How this tutorial works, and why every lesson looks the same

In this lesson you'll learn:

• what each of the seven slots in a lesson is for;
• which learning research each slot comes from;
• how to move through the tracks, whatever your starting point.

Sound familiar?#

• You have read three "complete guides" to agentic AI and still could not tell a colleague what to do on Monday.
• Each article defined the same terms differently, and none said where its ideas came from.
• You wanted the part for your role, and got a tour of everybody else's.

This tutorial is built against all three: it ends every lesson on something you can do, it credits every idea, and it lets you enter at your role.

The seven slots, and why each one is there#

Step 1 · The answer comes first#

The first box in every lesson answers the question in the title, completely, in one paragraph. Newspapers settled on this order, the most important fact first, around the turn of the twentieth century, partly because a telegraphed story could be cut off at any point. A reader arriving from a search is in the same position: they may stop after one paragraph, so that paragraph has to be the whole answer.

Step 2 · A picture, before the detail#

Every lesson carries at least one picture, and the rule is strict: a board, a figure or a diagram, never decoration. People learn more from words and pictures together than from words alone (the multimedia principle), provided the picture carries the idea rather than illustrating it. On the site the boards are live; on the wiki they are screenshots of the same boards.

Step 3 · The problem, stated rather than asked#

"Sound familiar?" lists three symptoms in plain statements. It is there so you can decide in ten seconds whether the lesson is for you, and so that the explanation that follows has something to attach to.

Step 4 · Step by step, one idea at a time#

The core of a lesson is short numbered steps, each carrying one idea. Splitting material into learner-paced segments helps people who are new to it (the segmenting principle), and cutting anything that would survive unchanged in a lesson on a different topic helps everyone: the coherence principle. That second rule is why these lessons are short.

Step 5 · Try it: commit before you look#

Each lesson ends its teaching with one problem, and the answer is hidden until you open it. Trying to recall or apply an idea strengthens memory of it more than reading it again (the testing effect), and the benefit holds even when your first attempt is wrong.

Step 6 · Three takeaways#

Three short, parallel statements: what to keep if you keep nothing else. They are written to be quoted, and to be the thing you check yourself against a week later.

Step 7 · Sources, with an origin mark#

Every lesson ends with a table crediting each idea as one of three:

The marks let you argue with the right things. A borrowed statistic is checkable in a minute; an original method is a starting point, not a standard. The full list of every claim and its standing is on Sources and Confidence.

Where to start#

Each lesson shows its reading time, level and position in its track at the top. Reading time is counted at 238 words a minute, the average for adults reading non-fiction silently, plus a few seconds for each picture.

Two copies of one source#

The lessons are written once and published in two places:

• On the site, under /learn/, with the live boards. This is the copy search engines index.
• On the GitHub wiki, beside the manual's full reference pages, with screenshots of the boards.

There are two because GitHub does not let search engines index a wiki unless the repository has at least 500 stars and its wiki is closed to public editing. The site copy is canonical, and each lesson is also available as plain markdown (add index.md to its address) for tools that read text.

Key takeaways#

1. Every lesson has the same seven slots in the same order, so you always know where the answer is.
2. Each slot comes from learning research: answer first, a picture, one idea at a time, try before you look.
3. Every idea is marked Borrowed, Adapted or Original, so you know what to check and what to tune.

FAQ#

How long does the tutorial take?#

Each lesson takes five to ten minutes, and the five interview question banks eleven or twelve. The start page shows each track's total, and the fundamentals take a little over an hour. There is no need to read them in one sitting; each lesson stands alone.

Is it free?#

Yes. The lessons, the manual, the simulator and the workbench are free and open source under the MIT licence. Keep the attribution if you reuse them.

Do I need an AWS account, or any particular tool?#

No. The agentic PDLC is method- and vendor-agnostic, and nothing here needs an account. The repository's hands-on curriculum uses AWS, and some examples name AWS services where they are the concrete case.

Can I use it with my team?#

Yes. The lessons work as pre-reading for a weekly session, one track at a time; the study plans include a twelve-week reading-group format.

Apply it in your role#

Across the enterprise. Use the tracks as role onboarding: two weeks per role, one lesson a day, each Try it worked aloud in a team session. The same shared vocabulary is what the gates and reviews rely on.

The ten-minute workflow. Make a lesson test you rather than inform you:

Read this lesson: <paste or link>. Write three new problems in the style of its "Try it", with
different numbers, and hide the answers. When I reply, mark my working, not only the result, and
tell me which step I would have got wrong at work.

Sources and credits#