Learn

By the end you will be able to:

Loading…

Why this is not the archive

The daily scan tells you what happened; this tells you how something works, and the two need opposite machinery. News is ordered by date and goes stale. A sequence is ordered by what you need to know first and nothing in it expires — so this stream is sequenced rather than reverse-chronological, and it does not reuse the scan's gate, its ordering or its ACT/WATCH/NOTE ladder.

How a step gets here

A person chooses the anchor; the machine only verifies it. That is the opposite of the daily scan, and it is forced by the corpus rather than chosen: automated slot-filling scored 12 correct out of 30, because a machine filtering on "is it free" cannot tell a paper that carries the argument from a paper that merely matches the words, and backfills the gap with off-domain work. The first attempt returned a history paper for what an agent is.

That is an argument about who does the choosing, not about what a reader should have to pay. Where a step's canonical paper turned out to be locked, the answer was to go and find free work that carries the same lesson — usually in ITcon, which is this field's fully open journal, or on arXiv — and rewrite the step's explanation to describe the paper that is actually there. Sixteen anchors moved that way. None of them lost their argument, and several got a better one.

So each anchor carries two labels, because can I read this and may TechScan quote it are different questions and one label used to answer both badly. free read is the first and the one that matters to you: a copy exists that costs nothing — publisher open access, a university repository, an author's accepted manuscript, or an arXiv twin. Every step in every track carries it. Where the canonical paper was locked, the anchor was replaced with free work that teaches the same thing, and the explanation rewritten to describe the paper that is actually there.

The second label is the licence, and it only decides what appears on this page. open means a CC licence, so we may quote from it. cited means we may not — you still read it free, but the summary here is ours rather than an excerpt, and never a quote reconstructed from metadata. An arXiv preprint is the ordinary case: free to everyone, not CC. needed is a step whose anchor has not been chosen yet; it is published with its search hint rather than hidden, because a curriculum that quietly skips a step is worse than one that admits a hole.

Every teaches and why line is written by a person. No model writes prose about a paper it has not read. A wrong news summary gets caught on the click-through; a wrong explanation gets believed and carried into a decision.

Track definitions live in learn/track-*.json; the resolved records are written by scan/learn.py into learn.json. Neither is edited by hand.