AI for Everyone: Downloads and Practice Resources

Keep the lessons and evidence together

Use these files to read offline, revisit an exercise, or inspect the source behind a teaching claim. The manuscripts were prepared and audited on September 8, 2026. Product features and external requirements can change after that date.

Read offline

Complete audited manuscript (Markdown): all 132 articles in one searchable file.
Reader guide (Markdown): learning paths, student guidance, and a plain-language glossary.
Complete audited download (ZIP): the fourteen batches, complete edition, guide, audit, source ledger, and four corrected labs. Keep the extracted files together so their local links work.

Four inspected practice labs

Lesson	Download	What it exercises
Article 100: first model	first_model.py · recorded result	Grouped data separation, fitting, baseline comparison, and measured classification results.
Article 118: MCP	mcp_lab.py	A small local client/server exchange and selected protocol errors.
Article 119: bounded agent	agent_lab.py	Controller behavior with simulated proposals and malformed records.
Article 123: retrieval	retrieval_lab.py	Keyword retrieval, explicit access filtering, exceptions, and conflicting evidence.

Read each article’s setup instructions before running its code. The model lab records its package versions. The other three labs use the Python standard library. These examples do not require a paid model API. They test defined local behavior, not every property of a production system.

Audit and sources

Audit report: corrections, checks performed, and the scope of verification.
Source retrieval ledger (CSV): sources, article references, check dates, and access observations.

Six source URLs had access or extraction limitations during the audit. The ledger reports those limitations; a successful retrieval is not a certification of every claim. Follow the citations within each lesson to inspect the evidence supporting the point you are using.

The revised model lab reports 30/30 correct test predictions on one small split. Its earlier edition used different test examples. The change is not evidence that the model improved.

A simple record for your own projects

Record the task, inputs and sources, assistance or code used, actual result, checks performed, and one remaining limitation. Students should also record an independent demonstration and follow the assignment’s disclosure requirements.

Choose a learning path · Return to the series