About this course
Follow the decision
Section titled “Follow the decision”A model asks to call a function. Python checks the request, decides whether it is allowed, and runs the code. The result goes back to the model. Another decision follows.
That small exchange raises bigger questions. Where does state live? What happens when a request fails? Who grants permission? How do we know the agent did the right thing?
Inside the Agent follows those questions through ten lessons in ordinary Python. Each lesson makes another part of the system visible: a structured response, a requested tool, a loop, an approval record, or evidence about what actually happened.
Understanding that stays useful
Section titled “Understanding that stays useful”I’m Max Braglia, the creator and curator of this course. These lessons are part of my own learning, and they have helped me tremendously in my development work.
I learn by following the details: what goes into a call, what comes back, which branch runs, and what changes before the next step. A concrete example gives me something I can reason about while working on real software.
Rereading matters too. I can return to an explanation after using the idea, notice a connection I missed, and absorb a little more. I want these lessons to remain somewhere I can find them, read them again, and keep building that understanding.
Publishing them also lets me share that experience with other developers who like to learn this way: patient explanations, small examples, and a walk through what actually runs.
A companion to Learn the Monolith
Section titled “A companion to Learn the Monolith”Learn the Monolith follows a request through a Python web application. Inside the Agent follows the boundary between model-generated decisions and application-controlled actions.
The courses share a teaching approach and a familiar place to read. They explore different questions about software, and either can be read on its own. Together, they reflect the same habit: follow the code until the system makes sense.
Authorship and review
Section titled “Authorship and review”ChatGPT wrote the original daily lessons. Codex helped review the technical details, correct the examples, reconcile the curriculum, and build the reading site. I reviewed the material for accuracy and for the teaching style that made it useful to me, and I decide what belongs in the publication.
The examples use fictional situations and data. Start here explains the prerequisites, technical baseline, and limits of the isolated checks used to review the code.
A companion to Learn the Monolith. Created and curated by Max Braglia.