Insights & Articles
A collection of technical deep-dives into AI, LLMs, and automation workflows.
Subscribe via RSS
AILLM
July 21, 2026
The Crucible #2: How Language Models Read Text
How raw text becomes the numeric representation a transformer can process. Building a tokenizer by hand, watching it break on an unfamiliar word, and fixing it the way GPT actually does: with Byte Pair Encoding and positional embeddings.
Read Article
AILLM
July 10, 2026
The Crucible #1: Why I'm Building an LLM From Scratch
A personal journey to understand large language models from first principles by building every major component from scratch. This post explains why I'm starting The Crucible, what I hope to learn, and the roadmap for the series.
Read Article