How we make Learn lessons

What every lesson has

• What you will learn, and which lessons to read first.
• The idea explained in our own words, with worked examples that include an edge case and a common mistake.
• For a lesson with code, the versions its examples run with, such as “Examples run with Python 3.14.8 and Pyodide 314.0.7”.
• References to the official documentation, standards or papers its facts come from.
• Links to the tools of this site that help with the topic, where there are any.
• A note on how the lesson was made, and a link to report a problem with it.

Written and checked with AI assistance

MySmartCoPilot writes its lessons with the help of AI models, and checks them with AI assistance too. A model drafts each lesson from an outline that sets its objectives, its examples and the official documentation it must follow, and the draft is revised until it passes every check on this page. Every lesson made with AI assistance says so on the page. No person signs off each lesson before it is published: what stands behind a lesson are the checks and the review described here, and the corrections you send us.

Lessons are our own writing. We do not copy or closely paraphrase other tutorials, books or courses, and we do not reuse their tables of contents. Every source we consult is listed in a licence register; material under non-commercial, no-derivatives, share-alike or all-rights-reserved terms is only cited and linked, never adapted. An automatic originality check compares each lesson with snapshots of those sources and with every other lesson, and fails it if it shares a run of twelve or more words with any of them outside a quotation.

Every example is run automatically

Example programs are never typed into a lesson page. Each one is a real file that our verification step runs on a pinned toolchain, the exact versions named on the lesson, in an isolated run without network access. The output under an example is what it printed, with the command that reproduces it. Every example runs twice, and an example whose output differs between the two runs fails, unless the lesson says why its output varies (a timing, say, or the name of the operating system). Files that an example only reads, such as a configuration file, and short commands quoted in the text, such as an install command, are not run on their own: like every other fact, they are checked against the official documentation.

Examples that you can also run in your browser are run in the same runtime your browser uses as well (Pyodide, for Python), and when the two print something different, the lesson shows both and explains why. The site’s build refuses to publish a lesson whose recorded output no longer matches its example files or its pinned versions, and when a lesson moves to a new release of its language, every example is run again on that release first.

Facts from official sources

Every fact is checked against official documentation, a standard or a primary paper, which the lesson lists under References. Claims that depend on a version name that version. Case studies about real companies use public information only and say that MySmartCoPilot is not affiliated with the company.

Checked before it is published

Before a lesson is published, it gets an adversarial review: a reviewer that did not write it, an AI model that works from the lesson and its sources only, tries to break it. It checks every factual sentence against the source the lesson cites, runs every example again and probes it with an extra edge case, and looks for outdated advice and for security problems in the sample code. Whatever it finds is fixed and checked again.

Automatic checks run on every lesson before it is published, too: the format of the lesson and its links; the wording rules (no dates, no promises of jobs, marks or results, no invented learners, ratings or reviewers, and inclusive terms); that the free lesson text offers nothing paid and hides nothing; that every diagram has a written description; and the originality check above.

The review is recorded with a fingerprint of the lesson’s content. If the lesson changes afterwards, even by one character, the fingerprint no longer matches, and the lesson needs a new review before it can be published again.

No invented people

Lessons are published under the name MySmartCoPilot. We do not invent authors, reviewers, credentials or photos, and lessons make no claims about experts who are not named. If a named person writes or reviews a lesson, the lesson names them and links to their author page, with links to their public profiles.

Free to read, for learners everywhere

Every lesson’s text, diagrams and examples are free to read, with no sign-up, and nothing on a lesson page is hidden behind a paywall. Your progress (lessons you read or marked complete) is kept in your browser; you can export it to a file and import it in another browser.

Learn is written for learners everywhere. Material that applies to one country only, such as campus placement tests, GATE or Indian personal finance, is labelled as such.

What the checks cannot promise

Run examples, automatic checks and an adversarial review catch many mistakes, but not every one. A lesson can still be wrong, unclear or out of date, most often just after a new release of the language or tool it teaches. Check facts that matter to you against the official documentation the lesson links to, and tell us when something is wrong.

Corrections

To report a mistake, use the contact form with the topic “Report a problem” or write to [email protected] , and give the lesson’s address and what is wrong. Every lesson also has a “Report a problem with this lesson” link that opens the form.

We check each report against the sources. When the lesson was wrong, we correct it and run its checks again, and a substantive correction is listed under the lesson in a Corrections section, so readers can see what changed.