I recorded the requests Codex generated for a 16-character prompt and measured what changed as it loaded instructions, exposed tools, read files, ran commands, received images, and compacted its history

How I go from a technical question I cannot answer to an explanation I can reconstruct without the model

A from-the-ground-up walkthrough of how modern LLMs work, from tokens to transformer blocks to the next-token loop

Why I'm shifting from vulnerability research to machine learning security

A walkthrough of CVE-2026-1678, a critical out-of-bounds write in Zephyr's DNS name parser caused by a stale bounds check

How CPU caches work, explained with a simple office analogy.