and innovation in LLMs

Large language models are probabilistic systems. While they're improving rapidly, they can still fail at basic tasks or hallucinate answers - sometimes without us even noticing. In this post, I explore why that happens, how we can mitigate it, and why these mistakes are fundamentally exciting.

Yet another connection reset

Envoy and Kestrel: not exactly love at first sight

Automating my gym bookings with a serverless assistant

My journey to building a serverless assistant that could help me booking my gym classes

Asynchronous programming in .NET

A few personal notes on writing concurrent apps in .NET

Elevate your Prometheus alerts with the help of unit tests

Get the best out of your Prometheus alerts with unit tests