← Blog

Welcome to the blog

A short hello and what to expect from the writing here.

This is the first post. The plan is short, opinionated notes on running large language models locally — what hardware actually matters, what to ignore, and where the gap between "fits in memory" and "useful at the desk" lives.

Posts are written as plain markdown in content/blog/. Each one has a title, description, and date in the frontmatter. The index sorts by date.

What to expect

  • Practical hardware notes: bandwidth, quantization, headroom.
  • Short reads. No 3,000-word listicles.
  • Occasional reproducible benchmarks, when there's something worth measuring.

If you have a topic you'd like to see covered, open an issue.