Guides
Not reference documentation. Each one covers a subject that is easy to get wrong, explains why, and shows the numbers or the tool that settles it. 5 so far, by subject.
Backend and web engineering, in detail.
These are the topics I keep having to re-explain: what a token actually contains, why a schedule fires more often than intended, which identifier to put in a primary key. Each guide covers one thing properly and hands you the tool that goes with it.
Models compared on numbers, not vibes.
Model marketing is a wall of adjectives. These guides plot the actual figures instead, so a claim is something you can look at. Every chart states where its numbers came from and when they were last checked.
All AI guides→The math that shows up in real code.
Most of the math a programmer meets arrives disguised as a bug. These guides work through the ideas visually, starting from the code that made them matter.
All math guides→Progress, revenue and mistakes get posted as they happen rather than saved up for a launch.