Lrnon

What's new

Every change to the learning content, dated, including the ones that make us look bad. A lesson that was wrong is more useful to you as a correction than as a quiet edit.

August 2026 · 20 changes

  1. Content updated Whole site

    Corrected how the site describes certification, which it had been promising in more places than it could deliver. There are two things, and they were being said as one: a Module Record, earned by finishing every lesson, which is issued today; and a Lrnon Certificate, which additionally requires passing that module's assessment, and which nothing on the site can issue yet. The landing page said 'free certificates', E7 said it was 'compulsory for certification', and the achievements page listed the two together. All of that now names the tier and says what it takes. Two words are retired: 'compulsory', because nothing on Lrnon is, and 'certified' as a verb, because it implies an authority we are not. The entry below of 26 August is left as it was published, wrong claim included — a correction belongs in a new dated entry, not in a quiet edit to an old one. Anyone who finished E7 under the old wording will be told directly when the assessment opens, and will not be charged to sit it.

  2. Accessibility Whole site

    Every lesson that names a tool now shows where the description came from and when it was last checked, with a warning when that is over 120 days old. A new About page states the position on brands and sources: no sponsorship ever, neutral rather than silent, and the vendor is always the authority over us.

  3. Content updated Explorer · E2 and E13

    Market review of the whole Explorer track. Added E2·L11 on open-weight and regional assistants — DeepSeek, Llama, Mistral and Qwen now have registry cards, closing a gap where E10 and E11 both referenced a tier the landscape module never covered. Added E13·L9 on whether AI use counts as cheating, which research showed is the question learners actually ask.

  4. New module Explorer · E14

    Health, Money and Legal published, completing the Explorer track at 14 modules and 112 lessons. Eight lessons on preparing for appointments, why symptom questions spiral, financial concepts versus financial advice, understanding legal documents, the four hard stops where AI is the wrong tool entirely, getting to a real person, and six rules for high-stakes questions.

  5. New module Explorer · E13

    Using AI to Actually Learn published — eight lessons on the illusion of understanding, getting explanations pitched at you, retrieval practice, spacing and forgetting, teaching it back, where AI is a bad teacher, what explanation cannot teach at all, and a study loop that survives a bad week.

  6. New module Explorer · E12

    AI You Didn't Choose published — eight lessons on AI embedded in tools you already had: spotting it, turning it off, the AI in your phone, features that read your messages, decisions made about you and your rights, workplace tooling, and a one-hour settings audit.

  7. New module Explorer · E11

    What It Actually Costs published — eight lessons on tokens as the unit you pay in, why bills move, right-sizing the model, the time budget nobody counts, getting real work done for nothing, paying from outside the dollar economy, spending caps before automation, and judging honestly whether it was worth it.

  8. New module Explorer · E10

    Working in Your Own Language published — eight lessons on why quality varies by language, testing your own, back-translation and what it cannot catch, code-switching, names and scripts, honorifics and register, learning a language, and running a two-language working life.

  9. New module Explorer · E9

    AI as an Assistive Tool published — eight lessons on reading, writing, seeing, hearing, focus and communication when those are barriers, plus how to make things other people can actually get at. Written after an audit found the curriculum had no coverage of assistive use at all.

  10. New module Practitioner · P1

    Advanced Prompting published — the first Practitioner module. Eight lessons on few-shot examples, structured output, splitting a prompt into single-purpose steps, chaining without compounding errors, persistent instructions, long inputs, testing a prompt against a saved evaluation set, and building prompts that survive a model update.

  11. New module Explorer · E8

    AI Ethics & Responsible Use published — eight lessons on authorship, disclosing AI use, where bias comes from and lands, who answers when something goes wrong, the environmental and labour costs, automating work that belongs to colleagues, how to refuse, and writing your own code of conduct. This completes the Explorer track: 64 lessons across eight modules.

  12. New module Explorer · E7

    Truth, Verification & Critical Thinking published — eight lessons on why fluency defeats scepticism, triaging which claims to check, a two-minute verification method, judging sources, handling disagreement, how manipulation finds you, your own bias, and a habit small enough to survive. This module is compulsory for certification.

  13. New module Explorer · E6

    Creative AI: Image, Video & Voice published — eight lessons on how diffusion models actually work, describing an image, iterating with seeds, the artist-style question, the current limits of video, voice cloning and consent, labelling what you made, and the categories that are not grey areas.

  14. New module Explorer · E5

    Documents, Spreadsheets & Presentations published — eight lessons on outlining before prose, drafting in your own voice, using an assistant as a critic, spreadsheet formulas you can test, recomputing every number, assertion-headline slides, rewriting for a different reader, and the final consistency pass.

  15. New module Explorer · E4

    Everyday Productivity published — eight lessons on where AI actually fits in your day, email, reading long documents, research that cites its sources, meeting notes, planning, a twenty-minute routine, and what not to hand over.

  16. Accessibility Whole site

    Reading preferences added: text size, line spacing, line length, contrast, a reading typeface and data saver. Dark mode also now follows your device — it had been built but never switched on.

  17. Fixed All quizzes

    Answer options are now shuffled. 94% of correct answers had been sitting in the middle slot, so always picking the middle passed every module quiz without reading the question.

  18. New module Prompting Fundamentals · E3

    Eight lessons on briefing, iteration, correcting a wrong answer, and the requests no wording can rescue. Adds a 22-question quiz bank and 24 review cards.

  19. Content updated Meet Your AI Assistants · L8

    NotebookLM was renamed Gemini Notebook by Google in July. The lesson title and the tool card now say so, and the documentation link it pointed at had been dead for six weeks.

  20. Lesson corrected Meet Your AI Assistants · L3

    Claude's tool card claimed no voice support. Voice mode has shipped on every plan, so the capability grid was showing learners a wrong fact.

Lessons you have finished get a note when they change

Progress is recorded against the exact version of a lesson you read. If a lesson you completed is corrected, the change is visible here — so you can relearn the fix rather than keeping the mistake. Your daily review →