Tired of your AI losing its edge halfway through a project? Here is a practical look at how I built a modular context harness using local models, YAML configs, Chroma semantic search over public domain literature, and LangChain to turn Lessy from a basic chatbot into a real collaborator.

The painting queue is finally complete! My Space Wolves Scout squad is battle-ready, and a pack of deadly Tyranid Raveners has emerged from the hobby desk looking every bit as terrifying as they should. The Raveners were an especially fun project, marking my first time painting Tyranids and giving me the chance to experiment with a vibrant color scheme that really brought the models to life. With both teams finished and the custom Killzone: Volkus terrain complete, it’s time to show off the results and get these forces onto the tabletop for their first Kill Team battles.

I am officially diving headfirst into the tactical skirmishes of Warhammer Kill Team 3rd Edition, and my hobby desk has never been busier! After putting the finishing touches on a rugged Space Wolves Scout squad, my brushes are now flying across a terrifyingly swift team of Tyranid Raveners. To bring these factions to life, I even 3D printed and painted a custom Killzone: Volkus replica, creating the perfect grimdark backdrop for the epic tabletop battles to come.

I tested a 0.2mm nozzle on my Bambu X1 Carbon to see if it could deliver resin-like detail without the hassle of resin printing. After printing, priming, and painting a detailed miniature in a single day, I was genuinely impressed with the quality and ease of the process.

I decode and “hack” a consumer weather station’s radio broadcasts using rtl_433 and Universal Radio Hacker to show how raw RF transmissions become readable temperature, humidity, and wind data. This project is my first step toward understanding the binary data encoded and transmitted over radio airwaves.

In this post, I explore the math behind Prometheus’s deriv() function, which calculates the derivative of time series using least-squares regression. This method estimates the slope of each series in a range vector, as explained in Prometheus: Up & Running, 2nd Edition.

Just finished Chapter 14 of Prometheus: Up & Running, which dives into Aggregation Operators. This chapter covers standard deviation and quantile functions, and as usual, I’m using Python to break these concepts down further for a deeper understanding.

In this blog post, I explore Prometheus subqueries, a powerful feature in PromQL that allows you to run instant queries over a range with specific resolution. After reaching Part 4 of Prometheus: Up & Running, 2nd Edition, I break down subqueries step by step to help you fully understand how they work and how they can be used effectively.

I recently 3D printed the Ulvheim Small Merchant House on my Bambu Labs X1 Carbon, and the results were incredible. The model’s intricate details, from wooden beams to stone textures, were perfectly captured, making it a fantastic addition to any D&D game.

Reading through the book Prometheus: Up & Running, 2nd Edition, highlighting its practical, no-fluff approach even for experienced users.