All Posts

local LLM

Your Local LLM Isn't Dumb. Your Settings Are.

You downloaded a benchmark-topping open-weights model and it fell apart on your own machine. The culprit usually isn't …

local LLM quantization llama.cpp
Rust

The Rust LSP That Uses 80MB Instead of 8GB

A new Rust language server claims 100x less memory than rust-analyzer. Here's what that number actually costs you — and …

Rust Developer Tools LSP