# Syllabus reading library

Updated 5 October 2026.

**Start with [the combined EPUB](Syllabus%20-%20Complete%20Reading%20Collection.epub).** It has consistent typography and navigation: three questions → twelve themes → readings. Each reading links to its theme, all themes, and adjacent readings. The combined EPUB contains reflowable text and optimized inline illustrations, without full-page PDF images.

- 87 syllabus reading entries: 85 captured readings, 1 preview, and 1 unavailable full text.
- Two supplementary author links are separate from the readings; one profile is captured and one remains unavailable.
- 87 individual EPUBs are in the twelve theme folders, alongside the combined edition.
- The contact email in the syllabus is not a reading.

## Supplied PDFs

The PDFs in `redown` completed the Axios AI statement, the TorontoToday article, and the published AI Regulation paper. The supplied OSI-inspired stack PDF is byte-for-byte identical to the previously captured public version; it is incorporated without duplicating the reading. The individual EPUBs and the combined edition use these supplied PDFs. Website menus, unrelated recommendations, and overlapping print lines were removed from the Axios and TorontoToday reading text. Original PDFs remain available in the source folders. The individual EPUBs retain their previous full-page reference images; the combined edition is compact.

## Conversion notes

The combined EPUB uses HTML-style reflowable text, with optimized inline illustrations extracted from raster images and vector regions. Whole-page PDF facsimiles are excluded. Figures are automatically identified; complex columns, tables, footnotes, and mathematical symbols may not transfer perfectly. Refer to the original PDFs for source layout. Images are limited to 1200 pixels on their longest side and use whichever is smaller: optimized JPEG or PNG. The combined edition is about 22.7 MiB.

Web articles include available local images. Missing illustrations remain online links. Interactive features and external linked pages are not included. The MCP specification and GitHub repositories contain their linked overview pages, not all subpages.

Items 58 and 76 use public author preprints and may differ from the published journal versions. Item 58’s latest arXiv version is titled “Large Language Models, and LLM-Based Agents, Should Be Used to Enhance the Digital Public Sphere”.

`index.html` is a clickable catalog; `download-index.csv` and `manifest.json` record every link and source. Original downloads are in `_sources` and `redown`. The offline build script is `_tools/build_library.py`.

## Remaining exceptions

- **PARTIAL — How AI Companies Can Pay Fair Rates for the Content They Need**: Only the publicly accessible summary and opening excerpt were available; the full article requires publisher access. [Source](https://hbr.org/2026/06/how-ai-companies-can-pay-fair-rates-for-the-content-they-need)
- **UNAVAILABLE — Restraining the formation of filter bubbles with algorithmic affordances: Toward more balanced information consumption and decreased attitude extremity**: The source blocks automated access (HTTP 403); no full text retrieved. [Source](https://asistdl.onlinelibrary.wiley.com/doi/epdf/10.1002/asi.24988?saml_referrer)
- **UNAVAILABLE — Rodrigo-Ginés**: The source blocks automated access (HTTP 403); no full text retrieved. [Source](https://www.researchgate.net/profile/Francisco-Javier-Rodrigo-Gines?_tp=eyJjb250ZXh0Ijp7ImZpcnN0UGFnZSI6InB1YmxpY2F0aW9uIiwicGFnZSI6InB1YmxpY2F0aW9uIn19)