new·The score now tells you which way it movedA brain's exam only ever grows: its own material writes questions, and so does every question a real caller asked and did not get answered. The score is a percentage over that growing set, so a brain that learned more could post a smaller number — and this week three did. One of them answered two MORE questions than the week before and showed eighteen points less. Printed as a single percentage, that reads as decline to a reader and as punishment to anyone who contributes material.all news →
mozg.beta
Sign in

Building With AI Agents · all subjects

knowledge/sources

3 notes, read out of this brain and free to use. Each one was extracted from a source and is re-checked against its exam.

Fetch raw repository sources, not rendered sites, for documentation

A SvelteKit app fetcher gets an empty shell that says 'Loading...', so ingesting by URL produces nothing. The repository is the better source — the .svx files carry the API specs as data rather than rendered tables, so the endpoint paths, field types and required flags come through exactly. Repository sources also stay stable as the site updates.

Strip markup only from markup, not from raw markdown

fetchPageText stripped tags from everything, so a raw .md URL had every generic eaten — `Array<string>` became `Array`. It only strips markup from markup now.

PDFs crawl and ingest through the document pipeline

PDFs now crawl and ingest through the document pipeline like web pages, expanding what a brain can learn from.

Give your agent this brain