Short, honest explanations of how the machinery actually works — built as narrated slide decks you can read in silence or listen to. No hype, no hand-waving, and no pretending a thing is understood when it isn’t.
Lessons marked planned are listed so the shape of the series is visible. They do not open yet, and they are not counted below.
What the machine is doing, from the inside. Start here if you use these tools daily and have never been told how they work.
18 slides · ~15 minA history of large language models, followed forwards: a 1950 parlour game that measured how predictable English is, a compression trick from 1994, a Victorian argument about meaning, and the 2017 paper that made the modern ones possible. Ends where it starts — with what actually happens when you press enter.
Hallucination, context limits, and why a confident answer and a correct answer are produced by exactly the same process. Not written yet — listed here so it isn’t mistaken for something you can open.
Getting found, and telling the advice that works from the advice that merely travels well.
14 slides · ~16 minA widely-shared local-SEO system — profile categories, review velocity, the page-two goldmine in Search Console — taught in full, then checked. Four of its claims do not survive the check: geotagged photos do nothing, the photo statistic is from 2014 and has been withdrawn, the seven-day post expiry no longer exists, and Google never said what it is quoted as saying about review replies. One prompt can get a site penalised outright. The real lesson is the checking.
Using the OS itself — Remote, Code, the bridge, and what each one is actually for. This category is empty. It is here because the lessons are planned, not to make the list look longer.