What is really inside a language model, why it works on numbers and not on meaning, how transformers are structured, and what distinguishes a model from an agent. From the perceptron to today, and the limits that no demo shows.
What a SQL injection really is, why prepared statements eliminate it, and how to use OWASP ZAP and Burp Suite to look for it in a systematic way. With ethics and hacker identity: the difference between those who break things and those who help, written without jargon.
A file with an appetising name, a false credential, a DNS that does not answer: none of this protects an asset, and yet it tells you who got in. The story and use of canary tokens, from cartographers to AWS containers.