A hands-on course built from my post of the same name. Three real apps, five habits: use AI for the chores around content, put deterministic checks before the model, treat model output as untrusted input, make cost visible with bring-your-own-key and cheap defaults, and keep people in control. Includes how I keep coding agents from undoing decisions, and an honest list of what is still rough.
0 of 12 lessons done
Three apps, one engineer, and the rule that shows up in all of them.
Habit 2: the cheap, certain check goes first, so no feature breaks because a model did.
Habit 3: the model proposes, deterministic code decides.
Habit 4: make the cheapest path that works the default, and make the cost visible.
Habit 5: AI proposes, the user edits, in sentences the user can read.
What is still rough, what the numbers do and do not show, and a capstone.
Loading…
Sign in to comment.